The Routledge Handbook of Philosophy of Agency [1 ed.] 9781138062849, 9781032182254, 9780429202131

Citation preview

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF AGENCY

One of the most basic and important distinctions we draw is between those entities with the capacity of agency and those without. As humans we enjoy agency in its full-blooded form and therefore a proper understanding of the nature of agency is of great importance to appreciate who we are and what we should expect and demand of our existence. The Routledge Handbook of Philosophy of Agency is an outstanding reference source to the key issues, problems, and debates in this exciting subject and is the first collection of its kind. Comprising 42 chapters by an international team of contributors, the Handbook is divided into eight clear parts: • • • • • • • •

The Metaphysics of Agency Kinds of Agency Agency and Ability Agency: Mind, Body, and World Agency and Knowledge Agency and Moral Psychology Agency and Time Agency, Reasoning, and Normativity.

A broad range of topics are covered, including the relation of agency to causation, teleology, animal agency, intentionality, planning, skills, disability, practical knowledge, selfknowledge, the will, responsibility, autonomy, identification, emotions, personal identity, reasons, morality, the law, aesthetics, and games. The Routledge Handbook of Philosophy of Agency is essential reading for students and researchers within philosophy of action, philosophy of mind, metaphysics, philosophy of psychology, and ethics. Luca Ferrero is Professor of Philosophy at UC-Riverside, USA. He works on the nature of diachronic agency and rationality, intentions, constitutivism, and personal identity. He is the editor of the Philosophy of Action section of the Stanford Encyclopedia of Philosophy.

ROUTLEDGE H A N DBOOKS IN PHILOSOPH Y

Routledge Handbooks in Philosophy are state-of-the-art surveys of emerging, newly refreshed, and important fields in philosophy, providing accessible yet thorough assessments of key problems, themes, thinkers, and recent developments in research. All chapters for each volume are specially commissioned, and written by leading scholars in the field. Carefully edited and organized, Routledge Handbooks in Philosophy provide indispensable reference tools for students and researchers seeking a comprehensive overview of new and exciting topics in philosophy. They are also valuable teaching resources as accompaniments to textbooks, anthologies, and research-orientated publications. Also available: THE ROUTLEDGE HANDBOOK OF PHILOSOPHY AND IMPROVISATION IN THE ARTS Edited by Alessandro Bertinetto and Marcello Ruta THE ROUTLEDGE HANDBOOK OF IDEALISM AND IMMATERIALISM Edited by Joshua Farris and Benedikt Paul Göcke THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF ECONOMICS Edited by Conrad Heilmann and Julian Reiss THE ROUTLEDGE HANDBOOK OF LOGICAL EMPIRICISM Edited by Thomas Uebel and Christoph Limbeck-Lilienau THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF AGENCY Edited by Luca Ferrero THE ROUTLEDGE HANDBOOK OF PROPOSITIONS Edited by Adam Russell Murray and Chris Tillman For more information about this series, please visit: https://www.routledge.com/ Routledge-Handbooks-in-Philosophy/book-series/RHP

THE ROUTLEDGE HANDBOOK OF PHILOSOPHY OF AGENCY

Edited by Luca Ferrero

Cover image: © Getty Images First published 2022 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2022 selection and editorial matter Luca Ferrero; individual chapters, the contributors The right of Luca Ferrero to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Ferrero, Luca, editor. Title: The Routledge handbook of philosophy of agency / edited by Luca Ferrero. Description: Abingdon, Oxon; New York, NY: Routledge, 2022. | Series: Routledge handbooks in philosophy | Includes bibliographical references and index. Identifiers: LCCN 2021034958 (print) | LCCN 2021034959 (ebook) | ISBN 9781138062849 (hbk) | ISBN 9781032182254 (pbk) | ISBN 9780429202131 (ebk) Subjects: LCSH: Agent (Philosophy) | Act (Philosophy) Classification: LCC B105.A35 R68 2022 (print) | LCC B105.A35 (ebook) | DDC 128/.4—dc23 LC record available at https://lccn.loc.gov/2021034958 LC ebook record available at https://lccn.loc.gov/2021034959 ISBN: 978-1-138-06284-9 (hbk) ISBN: 978-1-032-18225-4 (pbk) ISBN: 978-0-429-20213-1 (ebk) DOI: 10.4324/9780429202131 Typeset in Bembo Std by codeMantra

CONTENTS

Contributors

x

An introduction to the philosophy of agency Luca Ferrero PART 1

The metaphysics of agency

1

19

Introduction to Part 1 Luca Ferrero

19

1 Agency and causation Jesús H. Aguilar and Andrei A. Buckareff

27

2 Agency, function, and teleology Scott Sehon

37

3 Agency, events, and processes Matthias Haase

47

4 Negative agency Randolph Clarke

59

5 Bounded agency Elijah Millgram

68

6 Agency and games C. Thi Nguyen

77 v

Contents PART 2

Kinds of agency 87 Introduction to Part 2 Luca Ferrero

87

7 Minimal agency 91 Hans van Hateren 8 Animal agency 101 Helen Steward 9 Intentional agency 109 Lilian O’Brien 10 Rational agency 118 Eric Marcus PART 3

Agency and ability 125 Introduction to Part 3 Luca Ferrero

125

11 Agency, powers, and skills 130 Will Small 12 Expert agency 139 Barbara Gail Montero 13 Agency and mistakes 149 Santiago Amaya 14 Agency and disability 159 Kevin Timpe 15 Pathologies of agency 169 Lubomira Radoilska PART 4

Agency: mind, body, and world 179 Introduction to Part 4 Luca Ferrero

179 vi

Contents

16 Mental agency Matthew Soteriou

183

17 Agency and the body Hong Yu Wong

192

18 Agency, consciousness, and attention Wayne Wu

201

19 Material agency Matthew Noah Smith

211

PART 5

Agency and knowledge

221

Introduction to Part 5 Luca Ferrero

221

20 Epistemic agency David Hunter

226

21 Agency and practical knowledge Kim Frost

234

22 Agency and evidence Berislav Marušić and John Schwenkler

244

23 Agency and self-knowledge Brie Gertler

253

PART 6

Agency and moral psychology

263

Introduction to Part 6 Luca Ferrero

263

24 Agency, will, and freedom Thomas Pink

270

25 Agency and responsibility Pamela Hieronymi

279

vii

Contents

26 Agency and identification Agnieszka Jaworska

288

27 Agency and autonomy Andrea C. Westlund

298

28 Agency and (the limits of ) volitional conflict Sarah Buss

307

29 Agency and the emotions Carla Bagnoli

317

PART 7

Agency and time

329

Introduction to Part 7 Luca Ferrero

329

30 Diachronic agency Luca Ferrero

336

31 Planning agency Michael E. Bratman

348

32 Agency, time, and rationality Chrisoula Andreou

357

33 Artificial and machine agency Richmond H. Thomason and John Horty

366

34 Agency and personal identity Marya Schechtman

376

35 Agency, narrative, and mortality Roman Altshuler

385

PART 8

Agency, reasoning, and normativity

395

Introduction to Part 8 Luca Ferrero

395

viii

Contents

36 Agency, reasons and rationality Maria Alvarez

403

37 Agency and practical reasoning Jules Salomone-Sehr and Jennifer M. Morton

412

38 Agency and normativity Kenneth Walden

421

39 The aim of agency Kathryn Lindeman

430

40 Agency and morality Christine M. Korsgaard

440

41 Agency in the law Gideon Yaffe

448

42 Aesthetic agency Keren Gorodeisky

456

Index

467

ix

CONTRIBUTORS

Jesús H. Aguilar is Professor of Philosophy at the Rochester Institute of Technology in Rochester, New York, USA. His research and publications are mainly on action theory; recently, he has been working on a human-rights based educational model to be implemented in northern Mexico, his native country. Roman Altshuler  is Assistant Professor of Philosophy at Kutztown University of Pennsylvania. In his free time, he thinks about exercising free will and overcoming mortality. In his unfree time, he complains about lack of free will and the constraints of mortality. He has written a few papers about each of these, and co-edited (with Michael Sigrist) a book about the role time plays in agency: Time and the Philosophy of Action (2016). He is currently working on a book on the relation between nationalism, identity, and meaning in life. Maria Alvarez  is Professor of Philosophy at King’s College London. She’s the author of Kinds of Reasons: an Essay in the Philosophy of Action and works mainly on philosophy of action, reasons, and normativity and is currently working on a monograph on agency, choice, and moral responsibility. She is also a co-editor of Philosophy. Santiago Amaya is Associate Professor of Philosophy at Universidad de los Andes in Bogotá and works primarily in action theory and moral psychology. His articles have appeared in Noûs, Philosophical Studies, Synthese, and elsewhere. Chrisoula Andreou is a Professor in the Philosophy Department at the University of Utah and an Executive Editor of the Canadian Journal of Philosophy. Her current research projects lie in the areas of practical reasoning, action theory, ethical theory, and applied ethics. Carla Bagnoli is Professor of Philosophy at the University of Modena & Reggio Emilia, and Visiting Fellow at All Souls College, Oxford. She specializes in rational agency, practical reasoning, and meta-ethics. She is the author of Ethical Constructivism (Cambridge University Press, forthcoming) and the editor of Morality and the Emotions (Oxford University Press) and Constructivism in Ethics (Cambridge University Press). Her essays appear in journals such as Oxford Studies in Normative Ethics, Oxford Studies in Agency and Responsibility, Synthese, Theoria, x

Contributors

and The Canadian Journal of Philosophy. She is currently working on a dynamic account of responsibility, focused on rational and emotional self-governance. Michael E. Bratman is U. G. and Abbie Birch Durfee Professor in the School of Humanities and Sciences, and Professor of Philosophy at Stanford University, USA. He is the author of Intention, Plans, and Practical Reason (1987), Faces of Intention: Selected Essays on Intention and Agency (1999), Structures of Agency: Essays (2007), Shared Agency: A Planning Theory of Acting Together (2014), Planning, Time, and Self-Governance: Essays in Practical Rationality (2018), and Shared and Institutional Agency: Toward a Planning Theory of Human Practical Organization (forthcoming). Andrei A. Buckareff is Professor of Philosophy and Co-Director of the Cognitive Science Program at Marist College in Poughkeepsie, New York, USA. His recent work has focused chiefly on applying an ontology of causal powers to problems in the metaphysics of agency, causation, and philosophical theology. His work on agency has appeared in Canadian Journal of Philosophy, Philosophical Explorations, Philosophical Issues, Philosophical Studies, Synthese, and elsewhere. Sarah Buss  is Professor of Philosophy at the University of Michigan. She is the author of articles on autonomy, moral responsibility, practical rationality, and respect for persons. Her current projects address the normative significance of formal principles of practical rationality, the nature of reasons for action, the will’s contribution to action, and the moral implications of self-love. Randolph Clarke is Professor of Philosophy at Florida State University. He is the author of Libertarian Accounts of Free Will (2003) and Omissions: Agency, Metaphysics, and Responsibility (2014), and co-editor (with Michael McKenna and Angela M. Smith) of The Nature of Moral Responsibility: New Essays (2015). Luca Ferrero is Professor of Philosophy at UC-Riverside. He works on the nature of diachronic agency and rationality, intentions, constitutivism, and personal identity. He is the editor of the Philosophy of Action section of the Stanford Encyclopedia of Philosophy. Kim Frost is Assistant Professor of Philosophy at UC-Riverside. He works primarily on philosophy of mind, philosophy of action, and the metaphysics of powers. Brie Gertler is Commonwealth Professor of Philosophy at the University of Virginia. Her work lies at the intersection of epistemology and the philosophy of mind. She has written on self-knowledge, the mind-body problem, mental content, and the self. Keren Gorodeisky is Jane Dickson Lanier Professor of Philosophy at Auburn University. She works on the rational character of aesthetic judgment and feelings, on the nature of aesthetic value and agency, on emotions, pleasure, and on Kant. She is currently writing a book on aesthetic value, defending a neglected alternative between contemporary aesthetic hedonism and its non-affective denial. Matthias Haase is Assistant Professor of Philosophy at the University of Chicago. His research is focused on foundational topics at the intersection of ethics and philosophy of mind. He has a central historical interest in the tradition of German Idealism, especially aspects tied xi

Contributors

to Aristotle. He has also written on Wittgenstein and Frege. His current research project is devoted to the question whether there are specifically practical species of knowledge, reason, and truth—and what this means for the philosophical account of our fundamental concepts of ethics like good, ought, justice as well as action, character, and will. Pamela Hieronymi is Professor of Philosophy at the University of California, Los A ngeles. Her research stands at the intersection of the subfields of ethics, philosophy of mind, and philosophy of action. She is currently working on a manuscript seeking to unwind the traditional problem of free will and moral responsibility. John Horty is the author of a number of books and articles on logic, philosophy, and computer science. He teaches at the University of Maryland. David Hunter  is Professor of Philosophy at Ryerson University in Toronto. He works mostly on the nature of belief and epistemic agency but has research interests in desire and the varieties of goodness. His book On Believing: being right in a world of possibilities is forthcoming with Oxford University Press. Agnieszka Jaworska is Associate Professor of Philosophy at the University of California, Riverside. Her research lies at the intersection of ethical theory, medical ethics, and moral psychology. She is working on a book on the ethics of treatment of individuals whose status as moral agents and persons seems compromised or uncertain such as patients with dementia, addicts, psychopaths, and young children. Christine M. Korsgaard is the Arthur Kingsley Research Porter Professor of Philosophy at Harvard. She is the author of The Sources of Normativity (Cambridge University Press 1996), Creating the Kingdom of Ends (Cambridge University Press 1996), The Constitution of Agency (Oxford University Press 2008), Self-Constitution: Agency, Identity, and Integrity (Oxford University Press 2009), and Fellow Creatures: Our Obligations to the Other Animals (Oxford University Press 2018). Kathryn Lindeman is Assistant Professor of Philosophy at the University of South Carolina. Her main interests are in ethics, meta-ethics, and the philosophy of law, with an emphasis on questions about the source, content, and authority of normative standards. Eric Marcus works chiefly in the philosophy of mind and action, but also has active research interests in epistemology, metaphysics, aesthetics, and the philosophy of language. He is the author of numerous articles and a pair of books, Rational Causation (Harvard University Press 2012) and Belief, Inference and the Self-Conscious Mind (Oxford University Press 2021). He is Professor of Philosophy at Auburn University. Berislav Marušić is Senior Lecturer at the University of Edinburgh. He is the author of Evidence and Agency (Oxford University Press 2015), ‘Do Reasons Expire?—An Essay on Grief ’ (Philosophers’ Imprint 2018), and On the Temporality of Emotions: An Essay on Grief, Anger, and Love (Oxford University Press, forthcoming). He works on the relation between practical and theoretical reason, interpersonal epistemology, the temporality of emotions, and the philosophy of Jean-Paul Sartre.

xii

Contributors

Elijah Millgram  is E. E. Ericksen Distinguished Professor of Philosophy at the University of Utah. His research interests are centered on theory of rationality, and overlap ethics, metaphysics, and theory of agency. His most recent book is John Stuart Mill and the Meaning of Life. Barbara Gail Montero  is Professor of Philosophy at the City University of New York and a former professional ballet dancer. Her 2016 book, Thought in Action: Expertise and the Conscious Mind (Oxford University Press), explores how the body and mind intertwine in highly skilled actions. Jennifer M. Morton is Presidential Associate Professor at the University of Pennsylvania. She is the author of Moving Up Without Losing Your Way: The Ethical Costs of Upward Mobility (Princeton University Press, 2019) and of several articles in philosophy of action, philosophy of education, and moral and political philosophy. C. Thi Nguyen  is Associate Professor of Philosophy at the University of Utah. He has written on aesthetics, games, porn, trust, and echo chambers. He is interested in the way that technologies and social structures can shape our agency and rationality. His first book is Games: Agency as Art (Oxford University Press). Lilian O’Brien  is a fellow at the Helsinki Collegium for Advanced Studies and a docent at Practical Philosophy, University of Helsinki. She is working on the second edition of Philosophy of Action (Palgrave 2014) and has written several articles on the philosophy of agency. Thomas Pink is Professor of Philosophy at King’s College, London. He is the author of Self-Determination (Oxford University Press 2017) (the first of two volumes on The Ethics of Action), of Free Will: A Very Short Introduction (Oxford University Press 2004), and of many articles on ethics and moral psychology and on the history of these subjects. Lubomira Radoilska is Senior Lecturer in Philosophy, University of Kent, UK and a Senior Research Associate, Faculty of Humanities, University of Johannesburg. She works on issues at the intersection of philosophy of action, ethics, and epistemology. Radoilska is the author of Addiction and Weakness of Will (Oxford University Press 2013) and editor of Autonomy and Mental Disorder (Oxford University Press 2012). Jules Salomone-Sehr is Postdoctoral Researcher at the Centre de Recherche en Éthique and is affiliated with McGill University. He works in ethics, philosophy of action, and social philosophy. His doctoral dissertation explores the nature of cooperation and shared agency. The first chapter, “Cooperation: With or Without Shared Intentions,” is forthcoming in Ethics. Marya Schechtman  is LAS Distinguished Professor of Philosophy at the University of Illinois at Chicago, where she is also affiliated with the Laboratory of Integrative Neuroscience. She is the author of The Constitution of Selves (Cornell University Press 1996) and Staying Alive: Personal Identity, Practical Concerns, and the Unity of a Life (Oxford University Press 2017).

xiii

Contributors

John Schwenkler is Professor of Philosophy at Florida State University. He is the author of Anscombe’s Intention: A Guide (Oxford University Press 2019). His research concerns a number of topics in philosophical psychology, including perceptual experience, intention and intentional action, self-knowledge, and the linguistic expression of basic conceptual categories. Scott Sehon (PhD, Princeton, 1994) is Professor of Philosophy at Bowdoin College. He is the author of Teleological Realism: Mind, Agency, and Explanation (MIT Press 2005) and Free Will and Action Explanation: A Non-Causal, Compatibilist Account (Oxford University Press 2016), as well as numerous articles. Will Small is Associate Professor of Philosophy at the University of Illinois at Chicago. He works in the philosophy of action, and on related topics in ethics, epistemology, and the philosophy of mind. He has special interests in the topics of skill and practical knowledge, in the metaphysics of intentional action, and in the works of Elizabeth Anscombe and Gilbert Ryle. Matthew Noah Smith  is Associate Professor of Philosophy at Northeastern University. His primary areas of research are in political philosophy and philosophy of action. He recently co-authored The Spatial Contract: A New Politics of Provision for an Urbanized Planet (Manchester University Press 2020). Matthew Soteriou is Professor in Philosophy at King’s College London. He is the author of Disjunctivism (Routledge 2016) and The Mind’s Construction: The Ontology of Mind and Mental Action (Oxford University Press 2013), and co-editor (with Lucy O’Brien) of Mental Actions (Oxford University Press 2009). Helen Steward is Professor of Philosophy of Mind and Action at the University of Leeds. She is the author of The Ontology of Mind: Events, Processes and States (Oxford University Press 1997) and A Metaphysics for Freedom (Oxford University Press 2012) as well as many papers on free will, the philosophy of mind and action, ontology, and metaphysics. Richmond H. Thomason is a philosopher and logician who has also worked in linguistics and computer science. He teaches at the University of Michigan. Kevin Timpe  currently holds the William H. Jellema Chair at Calvin University. He is the author of, among others, The Virtues: A Very Short Introduction (with Craig Boyd, Oxford University Press 2021) and Free Will in Philosophical Theology (Bloomsbury 2013). He’s edited a number of books, including the Routledge Companion to Free Will (with Meghan Griffith and Neil Levy, Routledge 2017). His research focuses on free will, philosophy of religion, philosophy of disability, and virtue ethics. Hans van Hateren is Senior Lecturer in the Institute for Mathematics, Computer Science and Artificial Intelligence at the University of Groningen (the Netherlands). A biophysicist by training, he has done experimental and theoretical research on the neurobiology of vision. His current work focuses on mechanisms that can produce agency, intentionality, and consciousness. Kenneth Walden is Associate Professor of Philosophy at Dartmouth College. He works on ethics, moral psychology, and aesthetics. xiv

Contributors

Andrea C. Westlund is Professor of Philosophy at Florida State University, and the author of various papers on personal autonomy, narrative self-understanding, and the moral emotions involved in blame, forgiveness, and grief. Hong Yu Wong is Professor of Philosophy and Head of the CIN Philosophy of Neuroscience Group at the University of Tübingen, Germany. His research interests center around action, perception, and the body. Wayne Wu is Associate Professor in the Department of Philosophy and the Neuroscience Institute at Carnegie Mellon University. He works on attention, agency, perception, and related topics with special emphasis on integrating neuroscience and psychology with philosophical theory. Gideon Yaffe is the Wesley Newcomb Hohfeld Professor of Jurisprudence, and Professor of Philosophy and Psychology at Yale University. He has written books and articles on a variety of topics in the philosophy of law, and especially the philosophy of criminal law, including Attempts: In the Philosophy of Action and the Criminal Law (Oxford University Press 2010) and The Age of Culpability: Children and the Nature of Criminal Responsibility (Oxford University Press 2018).

xv

AN INTRODUCTION TO THE PHILOSOPHY OF AGENCY Luca Ferrero

1 A basic distinction Much of what happens, from the motions of the galaxies to the interactions of subatomic particles, is the mere product of the causal forces that operate in the universe—seemingly, these events have no point and they aren’t anybody’s doing: they just happen. But once in a while, especially closer home, some of the things that happen appear to have a distinctive character: they seem to unfold ‘with a view’ to some destination, to have a point to them, and be the doings of someone (or at least of something): a hummingbird drinking nectar from a flower, a dog chasing a ball, you reading this introduction, the clown making some children laugh, and your national soccer team celebrating their victory at the World Cup— just to give some examples. A similar distinction is also especially noticeable about yourself. Some things just happen to you—being knocked to the ground by a gust of wind, spontaneously bleeding from your nose, or falling asleep—but many other things you do: you get up (or at least try to) when you are knocked down, jump in celebration when your team scores, play the Hammerklavier Sonata, converse with your friends, build a house, and turn your head to admire the starry heavens above you—just to give some examples (not to mention the things that that you might be able to do, even if you are never going to do, such as, hopefully, knocking someone to the ground or punching someone so hard to make their nose bleed). For the most part, it seems that we can easily classify the things that happen, both in the universe at large and around us, between those that merely happen, with no point or purpose, and those that are done by someone or something, with some kind of point or purpose. Sure, there are going to be some ambiguous or borderline cases—when the children laugh at the clown, when you sneeze, or when you are in the grip of some compulsion, for instance. But the reason why we find these cases somewhat difficult to classify is that they seem to partake of some of the special features of the straightforward and uncontroversial examples of ‘doings’—or what we also refer to as acts, actions, or activities. The distinction between mere happenings versus doings is something that matters a lot to us. Whenever you are dealing with someitem (an object, a situation, or an occurrence)— whether by interacting with it, trying to explain it, or appraising it—you first need to determine where that item falls in the basic distinction between mere happenings and doings. Are

DOI: 10.4324/9780429202131-11

Luca Ferrero

you dealing with something that is just happening, or has no capacity to do anything, or is just suffering the effects of someone else’s doing? Or are you dealing with something that is a doing, or has the capacity to do things and possibly actively exercising this capacity? The very character and nature of your interactions, explanations, and appraisals—the category to which they belong—is affected by the preliminary classification of their objects into the two basic categories: mere happenings versus doings. Simply put, you would not try to make sense of a mere happening as ‘intelligible’ in its unfolding in terms of some of its own goals or purposes; you would not hold a mere happening or its source to any standards of success or appropriateness, and certainly not to any kind of responsibility. It would be both a category and a practical mistake to try avoiding a dog chasing you as if it were a flying ball, and vice versa. These differences become even more marked and significant when they bear on our relationship with fellow human beings (and with ourselves). This is because the explanation and assessment of, and interaction with the doings of human beings can be framed not just in purposive and teleological terms—that is, in terms of their goals or aims—but also, and much more importantly, in terms of such notions as ‘reasons,’ ‘rationality,’ ‘justification,’ ‘responsibility,’ and ‘morality.’ A dog chasing you might have some reasons to do so but it would not be cognizant of them as such. Hence, it would in principle be impossible to ask the dog for a justification of its action, and inappropriate to hold the dog accountable for it. But a human being chasing you is, in principle, expected to ‘respond’ for this action of theirs: they are in principle under a demand to offer some plausible reason in support of their behavior. If this subject were unable to provide such reasons, they should stop engaging in it or, if it is too late, they should be held accountable for that performance and its consequences (and the use of the modal ‘should’ is a major element of what is ultimately at stake here). This is not to say that everything we do is appraised and ‘held’ to such standards and expectations. But when it is not, it is because our conduct should somehow be exempted, exonerated, or excused—that is, a special dispensation is to be made against the default expectation that we are dealing with ordinary purposive and intentional conduct, with its full range of normative, legal, and moral implications.

2 Agency and philosophy The distinction between the different kinds of occurrences that I have just sketched is a fairly intuitive and familiar one. But matters get more complicated and controversial once we look at what philosophy has to tell us about this distinction. The simplest way for a philosopher to explain this intuitive distinction is in terms of the concept of ‘agency.’ The philosopher would say: some entities—agents—are capable of agency, whereas non-agents aren’t. For this to be informative, however, we would need to know more about the nature of the capacity of agency. If we just say that agency is the capacity to act or to do things, we have hardly explained anything. What we are looking for, after all, is an account of those distinctive modes of behaving that we call ‘acting’ or ‘doing.’ Things might not look much better if we consider some of the standard ways in which the concept of agency is introduced by philosophers, such as agency as the capacity to ‘make things happen,’ ‘make a difference,’ or ‘cause some kind of change.’ The trouble is that these characterizations can be used in ways that appear too broad. These expressions might refer, after all, to such things as the sun causing sunburn or the vinegar curdling the milk in the paneer. 2

Introduction to philosophy of agency

So, any quick characterization of the capacity of agency seems either uninformative or too broad. Can we do better? I suspect not, at least not at this stage. But we should not despair. As Pamela Hieronymi correctly says at the opening of ‘Agency and responsibility’ (Chapter 25), “‘Agency’ is a term of art. Its meaning and use might be discovered by reading and studying this volume.” This is indeed the spirit in which we should approach our philosophical investigation: by reading and reflecting on the topics discussed in this handbook, we should be able to better articulate the various features, dimensions, and implications of the capacity of agency. In doing so, we could temporarily rely on a rough-and-ready definition—agency as the capacity to do things or to make things happen, say—but any such definition should only be used as a stand-in for the capacity that will eventually be shown to account for the intuitive distinction between purposive (and possibly intentional) conduct and mere happenings. This is why this introduction is not going to offer a definition of agency (and why you won’t find any chapter in this handbook with such title as ‘What is agency?’ or ‘The nature of agency’—but only chapters with titles such as ‘Agency and x’ or some qualification of agency, say ‘Animal agency’ or ‘Aesthetic agency’). So the best I can do in this introduction is to point to a cluster of questions and issues that arise when discussing the differences between doing and mere happening, between making things happen rather than having things happen to you, between being active rather than passive, and so forth—with an understanding that these are just intuitive distinctions in need of further refinement as we proceed with our philosophical investigation. I will continue to use ‘agency’ unapologetically but always with the implicit understanding that it refers to what we are trying to explicate in raising these various philosophical questions.

3 The exercises of agency 3.1 Agency as a capacity Agency is a capacity: it can be possessed without being exercised or manifested (e.g., when we are temporarily unconscious, we possess the capacity but we are not exercising it). Those who possess this capacity are ‘agents’ (as opposed to non-agents), but they are not playing the role of agents when they are not exercising it (at those times, agents are rather playing the role of patients). For instance, if my arm moves because of a spasm or because someone has grabbed my arm and moved it, I am not the agent (as role) of that movement, even if I am still an agent (as a kind of being). (The distinction is in principle straightforward but there are cases where the exercise of one’s agency is partly under the control of someone else’s agency, as it might happen when someone is under the spell or the authority of another agent. In these cases, which role is one playing? Agent, patient, or a mix of both? The problem arises because, in these cases, one’s agency does not appear to be entirely bypassed but only severely constrained in its exercise.) To begin articulating the notion of ‘agency,’ one might start by looking at some paradigmatic exercises of this capacity. A common suggestion is to look at intentional bodily movements, such as raising one’s arm or crossing the street. Here are some typical features of these cases: (a) the movements are goal-directed; (b) the goals cannot be immediately achieved, (c) one directly engages in these activities (i.e., one does not make these movements by doing something else that results in them), and (d) the movements are ‘productive,’ that is, they make a difference to the world (both in their unfolding—the rising of the arm as I am raising it—and when they reach their completion—the arm being raised once I have raised it). Let’s briefly consider these features in turn. What do they tell us about agency? Should they be considered paradigmatic of it? If not, why not? 3

Luca Ferrero

3.2 Goal-directedness A goal-directed movement has a certain ‘direction’ or ‘orientation’ toward the goal in question, which means that only some of the ways in which the agent might behave would make her attain the goal or at least move her closer to attaining it. When the agent is acting, she is behaving in ways that take her closer to reaching her goal (or at least that she believes would take her closer to it). Additionally, she is ready to attempt some compensations or adjustments if she encounters obstacles, interferences, or set-backs in her progression toward her goal. The pursuit of a goal seems to require some minimal persistence: the agent does not necessarily and immediately give up the pursuit in the face of any interference or perturbation. When she encounters some obstacles, she is normally expected to try to get around them. Relatedly, once the agent has attained the goal, she is supposed to stop the movement that was conducive to it. The insistence on the goal-directed or purposive character of paradigmatic examples of agency seems correct. If there is an issue here, it is with the proper characterization of the expected adjustments and compensations, since we might want to exclude from purposiveness the simple behavior of physical systems that are naturally tending toward some equilibrium (and thus might seem to ‘persist’ in their behavior as if they were genuinely aiming toward that condition). This raises the question of whether goal-directedness could be characterized exclusively in ‘extrovert’ terms, that is, just in terms of patterns of observable external behavior, or it should appeal to some internal representation of the goal. Either way, some kind of purposiveness seems indeed a necessary and distinctive feature of agency.

3.3 Agency as a productive power Let’s now consider the suggestion that agency is paradigmatically productive, that is, that it is a matter of interventions that make a difference in the world. This suggestion has some intuitive appeal but it seems to exclude witting or intentional omissions and refraining—an exclusion that might seem unwarranted. (See ‘Negative agency,’ Chapter 4.) Regardless of how this question is ultimately resolved, notice that even in standard exercises of agency of the productive sort, there might often be moments of ‘negative’ agency, where an agent takes advantage, at least temporarily, of already favorable conditions. That is, an agent might just let the events unfold naturally without any intervention for as long as this unfolding is getting her closer to her goal. One might just go with the flow, roll downhill, or let the wind blow in one’s sails. At these times, agency is not exercised by antagonistic and effortful interventions (including bodily movements) but simply by monitoring and supervising the unfolding of the events, while being able and ready to intervene in case the natural course of events might need to be nudged or corrected. So it seems plausible that agential guidance and control need not necessarily (and maybe not even paradigmatically) take the form of continuous, sustained, antagonistic, and productive interventions. At times, all that one must do to exercise one’s agency is to coast or drift. Relatedly, one should be careful about picturing the exercise of agency as being always or paradigmatically about ‘initiating’ some motion or change, as if acting were modeled on such scenarios as getting our inert and slouching bodies off the couch. By contrast, many (if not most) exercises of agency seem to take place against the background of the directionless restlessness of the world (a restlessness that can also take place inside our bodies). The basic job of agency might thus be better characterized as giving some direction and shape to these underlying motions—something that can be done in a variety of forms, from pugnacious antagonistic interventions to hands-off supervision. 4

Introduction to philosophy of agency

In addition, an agent might often be aiming at the maintenance of some state or condition, including one’s own health and existence, rather than at the creation of some novel items or conditions. In such a case, the agent’s busyness with antagonistic interventions is not constitutive of agency as such but it is only forced on the agent by the need to counter the (possibly constant) perturbations to her desired stasis or equilibrium.

3.4 Telic and atelic goals The distinction between production and maintenance also relates to the paradigmatic structure of the agent’s goals. A tempting picture of standard action is one in which we reach for or move toward some object, condition, or state of affairs that is at some spatial and temporal distance from us—so that there is necessarily an interval in which one is pursuing a distal goal which has not yet been achieved. In these cases, the action concludes only when the goal has been reached. However, there are many pursuits in which the end is achieved at the same time as one successfully engages in them. For instance, when one is leisurely strolling— walking ‘aimlessly’—one succeeds at strolling at the very same time as one is taking the appropriate steps. As one strolls one has thereby strolled. And this success does not necessarily call for the termination of that activity, though it might propel its continuation. Some maintenance activities might have a similar temporal character. The agent might make repeated and possibly continuous adjustments that make her succeed at maintaining a given condition right at the time when she makes the adjustment. But her stasis or equilibrium might be immediately and continually perturbed, hence calling for continuous and potentially indefinite maintenance (the activity of keeping oneself biological alive or healthy, for instance, might have this temporal profile, which is very different from the profile of raising one’s arm or crossing the street). (See ‘Agency, events, and processes’ Chapter 3, and ‘Diachronic agency’ Chapter 30.)

3.5 Agency and embodiment Let’s now consider the role of the body. Does all agency necessarily or paradigmatically involve the body? As physical beings, it is unsurprising that much of what we do involves our bodies. But this should not entail that there couldn’t be any mental agency. (See ‘Mental agency,’ Chapter 16.) A lot of our mental life, after all, appears to be active in recognizable ways: there are times where we form beliefs and intentions, engage in reasoning, direct our attention to certain matters, and so forth. These mental episodes seem to be correctly described as intentional, deliberate, or voluntary, and we might be held accountable and responsible for their occurrences and consequences (although it is still an open question whether there is a distinctive kind of epistemic agency, see ‘Epistemic agency,’ Chapter 20.) In addition, an important part of our mental and affective life is a matter of our being (defeasibly) responsive to reasons. It could be plausibly argued that this responsiveness is also a manifestation of agency (maybe even the primary manifestation of our distinctive kind of agency). If that is correct, the ‘at will’ character of voluntary bodily movements might offer a misleading model of reason-responsiveness, with its insistence on discretionary choice rather than on the conformity of our conduct to reasons. (See ‘Rational agency,’ Chapter 10, and ‘Aesthetic agency,’ Chapter 42.) Regardless of where one stands on the issue of mental agency, there are still questions about the role of the body in paradigmatic modes of agency. In particular, one might wonder whether agency might not be directly realized and exercised in other physical but 5

Luca Ferrero

non-biological forms, such as various technologies, and what implications this might have for the individuation of the realizations of agency, including its social and political dimensions. (See ‘Material agency,’ Chapter 19.)

3.6 Agency and intentionality The last element of the allegedly paradigmatic model of action is the ‘intentional’ qualification of productive bodily movements. The term ‘intentional’ can be read in at least two ways. According to a weak reading, it just stands for a generic kind of purposiveness or goal-directedness (possibly the one that requires at least a mental representation of the goal). As such, it can be applied to a variety of agents, including nonhuman animals. According to a stronger reading, it refers to the much more demanding form of agency that is characteristic of human beings—a form which includes such things as the understanding of goals as goals, a distinctive kind of self-knowledge about one’s goals and actions, planning abilities, and a willingness to offer and ask for reasons for action (for more on this, see the discussion of fullblooded agency in this introduction below). Both uses of ‘intentional’ and cognate expressions are legitimate but, when using this term in the philosophical investigation of agency, one needs to make clear whether one is trying to characterize a notion of agency that has broader application across a variety of entities, including various organisms and possibly some artifacts, or one is primarily focused on the distinctive properties of human agency.

4 Kinds of agency 4.1 The varieties of agency Unless one is skeptical about the existence of any genuine agency at all, there is a general philosophical agreement that normally developed human beings possess agency. But what about other animals, plants, and unicellular organisms? What about artifacts such as robots, or supernatural beings such as angels, demons, and gods? There are also important philosophical questions about the presence of genuine (and non-derivative) agency in sub-personal systems or functional components (say, the digestive system, the heart, a single cell in a multicellular organism) and in super-personal entities, such as social groups and institutions. Part of what is at stake in addressing these questions is whether there are different levels or degrees of agency. It might be that some dimensions of agency could be attributed only to more complex organisms but not to simpler ones. If so, what are these dimensions? How are they related to each other? What are the normative implications of these attributions? (See Part 2, Chapters 7–10.) Even when there is some agreement on the different levels or degrees of agency, there is still one additional methodological issue: what is the relation between the various kinds of agency? Are they set along an additive sequence, so that one can go from one level to another by simply adding some powers or capacities to the simpler forms (thereby retaining the basic operations of the lower/simpler levels)? Or are the various kinds different species of the same genus, so that possessing some different capacities has a ‘transformative’ rather than a merely ‘additive’ effect on the various capacities and powers? (See ‘Rational agency,’ Chapter 10.) Relatedly, there is a question whether some of the more complex forms of agency might be taken to be better along some evaluative dimension, and whether agents might be under 6

Introduction to philosophy of agency

a demand to acquire and exercise better forms of agency, when they have the option to do so. One might also wonder whether there are any normative pressures to relate differently toward simpler forms of agency. Concerns of this sort become particularly poignant in the case of human agency. Human agency exhibits various dimensions and layers of complexity, which are acquired as we grow up, and that we might lose or fail to develop and cultivate. How should we both conceptualize and relate to those instances of human agency that do not exhibit the standard combination of powers and capacities? (See Part 2, Chapters 7–10; ‘Agency and disability,’ Chapter 14; and ‘Pathologies of agency,’ Chapter 15.)

4.2 Full-blooded agency Much of the philosophical reflection on agency is centered on human agency, which is often referred to as ‘full-blooded agency.’ This is also the term that I will use throughout this handbook to indicate that this agency exhibits a distinctive combination of features. In addition to purposiveness, full-blooded human agency exhibits the following features: 1

2

3

4

There is structural complexity both in our pursuits and in our ends. Much of this complexity stems from the extended temporal character of our agency in the mode of planning—we have psychological, conceptual, and reasoning capacities to engage in integrated long-term plans, often directed at ends that we could not pursue and conceive in the absence of these very capacities (see ‘Planning agency,’ Chapter 31, and ‘Agency, time, and rationality,’ Chapter 32). We occupy a privileged epistemic position with respect to our own intentional conduct: we have a distinctive kind of self-knowledge about what we are doing and why we are doing it (a kind of self-knowledge that might extend to or explain other forms of self-knowledge). (See ‘Agency and practical knowledge,’ Chapter 21, and ‘Agency and self-knowledge,’ Chapter 23.) In explaining, assessing, and interacting with the intentional conduct of fellow human beings, we expect them to have reasons that justify and make sense of their conduct (reasons that are in principle expected to be explicitly offered by the agent when one is asked for them). In this way, we make explicit what a constitutive feature of intentional conduct— being the object of a distinctive kind of explanation: explanation in terms of reasons, rationality, and intelligibility. The offering and the asking for reasons, moreover, is not an idle kind of external appraisal. We have a basic disposition to modify and adjust our behavior in light of our appreciation of the reasons that purport to justify it. Hence, we operate with the expectation that our conduct conforms to the reasons that purportedly justify it. The same disposition and expectation operate in response to the explicit requests of reasons. To the extent that our conduct does not conform to these reasons, we are expected to modify it (if still possible) or to be held accountable for the failure to do so. (See ‘Agency, reasons and rationality,’ Chapter 36, and ‘Agency and practical reasoning,’ Chapter 37.) The accountability to rational justification is at the basis of what might be called ‘ethical responsibility.’ We hold each other responsible for our intentional conduct, its consequences, and its results (something that takes place in various settings, including interpersonal relationships, institutional and legal contexts, and morality). It is because of its intimate relation to the practice of holding agents responsible that full-blooded agency is the natural home for the application of such qualifications and concepts as voluntariness, willingness, duress, compulsion, coercion, reluctance, consent, innocence, and guilt. (See ‘Agency and responsibility,’ Chapter 25.) 7

Luca Ferrero

5

6 7

8

9

Full-blooded agency is the ground of the possibility of attributing normative or deontic statuses: only full-blooded agents can be regarded as genuine and full bearers of duties, obligations, rights, and responsibilities. Full-blooded agency is thus a precondition for both legal and moral personhood. (See ‘Agency and personal identity,’ Chapter 34.) Traditional questions about freedom and free will seem to be appropriate only once fullblooded agency is on the scene. (See ‘Agency, will, and freedom,’ Chapter 24.) Questions about identification, self-governance, authenticity, and autonomy only arise in the presence of full-blooded agency. (See ‘Agency and identification,’ Chapter 26, and ‘Agency and autonomy,’ Chapter 27.) Full-blooded agency comes in a distinctive social form. This sociality is not limited to the capacity to engage in social activities, but it is also a constitutive aspect of the practices of explanation, justification, and accountability. Finally, all these dimensions, features, and implications of full-blooded agency either go together with, or are partly made possible, constituted, or enhanced by our critical and reflective capacities. Minimally, we have the linguistic, conceptual, and reasoning resources to articulate our goals and the reasons in their support, and to understand the very notion of goals and reasons. This is our basic critical capacity: we can represent and understand the basic rational structure of our intentional agency and its normative implications. An important consequence of this critical capacity is that rational scrutiny and justification are not limited to our instrumental conduct but also extend to our ends. Higher-order reflective abilities can then be turned toward all features and implications of our full-blooded agency, in its psychological, conceptual, rational, normative, and moral dimensions—as evinced by the very existence of the philosophy of agency. (See Part 8, ‘Agency, reasoning, and normativity,’ Chapters 36–42.)

Much of the work in the philosophy of agency is centered on full-blooded agency, given that we are supposed to be agents of this kind. Unsurprisingly, many of the chapters in this volume have the same focus. The appearance of the word ‘agency’ in them (including in some of their titles) might often be better interpreted as ‘full-blooded agency,’ which is something that should be apparent to the attentive reader.

5 Four pictures of agency This volume does not try to advocate for a specific positive account of agency. Its aim is rather to introduce the reader to several central issues and some different approaches in the philosophy of agency. The reader is invited to explore various aspects of agency, their connections, and their implications. There is no privileged place of entry or re-entry, although several chapters are loosely organized around some larger set of questions, as described in the introduction to the sections of this handbook. To provide some initial direction, however, I’d like to offer four general pictures of agency that can seem to inform, often implicitly, various specific philosophical views about agency. By ‘picture’ I mean an approach to the philosophy of agency that tends to provide some basic orientation in selecting the questions that are considered worth asking and in setting the desiderata and constraints on their answers, including judgments about salience and irrelevance, the style of argumentations, and the examples that tend to be used (including the choice of paradigmatic illustrations of agency). This orientation might not be self-conscious or explicit, and it might originate in a pre-theoretical view on the nature of agency, which 8

Introduction to philosophy of agency

might then be elevated into a more systematic philosophical account. A picture is not the same as a full account or theory that aspires to consistency and coherence. Rather, a picture tends to motivate certain lines of inquiry and to frame the questions to be raised. It might not exclude other approaches. One might find strands of different pictures within more elaborate philosophical accounts of agency, some of which might actually and explicitly try to reconcile different pictures. This should not be surprising. After all, each of the pictures that I am about to present appears to highlight some intuitively plausible aspects or dimensions of agency. Thus, as we start to reflect on the nature of agency, we should try to become aware of the role that any of these pictures might implicitly play in framing our investigation. And we should not be too cavalier in dismissing (or privileging) any of them. Given my understanding of what a ‘picture’ of agency is, it should not be surprising that what I am about to offer are only some impressionistic sketches, where I hint at some variations on what I take the distinctive theme of each of these pictures (which also means that it is unlikely that one could find a philosopher that fits neatly into a single picture and that exhibits all of the traits that I am about to present). Notice that I am not claiming that these are the only four possible pictures of agency, but this is my best attempt to begin to articulate some of the approaches that can be encountered in the philosophy of agency literature.

5.1 Agency as creation According to this picture, agency is the capacity to create or produce, to bring about something new, such as the initiation of an action. A down-to-earth example of this creative power is the ‘at will’ raising of one’s arm from a position of rest, which is prompted by nothing other than one’s choice or decision to do so. A similar example is the case of the arbitrary selection between two or more open paths, especially in the case in which they appear to be equally desirable (think about Buridan’s ass scenario, for instance). This picture emphasizes the role of the agent as the source or origin of action, where the action is added as something new to the world. This is why I call this approach ‘agency as creation’ (rather than agency as the mere power of initiation or selection). I also suspect that, for some proponents of this picture, the ideal or model of agential power might be something like a divine ‘fiat’—a divine ex-nihilo creative act. This is not to say that raising one’s arm or selecting from among open paths is without constraints. But within those restrictions, for this picture, agency operates unfettered, hence its ‘discretionary,’ ‘at will,’ or ‘arbitrary’ character. According to this picture, exercises of agency are ultimately manifestations of a radical kind of freedom (possibly as radical as the ‘liberty of indifference,’ which is unfettered even by the constraints imposed by responsiveness to reasons and thus open in principle to the possibility of perverse action). The insistence in this picture on the (locally) unconstrained creative or productive powers of the agent might ultimately reflect and expose a deep concern with our possible ‘captivity,’ with the worry that we might ultimately lack freedom. Relatedly, in this picture, the attribution of some conduct to the agent is primarily a matter of tracing the source of that conduct back to the agent as the place where one cannot go any further. The investigation of the internal structure of the agent appears to be a less urgent task, possibly because of the expectation that this investigation could not reveal much. This creative power might thus be a distinctive but irreducible feature of agency. This irreducibility, however, raises the worry that this picture might be difficult to reconcile with any naturalistic view of the world—especially if one thinks of the basic exercises of agency as ‘inserting’ themselves ex nihilo, so to say, into the natural causal order. 9

Luca Ferrero

5.2 Agency as self-constitution According to the second picture, agency is ultimately a matter of self-constitution or self-maintenance. In this picture, the paradigmatic example of the exercise of agency is the conduct of an organism, which is ultimately directed at the organism’s self-maintenance, that is, at securing its continuous survival in response to the ultimate existential threat: that of dissolution and death. For this picture, agency is ultimately the same as the capacity of life: agents are first of all organisms (self-constituting and self-maintaining entities), and different kinds of agency reflect different kinds of life-form. The sense of life (and organism) in this picture need not be restricted to the ‘biological,’ to material organisms with a metabolism. In principle, it seems possible to extend the idea of self-constitution and self-maintenance to rational life, to the life of a rational subject as a rational subject, where the existential threat arises within the rational order rather than within the causal one. Inconsistency and incoherence might be to rational life what material disintegration is to physical life. In both cases, the threat is to the maintenance of the kind of integrity and unity required to sustain continuous existence either as a rational or a biological agent (or both). (Notice that this description of rational life concerns the characterization of the kind of unity and existence at stake in the power of agency—it is a separate issue whether a rational life-form could exist in a disembodied form or as embodied in a non-biological form, say in a material artifact.) This picture of agency gives expression to the idea of agency as self-motion, as self-originating conduct with a necessary teleological orientation. Within this picture, there is no question about the source of the agent’s basic busyness, since self-maintenance comes with the built-in preoccupation to fend off any potential fatal threats—which the agent might constantly encounter in its path to survival. Unlike agency-as-creation, the basic concern here is not captivity, but mortality. Agency-as-self-constitution offers a straightforward account of the attributability of the action to the agent. As an exercise of self-motion, any piece of conduct is necessarily of the ‘self,’ that is, of the entity that is, by its nature, in the business of constituting itself—of making its ‘self.’ This invites an investigation in the conditions that make possible for the entity to have both adequate internal integrity and sufficient separation from the external world to constitute a distinct and viable unit of agency. Likewise, in this account, there is a straightforward path to the attribution of agential conduct to the agent ‘as a whole.’ Agency is to be primarily attributed to the self-maintaining entity rather than to any of its functional components, such as an organ or a subsystem. The operations of these components might have their own teleological orientation, but if these components are not self-sustaining in relative isolation from the larger entity, they might be said to have agency only in a derivative sense, if at all. Likewise, larger aggregates of self-maintaining agents would not exhibit genuine agency unless they reach a level of integration such that, at that level of organization, they would count as self-maintaining units in their own right. (This is not to say that self-maintenance is maintenance in isolation from the external world or from other agents. The picture of agency-as-self-maintenance can actually be well-positioned to account for the mutual interdependence of agents, including various forms and levels of sociality and the nesting of different levels of agents within each other.) In agency-as-creation, the creative power is primarily manifested in each individual exercise of agency, as directed at bringing into existence the particular object of that individual act of creation. In agency-as-self-constitution, agency still has a creative aspect, but what is ultimately (and constantly) brought into existence is the agent itself. The agency of self-constitution is the agency of continuous self-(re)creation. 10

Introduction to philosophy of agency

This does not mean that self-constituting agents must explicitly represent their own maintenance as the object of any of their particular activities. Self-maintenance might often be performed by the explicit pursuit of instrumental means, without necessarily representing or appreciating them as means to self-maintenance. An organism might never seek self-maintenance as the direct object of any of its goal-directed pursuits, whose proximal object is always some specific (and usually telic) way of contributing to its survival, such as getting food, avoiding predators, and seeking shelter. It is only for a reflective self-maintaining agent, who becomes cognizant of the fundamental structure of their agency, that (atelic) self-maintenance might become the primary and explicit object of concern (for the telic/ atelic distinction see above and ‘Diachronic agency,’ Chapter 30). Here are some of the attractive features of this picture. First, it is easy for it to emphasize the temporal dimension of agency. This is because self-constitution is a continuous process, which is terminated only by the agent’s death. The continuity of self-constitution should not be confused, however, with a relentless process of positive intervention in the world. There is nothing in the idea of a self-maintaining entity that stands opposed to taking advantage of already favorable conditions by coasting and drifting. If anything, the negative mode of agency—letting things unfold naturally—could be seen as the preferred mode of self-maintenance, if only one were not under constant threats. After all, for this picture of agency, positive interventions are ultimately only in response to present or anticipated threats. It is only the indifference or the hostility of the world that might force the selfmaintaining agent into a relentless series of mostly positive interventions. In agencyas-self-constitution, there might be no special value attached to spontaneous or ‘unforced’ individual acts of positive creation or production as such. Another interesting feature of this picture is that it might offer the basis for an account of the connection between agency and normativity—by leveraging either the built-in aim or function of self-maintaining entities or the force of the existential threat that always hangs over the heads of self-constituting entities. (See ‘Agency and normativity,’ Chapter 38, and ‘The aim of agency,’ Chapter 39.) Finally, this picture might make it easier to account for the natural history of agency and the existence of different kinds of agents, since this history and variety might just reflect the unfolding of the tree of life on Earth. But for this same reason, this picture might run into problems when trying to account for the rational aspect of agency, especially in the form of self-conscious reason-responsiveness. Although agency-as-self-constitution might explicitly contemplate ‘rational’ life-forms, there is a legitimate worry about whether the notions of self-maintenance and self-constitution can really apply to rational agency, especially of the critical and reflective kind. Agency as self-constitution seems especially apt at accounting for the agential character of the operations of simpler forms of life. The concern is that it might become much less plausible when applied to the distinctive features of full-blooded agency. In a related vein, one might wonder whether an excessive reliance on a picture of agency inspired by the structure of biological life might rule out the possibility of genuine agency in the absence of either metabolism (say, in robots or androids) or of any existential threats (say, in immortal beings).

5.3 Agency as psychological causality According to a third picture, agency is fundamentally the psychological capacity to bring about bodily movements that intelligibly fit with the agent’s desires, cares, concerns, or commitments. A straightforward illustration of this picture is found in what is known as 11

Luca Ferrero

the ‘standard story about action’ (see Introduction to Section I). According to this story, an action is a bodily movement caused (in the right kind of way) and rationalized (that is, made intelligible) by the agent’s desire for a certain end and her belief that moving her body in that particular way will bring about that end. There is something intuitively appealing about the standard story, since it seems to conform to ordinary folk-psychological explanations of action, in terms of the so-called belief/ desire psychology. For instance, one explains the action of reaching for a glass of water out of the combination of one’s desire to quench one’s thirst and one’s belief that one can satisfy that desire by extending one’s arm in the direction of the glass of water. This arm movement ‘makes sense’—it is made intelligible or ‘rationalized’—by that particular combination of mental states. (By contrast, it would not make sense to reach for the glass of water if, say, one believed that the glass was outside of one’s reach or one had no desire for the water.) In this account, the elements of one’s psychology play a double explanatory role: they make the action intelligible by providing reasons for its performance; and they bring about the action by causing it (in the right way—this qualification is always implicit because of the so-called problem of ‘deviant causal chains,’ on which see ‘Agency and causation,’ Chapter 1, and ‘Agency, functions, and teleology,’ Chapter 2). The standard story is not the only version of what I am calling agency-as-psychologicalcausation. For instance, one could offer a much more elaborate characterization of the psychological structure (or ‘psychic economy,’ as this is sometimes called) that is explanatory of the action, including the existence of different kinds of mental states (including intentions, plan states, and higher-order mental states) and more complex ways in which they might relate to each other. (See ‘Intentional agency,’ Chapter 9, ‘Planning agency,’ Chapter 31, and ‘Agency and identification,’ Chapter 26.) What makes these various accounts versions of the psychological-causation picture is their focus on an account of action in terms of its psychological (as rationalizing) causal antecedents. Unlike the pictures of agency-as-creation and agency-as-self-constitution, this picture makes the presence of a mind—of a psychic economy—central to the nature of agency; it does so by articulating the internal structure of the psychological structure and its contribution to bringing about genuine exercises of agency. As I remarked earlier, this articulation appears to be a problem for agency-as-creation, especially if that picture insists on the sui generis character of the agential powers. The agency-as-creation picture is at risk of locating agency outside of the natural causal order. A commitment to a naturalistic explanation appears to be a major motivation behind the agency-as-psychological-causation model, hence its insistence that both the internal operation and the external outputs of the psychology be accounted for in terms of the generic bond of ordinary causation. Psychological causation is ordinary causation by elements of one’s psychology, not some kind of supernatural, mysterious, or spooky power. Agency-as-creation is not necessarily guilty of invoking some kind of magic, but it can lend itself to such invocation. For, unlike agency-as-psychological-causation, it does not start from the very idea that agency is a capacity to be explained in terms of the naturalistic operation of some underlying causal structure. Notice that this is not a concern with agency-as-self-constitution. A self-maintaining entity is one with an internal structure and organization that is in principle naturalistically explainable. What differentiates agency-as-psychological-causation from agency-as-self-constitution is that the former starts from a psychological structure. This appears to be an advantage when one is to account for the distinctive features and implications of full-blooded agency, which is imbued with psychological and rational 12

Introduction to philosophy of agency

attributes. The psychological and rational dimensions of agency, instead, are only a secondary feature of agency-as-self-constitution, given that self-maintenance is, first of all, an ontological rather than a psychological or rational property. By starting with life rather than with mind, agency-as-self-constitution might end up being too generous in the attribution of agency to simpler kinds of organisms while struggling to account for the distinctive rational dimension of full-blooded agency at the other end of the spectrum. A related advantage of agency-as-psychological-causation is that, in principle, it faces no difficulty accommodating the possibility of genuine agency in entities that are not self-maintaining—such as artificial, non-biological machines—as long as they are realizations of a sufficiently complex psychic economy. Agency-as-self-constitution, however, might fare better than agency-as-psychologicalcausation when we come to the question of how to attribute actions to their agents. As we have seen above, in agency-as-self-constitution, the self-originating character of the exercises of agency is guaranteed by the very notion of self-constitution. The psychological causation picture, instead, might struggle on this front. According to psychological causation, where exactly is the capacity of agency located? Who or what is exercising this capacity? ‘The psychic economy!’—one might say. But what does make this economy that of a genuine agent—in the sense of a sufficiently integrated unit to count as the proper locus of the exercise of agency? The concern is that the elements of the psychic economy might be doing the work by themselves, without any discernible role for the agent (this is known as the problem of the ‘disappearing agent’). The risk is that the psychological-causation picture might either explain the agent away or surreptitiously and uninformatively assume some homuncular unity within the internal working of the psychic economy. True, the psychological causation story is framed in terms of the operation of a system—a psychic ‘economy’—but there seems to be nothing in the basic idea of psychological causation that, by itself, explains how to individuate a system with the kind of unity and integration that we expect of the idea of a genuine agent. In addition to the worry of the disappearing agent, many objections have been raised against the standard story of action. These objections target its metaphysics of mind and of causation, its seeming inability to account for extended guidance rather than just the initiation of actions, and its troubles with the putative agential character of omissions and refraining. (See the introduction to Part I, ‘The metaphysics of agency,’ and Chapters 1–6.) However, the standard story is only one possible version, even if an intuitive one, of agency-as-psychological-causation. Its problems need not necessarily invalidate the picture in all of its possible forms. After all, the basic idea of psychological causation seems most welcoming to different characterizations of the psychic economy, including a variety of mental architectures, with very different components and structures (including different metaphysical accounts of causation and the appeal to psychological elements belonging to different metaphysical categories—such as states, events, and processes). In my view, what makes these different accounts instances of the psychologicalcausation picture is not a commitment to the specific features of the standard story but the aspiration to provide a naturalistic account of how a psychology can operate causally and yet stay within the rational order, so as to give rise to full-blooded agency. This is what makes agency-as-psychological-causation especially attractive but also potentially most vulnerable. For there is always the worry that the causality invoked by this picture to explain the operations of the psychic economy might not be able to explain the truly distinctive aspects of full-blooded agency, including its normative character and its self-conscious and reflective dimensions. 13

Luca Ferrero

5.4 Agency as reason responsiveness According to this last general picture of agency, agency is primarily the capacity to respond to reasons. Agency is first of all exercised in making up our minds on the basis of normative or rational considerations about how we ought to make up our minds. In the paradigmatic cases, in making up our minds about what to do, we thereby also give shape to our conduct in the material world. Our agency, however, is not primarily exhibited in the ‘at will’ shaping of our physical actions but in our avowals or disavowals of reason-responsive attitudes. It is useful to compare this picture to agency-as-creation. In agency-as-creation, agency is ultimately a matter of making a difference within the causal fabric of the world. This kind of difference can be modeled in terms of changes in the physical world, such as selecting which path to take at a junction or setting into motion an inert body. In agency-as-reasonresponsiveness, instead, the change is first of all within a normative rather than a physical space. In agency-as-reason-responsiveness, the agent takes on a new shape because of the new status acquired by some of one’s own attitudes (paradigmatically, one’s judgments) in response to one’s sensitivity to normative considerations. Crucial to both pictures is the idea of the agent as the direct source of one’s own agential conduct. But the two pictures drastically differ in the character of the proposed source. This difference can be illustrated by the distinction between ‘authorship’ and ‘authority.’ According to agency-as-creation, the agent is the source of agential conduct. It is so because the agent is the author—the creator or originator of this conduct. This authorship, which retains a voluntary character, is ultimately accounted for in causal terms. By contrast, in agency-as-rational-responsiveness, the agent is a source in the sense of being the authority that endorses, judges, or avows something. Being the agent is, first of all, a matter of putting a stamp of approval (hence the change in normative status within the space of reasons) rather than of directing one’s conduct into some physical direction instead of another (even if one’s conduct is by default expected to take on a specific direction in physical space as a result of the change in one’s normative status). The two pictures also differ in their interpretation of what they fear the most. Both are worried about the loss of freedom, especially in the form of being pushed around by mere causality. But agency-as-creation might respond to this concern by encouraging a radical resistance to any form of ‘constraining,’ even the one induced by responsiveness to reasons. Thus, agency-as-creation might end up encouraging an extreme view of agency as the defiance of reason, as epitomized by choices done in an arbitrary matter or in the absolute refusal to be shaped by any rational consideration. The defiance of reason, instead, is contrary to the very spirit of agency-as-reasonresponsiveness. The kind of freedom that the latter view wants to affirm is that of autonomy— of self-determination in response to one’s appreciation of the reasons for action, not that of arbitrary indifference (if not even defiance) of one’s assessment of which reasons bear on one’s practical circumstances. What counts in favor of the agency-as-reason-responsiveness? First, this picture accounts for the seeming agential and active character of aspects of our lives that go beyond physical actions, including the vicissitudes not only of our cognitive attitudes but also of some of the conative and affective ones. This is shown by our practices of explaining, appraising, and holding agents responsible on account of the reason-responsive (rather than ‘at will’) formation, retention, and revision of these attitudes. Second, the view does not have problems with the attribution of a piece of conduct to its own agent. This is because, at least in the paradigmatic cases, the subject’s judgment that so-and-so is to be done constitutes the fact that 14

Introduction to philosophy of agency

the doing of so-and-so is the doing of that subject. Third, this picture has no trouble accounting for the relation between agency, reasons, rationality, and, more broadly, the normative domain. This is because the connection between agency and normativity is built right into this picture of agency. These same features, however, might also constitute the major vulnerability of agencyas-reason-responsiveness. First, by being centered on reason-responsiveness, this picture might be unable to account for the simpler kinds of agency. Second, it is unclear how this picture fits with a naturalistic outlook. It is not that this picture is in principle anti-naturalistic, but, unlike agency-as-psychological-causation, it does not seem to be bothered with securing foundations with impeccable naturalistic credentials (which is unsurprising since it primarily trades in the abstract space of the normative).

5.5 Conclusion This concludes my initial and very rough sketch of four basic pictures of agency that might, at least implicitly, motivate and shape many philosophical investigations of agency. As I remarked at the outset, these are general pictures that might be neither fully explicit nor articulated by those who are under their influence. And it would not be surprising if many philosophers are attracted by more than one picture at the same time and possibly try to reconcile them, given that each of these pictures seems to highlight some distinctive and attractive features of our full-blooded agency. However, it is too early to say whether any such reconciliation is possible or whether one of these pictures (or possibly even a different one altogether) will eventually emerge as the correct one.

6 What’s next? 6.1 ‘Philosophy of agency’ The questions and pictures that I have presented in this introduction are a sample of the philosophical issues raised by the notion of ‘agency.’ Many additional topics will be presented in the following chapters. Taken all together, they should give the reader a good (although far from comprehensive) overview of the ‘Philosophy of Agency.’ Notice that this label, although it is gaining prominence, has not yet replaced ‘Philosophy of Action’ as the standard name for this area of philosophy (whereas the old-fashioned and far too restrictive label, ‘action theory,’ seems to be, fortunately, almost forgotten). Whenever it is in my power to do so (such as editing this handbook), I will insist on the use of ‘philosophy of agency.’ This is, I believe, for very good philosophical reasons. The primary topic of investigation is not ‘actions,’ which are just among the possible exercises of the capacity of agency. True, actions (together with acts and activities) appear to be the primary manifestations of agency, and, as such, they might correctly get the bulk of the attention. But this does not justify making actions, rather than the capacity that underlies them, the primary object of philosophical reflection. And even if one were to argue that agency can only be exercised in actions, this would still be a substantive view within the philosophy of agency, not a neutral definition of an area of philosophical inquiry. Even those who might profess neutrality on this matter should be wary of the surreptitious effects of the use of the label ‘philosophy of action.’ This name might favor a potentially tendentious philosophical agenda, which could give undue prominence to actions over other manifestations of agency. Just to give an example about some hidden dangers: by starting the 15

Luca Ferrero

investigation with talk of ‘actions’ one might privilege certain phenomena in their nominalized form rather than in a verbal one, including variations in tense and aspect. But who is to say that the focus should be primarily on actions rather than on acting, say? (Likewise, why should one privilege deeds over doings? Intentions over intending? Plans over planning?) In addition, the primary focus on actions might implicitly suggest that they could stand as separate items, which could be studied piecemeal, in relative isolation from each other and from their sources, and used as building blocks in the reconstruction of extended exercises of agency. Last but not least, the primary focus on the exercises of a capacity also risks obscuring the role played by agents, as the loci of this capacity. The philosophy of agency should be, at the very least, a philosophy of agency, agents, and actions. These are not idle worries, since important and foundational philosophical matters might lie behind these seemingly innocent linguistic choices, especially in the case of the philosophy of agency. (In further support for my terminological choice, consider how odd—and philosophical prejudicial—it would sound to refer to other philosophical areas by the analog of ‘philosophy of action’: for example, ‘philosophy of belief ’ or ‘belief theory’ instead of epistemology; ‘philosophy of mental states,’ or ‘propositional attitudes theory’ instead of ‘philosophy of mind;’ ‘philosophy of moral conduct’ instead of ‘ethics’ or ‘moral philosophy,’ etc.)

6.2 How to read this book This handbook is not going to propose any positive characterization of the nature of agency but rather to represent some of the variety of issues and viewpoints to be found in the philosophy of agency (not to mention that many of the central questions in this area revolve quite directly with addressing the very question of the nature of agency). This is also why there is no specific chapter in this volume titled ‘The Nature of Agency.’ The reader is rather invited to try to discover and reflect on the various questions by sampling different topics in this handbook, with no specific point of entry or progression. At the end of each chapter, the reader will find some additional suggestions for further related topics, which one might follow up at their discretion. Some chapters are organized around a set of common questions, and the reader is invited to read the introduction to these sections to learn more about how the topics might be related. This organization, however, is not a rigid one. Many chapters could have as easily been assigned to different sections, and different groupings of questions might have been proposed altogether. The volume aims to be comprehensive but it is by no means exhaustive. Many topics are not covered here. The absence of any particular topic is hardly an indication of its lesser philosophical importance. Many are missing simply because this volume is the product of the coordination of the limited and constrained agency of the editor, the authors, and other potential contributors. (Consider, for instance, some of the topics that went through some advanced stages of planning but had to be dropped at some point: ‘Agency, desire, and motivation,’ ‘Agency and the first person,’ ‘The logic of agency,’ ‘The sense of agency,’ ‘Agency and the virtues,’ ‘Agency and perception,’ just to name a few.) In addition, there are for sure many unwitting omissions that just reflect the ignorance and imperfection of this editor. But there are a couple of noticeable intentional omissions that require some justification. First, except for the brief discussion in ‘Agency, will, and freedom,’ Chapter 24 (and indirectly in ‘Agency and autonomy,’ Chapter 27, and ‘Agency and responsibility,’ Chapter 25), there are no chapters that deal directly with the question of freedom and free will, despite the centrality of these topics to the philosophical reflection on agency. This ultimately reflects a major division of philosophical labor that is now well-established within contemporary 16

Introduction to philosophy of agency

Western Anglophone philosophy. There is an entire philosophical area specifically devoted to questions of freedom, free will, and responsibility, and several introductory and more advanced resources are already available (see, for instance, Griffith, M., Levy, N., & Timpe, K. (eds.), 2017, The Routledge Companion to Free Will.). I have thus decided that there was no point in duplicating in this volume what is already available elsewhere, and I invite the readers specifically interested in issues about freedom and free will to consult these other resources. Likewise, there is no chapter devoted to shared, joint, collective, or social agency. This is not because these are not important aspects of agency. A reasonable claim can be made that agency, especially full-blooded agency, must necessarily come in some kind of social form. Some of these social aspects get explicitly discussed in some chapters (such as ‘Material agency,’ Chapter 19, ‘Planning agency,’ Chapter 31, ‘Agency and the emotions,’ Chapter 29, and ‘Agency and responsibility,’ Chapter 25). But a single introductory chapter on social agency would hardly add any new information which couldn’t already be available in such publications as Jankovic, Marija, and Kirk Ludwig (eds.), 2017, The Routledge Handbook of Collective Intentionality or the relevant entries in the Stanford Encyclopedia of Philosophy (https:// plato.stanford.edu). As for the case of freedom, readers specifically interested in collective or social agency should consult these resources. Many readers of this handbook might also benefit from reading it in conjunction with a systematic introduction to the philosophy of agency, such as the excellent volume by Sarah Paul, (2021) The Philosophy of Action, Routledge (see also Lilian O’Brien, 2015, Philosophy of Action, Palgrave and Rowland Stout, 2005, Action, Routledge). Finally, this handbook does not specifically discuss any of the long history of the philosophy of agency, although many of the topics have been the object of philosophical reflection and discussion for centuries. This is, once again, a major editorial choice forced by practical constraints in putting a volume of this kind together. It does not reflect any negative assessment of the philosophical value of a historical look at this discipline. On a related note, this handbook does not want to deny or hide its historical, geographical, and cultural situatedness. The general philosophical outlook presented in this volume is the one informed by the scholarly discussion in Western Anglophone philosophy from about the 1950s, with some occasional influences from the so-called ‘continental’ tradition and some glances at selected portions of the history of Western philosophy. Much of the philosophical discussion that has taken place within this tradition in the last seven decades or so is done with the aspiration that the questions that are raised and the answers that are offered might, at least in part, transcend the confines of the particular philosophical outlook within which they have been raised and offered. Whether this aspiration succeeds (or is ultimately in vain), however, is not something that could be addressed in this volume. My hope, however, is that much more work will be done to investigate the nature of agency within other philosophical traditions and other disciplines. Although I believe that any reader should feel confident about beginning the study of the philosophy of agency by reading this handbook, the reader should never be under the impression that the philosophy of agency either begins or ends with the contributions collected in this volume.1

Note 1 This handbook is the product of the joint, telic, and extended intentional activity of many agents. Whether or not they want to admit it, they all bear some responsibility for it (with the possible exception of their unwitting omissions). But I suspect that I am the ‘most agent-of ’ this final product and so I will take most of the responsibility for it. The least burdensome and most welcome aspect of this responsibility is my duty (and desire) to thank all who contributed to the final attainment

17

Luca Ferrero by playing very different roles at different places and times. To begin with, I want to thank the editorial staff at Routledge: Rebecca Shillabeer—who first commissioned the handbook—, Gabrielle Coakeley, and Adam Johnson—who has overseen its further development and completion. This book could not have been conceivable without the existence of a thriving community of philosophers, who continue to engage in exciting conversations and reflections on the nature of agency. A fair number of them have directly participated in this volume, so I thank them all, both individually and jointly, for their chapters (and for their patience and promptness in dealing with my slow and desultory work toward the final success and closure of our joint pursuit). I also want to thank the many friends and colleagues who have not written a chapter for the handbook but have provided essential and invaluable support, advice, and above all philosophical inspiration over the years. In particular, John Fischer, David Horst, Richard Moran, Michael Nelson, Alexandra Newton, Sarah Paul, Andrews Reath, Karl Schafer, Tamar Schapiro, Sergio Tenenbaum. I also thank my colleagues and students at the University of Wisconsin-Milwaukee and, more recently, at the University of California-Riverside. A special thank goes to the organizers and participants of the many meetings of the DFG Netwerk ‘Practisches Denken und Gutes Handlen’ and its descendant (aka ‘The Action Network’) who have done so much over the years to keep the philosophical conversation about agency both alive and lively. Last but not least, I have to thank the members of my small extended family—Marilena, Carla, and Susan—who have paid many of the direct costs of this project, by being around me during the ups and downs of the writing and editing process, but are only going to gain very indirect benefits from its eventual success: as much as it might surprise the readers of this handbook, not every full-blooded agent turns out to be also a philosopher of agency—something which, after all, we should all be thankful for.

18

PART 1

The metaphysics of agency Introduction to Part 1 Luca Ferrero

It would not be unreasonable to expect that a handbook in the Philosophy of Agency would open with a chapter titled ‘What is agency?’ or more positively ‘The nature of agency,’ especially if the first part of the book is devoted to nothing less than the metaphysics of agency. Alas, you won’t find this chapter here. As I explained in the general introduction, although I think we have a fairly good intuitive grasp of what the core instances of fullblooded agency look like, it is far from clear whether we could offer some uncontroversial and straightforward statement about what the concept of ‘agency’ stands for. Our best bet, as I suggested in the introduction, is to try to unpack and explore various more specific issues, attacking the question from different angles, in the hope of eventually gaining a better articulation and understanding of agency. So the discussion of the metaphysics of agency will begin with the exploration of four distinct but not unrelated issues: what kind of causation underpins the operation of agency, the role of teleological explanations, the kinds of entities that might be at stake in the temporal exercises of agency, and whether omissions and refraining are genuine instances of agency, even if only of a ‘negative’ kind. The first four chapters exemplify some of the important metaphysical dimensions that need to be explored to make some progress toward a comprehensive account of the nature of agency. They do not purport to be part of a single and unified view, and some of them might push in different and possible incompatible directions. But it is worth remarking that all the authors of these chapters explicitly frame their contributions by criticizing or distancing themselves from what is often known as the ‘standard story of action,’ which in its basic form claims that an action is a bodily movement caused (in the right kind of way) and rationalized (that is, made intelligible) by the agent’s desire for an end and her belief that moving her body in that particular way will bring about what she desires. It is hard to tell whether something like the standard story enjoys the status of being the received view among the current philosophers of agency—which is, after all, a rather diverse group, with very different philosophical backgrounds, interests, and temperaments (as it should be evident by the contributions to this volume). But it seems undeniable that there is something at least initially attractive, if not even intuitive, about the standard story, so any view of the metaphysics of agency, at some point, might have to contend with it. Hence, it is worth pointing out what might make this picture attractive in the first place: first, the picture does not divorce agency from causation but actually makes agency fit within DOI: 10.4324/9780429202131-219

Luca Ferrero

a naturalistic understanding of our conduct; second, it recognizes the specific role of our psychology in shaping our agency, while also acknowledging the special role played by our bodily movements; it uses a metaphysics of events that seemingly fits both a standard understanding of causation and a plausible account of the individuation and location of actions, and how they relate to their consequences and results. The first four chapters, however, take issue with some of the central features of the standard theory: how to understand causation, whether genuinely teleological explanations can really fit with naturalistic explanations, whether a metaphysics of events is at all adequate to account for the temporal extension and the dynamics of agency, and whether the standard story can really account for the seemingly genuine agential character of omissions and refraining. Some of these criticisms might be taken as ultimately friendly to the standard story, suggesting how it could be improved. Others might be taken as inviting for a pluralism of approaches. Others still might be seen as criticizing the standard story as deeply misleading in its basic assumptions, possibly even in how it conceives of the basic desiderata for an account of agency. At this point, all these options are on the table. The point of this handbook is not to adjudicate the standard story (or any particular view of agency) but rather to explore the variety of approaches in the philosophy of agency. This exploratory spirit should be especially evident in the last two chapters in this section, on bounded agency and games, which might seem an unlikely fit for a section on the metaphysics of agency. Let me explain why they appear here. First, the chapter on ‘Bounded agency’ by Elijah Millgram contains a sketch of a basic account of the nature of agency (‘agency as determination’) that has the potential to be useful to cover, at a sufficiently high level of abstraction, all kinds of agency, not just full-blooded agency. As such, it already points in the direction of the metaphysical questions highlighted in the next section on ‘Kinds of agency.’ But at the core of the paper is an intriguing methodological claim: that philosophical reflection about agency should focus on ‘bounded’ rather than ‘ideal’ agency. That is, it should focus on real-world agencies with their distinctive architectures that operate in response to localized challenges, constraints, resources, rather than trying to account for agency as a generic capacity for determination that could in principle operate across widely different locations. The closing methodological suggestion of Millgram’s paper is that philosophers of agency should investigate various forms of agency “in the piecemeal manner of old-time botany and zoology.” It is partly to follow this recommendation that I have added the chapter on ‘Agency and games’ by C. Thi Nguyen at the closing of this section. The paper raises interesting metaphysical questions about the internal structure of our agency (especially about its internal layers and how we can navigate fluidly across them). But it does from an entry point—game playing—that might be unfamiliar to many philosophers of agency. I think that the topic of games and agency could help us wear the old-time naturalist hat, as Millgram urges us to do. This is because, as Nguyen argues, games contain a ‘library of agencies’—a rich repository of different codified forms of agency that might be used as an entry point for further explorations into the thick, deep, and lush philosophical jungle of agency.

Agency and causation What is the best way to characterize the relationship between the exercise of agency and causation? This is the topic of Jesús H. Aguilar and Andrei A. Buckareff’s ‘Agency and causation.’ As a preliminary matter, they characterize the power of agency as the power to settle whether a state of affairs obtains or not, rather than a power to act. In their view, this 20

The metaphysics of agency 

is preferable because we manifest the power agency not just in acting but also in omitting to act and in the intentional outcomes of our actions and omissions. There are three possible general approaches to the nature of causation in agency. First, agency as causal initiation: agency is exercised either in a basic mental action (an act of will or a trying) or by the agent-as-cause. One major problem with this approach is that it does not account for the guidance of extended pursuits beyond their initiation. Second, agency as outcome: agency resides in the effect or outcome of mental causes (as in the standard causal theory of action). The problem with this approach is that the agent seems to play no role: the mental attitudes do all the agential job. Third, agency as causal process: the process includes among its part the causing items (such as the intention to A) and their outcome (the A-ing). The process can either be Russellian (a diachronic series of discrete items in counterfactual dependence) or neo-Aristotelian (a process that has unity and direction in virtue of a constellation of causal powers, which are activated in mutual interaction, and whose collective outcome can be simultaneous with the causing). This latter approach is the one defended by Aguilar and Buckareff. They claim that it has a compelling advantage: it accounts for the exercise of agency in a way that by default involves the agent and takes place over time as an extended activity. A second important (but often neglected) question about the metaphysics of causation that bears on agency is whether causation is a relation of production rather than mere difference-making. According to the latter view, causation is non-productive; it is just a matter of counterfactual dependence or nomic regularity. According to causation as production, causings are the results of actual rather than counterfactual features of objects. Aguilar and Buckareff argue that causation as production is the correct view because it can satisfy the conditions of spatiotemporal locality and contiguity between action and its causal antecedents, which appear necessary to account for non-inferential epistemic authority, the sustaining of extended activities, and the context-sensitivity of agency. Difference-making theories, by contrast, cannot account for the causal connectedness between the agent and the outcome of their exercise of agency. Finally, if agency is causation by process, the question arises about the location, within the process, of the effective causes. There are three options. First, they could be located in the events. But this is unsatisfactory because it obscures the role of the capacity of the agent to make a causal difference. Second, they could be located in substances. This is a better option since the exercise of agency is now a manifestation of the agent’s causal powers. But there are still two concerns: the focus is on the agent qua substance (but not on how that substance causes what it does) while ignoring the contribution of other causal powers. The authors argue in favor of a third option: the effective causes are the powers manifested in the process. According to their neo-Aristotelian account, the powers are manifested in a unified teleological process, in a constellation of powers that collectively cause the outcome (a solution that shifts the focus away from the substance as a whole to the specific relevant ways in which a specific substance operates).

Agency, functions, and teleology At the core of the concept of agency is the idea that agents act for purposes or reasons: unlike the motions of rocks, planets, or elementary particles, actions are goal-oriented behaviors. Hence, it is no surprise that we ordinarily explain actions in teleological terms: we explain why the agent did what she did by citing the state of affairs toward which her behavior was directed. But teleological explanations raise a philosophical problem, as discussed by Scott Sehon in ‘Agency, functions, and teleology.’ Teleological explanations do not seem to have 21

Luca Ferrero

a place in the natural sciences. Does this mean that agency, especially human agency, stands mysteriously outside of the natural order? This conclusion is not warranted provided that we can find a way to reduce teleological explanations to other kinds of explanations available within the normal sphere of naturalistic science. Unfortunately, as Sehon argues, the two standard strategies of reduction fail. The first strategy reduces teleological explanations to straightforward causal explanations in disguise. In outline, actions are those behaviors that are caused by appropriate mental states (such as a combination of belief and desire). The trouble with this standard causalist explanation of action is the well-known problem of deviant causal chains. Not every possible way in which the mental states might cause the behavior in question would make the behavior a into genuine action. However, as Sehon argues in detail, attempts at specifying what counts as the proper, non-deviant way of causing an action have all failed. The second strategy is to reduce teleological explanations to functional or selectional explanations that are common in evolutionary biology. In rough outline, one could explain the action of a given organism in terms of the dispositions and abilities, or the corresponding mental states, that have been acquired via natural selection by organisms of that kind in virtue of the contribution of those dispositions/abilities/mental states to the past reproductive success of those organisms. Sehon argues that there are two main problems with this proposal. First, it is dubious that this analysis could be applied to many of our actions when they are quite removed from reproductive success. Second, the deviant causal chain problem arises for this reduction as well. Should we be troubled if human action cannot be fitted with in the realm of naturalistic science? The irreducibility of teleological explanations does not mean that actions have no causal explanation and that teleological explanations replace causal ones. We should rather see teleological explanations as addressing different explanatory questions. This still leaves open the issue of how to reconcile agency with the closure of physical explanations. Should we take agency as inserting itself into the causal order? Or should we rest content with some kind of cosmic coincidence between the two kinds of explanations? Both choices might seem unpalatable but, as Sehon concludes, “if no reduction [of teleological explanations] is forthcoming, then we must either become eliminativists about agency or conclude that the world is rather more mysterious than we might have thought or hoped.”

Agency, events, and processes The so-called standard story of action takes actions to be events with a special etiology: events caused in the right way by the relevant mental states, such as belief-desire pairs. This story, however, might be unable to accommodate the temporality of human agency—the fact that most of the things we do unfold over time and through several steps. We usually must progress toward the completion of our pursuits, and this progression does not guarantee eventual completion: pursuits can be interrupted before their successful termination. But, as Matthias Haase argues in ‘Agency, events, and processes,’ the standard story only makes room for complete actions as events that are fully determined particulars, and thus gives rise to the so-called imperfective paradox: it fails to explain why progression doesn’t necessarily imply completion. Haase considers several recent alternative accounts to the standard story. What they have in common is the introduction of processes into the ontology of agency. But, he argues, they are still problematic because they inherit some of the conceptual framework of the standard story. 22

The metaphysics of agency 

A radical alternative to the standard story is found in Michael Thompson’s development of Anscombe’s suggestion to focus on the aspectual contrast between the imperfective of the present progression (I am doing so-and-so) and the perfective of the past tense (I have done so-and-so). In Thompson’s view, the main metaphysical contrast is that between imperfection and perfection, between the pursuit of an end and its achievement. But this account, Haase argues, runs in the perfective paradox: the failure to explain how progression can lead to completion. A family of views has been recently developed to try to address these issues. What they have in common is the suggestion that what we describe as being in progress has some actual presence in time and space. This can be done in two basic ways. First, there is a single particular entity with two mutually exclusive properties: being in progress or being complete (what Haase calls the ‘two properties view’ or TPV). Alternatively, the progressive is taken to introduce another kind of temporal entity (the ‘two entities view’ or TEV). According to TPV, when a pursuit is in progress, there is already a concrete particular, which is the same concrete particular as the action (once the pursuit is properly completed). In the process version of TPV, this particular is a concrete unrepeatable individual process that has temporal parts and that can add more additional parts as it unfolds. TPV makes space for interruption of actual progression and avoids both the imperfective and progressive paradoxes. But it runs into the perfective paradox: it fails to explain why completion excludes current progression. The TEV seems to avoid this troubling implication. According to TEV, the contrast and the relation between progression and completion is due to two temporal entities of irreducible distinct kinds (sometimes called processes vs. events). Neither the imperfective nor the perfective paradoxes arise for TEV since progression and completion are associated with distinct entities. But this separation raises the question of how progression and completion are related and why completion requires the prior existence of progression. The proponents of TEV claim that the solution rests on the idea that the entity associated with the perfective is ‘composed’ or ‘comprised’ of the entity associated with the progressive. Everything turns on the thesis that the entity associated with the perfective is composed or comprised of the entity associated with the progressive, where the latter can be ‘wholly present’ at each time of its existence (that is, it does not have temporal parts, unlike processes for TPV). But Haase argues that the separation between the two entities gives rise to a version of the progressive paradox that already plagued the standard story: the conception of process suggested by TEV does not explain how it would be unfolding toward completion. In sum, neither the standard story nor its alternatives can escape the paradoxes related to the temporality of agency. For Haase, this is ultimately because all these accounts fail to account for the dynamic presence of action. There is nothing dynamic about the entities or properties that are at the core of these accounts. Haase conjectures that at the heart of the problem is a problematic assumption shared by all the existing accounts: that the metaphysic of (temporal) action is to be approached via the non-dynamic notion of instantiation and the correlated distinction between type and token.

Negative agency Actions are standard ways in which we exercise and manifest our agency. But what about omissions and refraining? As Randolph Clarke argues in ‘Negative agency,’ omissions and refraining are agential in important respects: we can omit and refrain intentionally, for reasons, and we can be held morally responsible for them and for their results. A comprehensive 23

Luca Ferrero

theory of agency therefore should encompass omitting and refraining as well as acting. Clarke doubts, however, that we can have a uniform theory that covers them all, especially a single causal account along the lines of the so-called ‘standard story of action.’ But, he claims, we can still offer sets of sufficient conditions to characterize both omitting and refraining, without committing to some ontically negative entities or questionable causal claims. For a causal theory, the easiest cases to handle are those in which a refraining is identical to the performance of a (positive) action that is negatively described. For instance, a mime who is refraining to move is actually holding the pose—he is standing still. Standing still is an action; a standard instance of ‘positive’ agency. If a similar strategy could be applied across all cases of negative agency, there would be a uniform account of all agential events in terms of the satisfaction of the same causal requirement. But Clarke argues that this is not possible. First, there are unintentional or unwitting omissions, for example, intending to stop to get milk on the way home but forgetting it. These are failures to act but not actions (and the results of some of these omissions might happen to fall outside of the possible sphere of causal influence of whatever the agent is doing when the omission takes place). Second, there are intentional omissions or refrainings, for example, refraining to reach for the chocolate, which one can do simply by letting things stand as they are. All the subject has to do is not do something—not move her arms, not move the chocolate. There might be no act of keeping her arms from moving since her arms might simply be hanging on her side. Relying on the gravitational forces is all it takes to keep the arms there. Sure, the agent’s position is under her control but that does not require that she is controlling it at that time by the performance of some action (for instance, the rate of breathing can be under our control since we can alter it at will, but most of the time we are not controlling it). Clarke, then, offers two sets of sufficient conditions for unwitting omitting and refraining. In both cases, even if one is exercising one’s capacity to do things intentionally, there need not be something that counts as an act of unwitting omitting or an act of refraining. As a final example, Clarke considers extended refraining, such as the resolution not to smoke. Unlike some cases of episodic refraining and possibly the first episodes of the resolution, one might carry out the resolution without performing any particular inhibitory act (or a string of them), such as resisting temptation. As time goes by, one might never be tempted to smoke again and yet one can still get credit for continuing to refrain from smoking and living up to one’s resolution. The morale that Clarke draws from his account of negative agency is that “agency is diverse.” There are agential phenomena that are not actions, and to which a simple uniform causal account does not apply. But the phenomena, even if ontologically and causally diverse, still form a coherent subject matter. As Clarke concludes, “reasons for action bear on omitting and refraining as well as on acting. An episode of any of these can reflect an agent’s practical identity. Each can be attributable to one as an agent, justifying ethical evaluation, reactive attitudes, and overt responses, as can their results. We have a coherent practical interest in this variety of things.”

Bounded agency Much work in the philosophy of agency tries to characterize the ideal agent in the sense of an agent who is always an agent. Real-world agency is then conceived as one that falls short of this ideal, but it has to be understood through the ideal form. Contrary to this picture, in ‘Bounded agency,’ Elijah Millgram argues that not only is real-world agency always bounded but that we cannot even make sense of an ideal and unbounded kind of agency. 24

The metaphysics of agency 

Millgram introduces the notion of ‘bounded agency’ by analogy with the concept of ‘bounded rationality:’ because of resource constraints, real-world agents inevitably face performance trade-offs, and whatever agential architecture they adopt, that architecture will only perform well in some environments but poorly in others. In his account, Millgram uses the notion of ‘agency as determination,’ as the capacity to stay on track by absorbing various kinds of noise and buffeting from the environment. Staying on track is to be understood not as the pursuit of a goal or end but as ‘carrying out a mission.’ Determination is a matter of absorbing prima facie defeaters (physical and rational alike) that would derail the mission. Supporting determination is expensive and in the real world always calls for some trade-off. The trade-offs that work in some locations, however, need not work in other locations (similarly to the use of heuristics in bounded rationality, where all heuristics are location-specific and no heuristic works well regardless of where it is used). A philosophically interesting example of this location-specificity are the different possible agential architectures (and related theories of practical reasoning): a goal-driven architecture does not work if the desires that the agent happens to have are moot or irrelevant; a plan-driven architecture is successful in a managerial society but not in the chaos of a failing state; a feedback-drive architecture selects new goals in terms of affective feedback and can climb successfully in certain welfare landscapes but can get stuck in addiction. An ideal agent would not face these problems if it could endlessly shift their agential architecture always to match their current environment. The ideal agent would be endlessly polymorphic. But Millgram argues that we cannot understand what this agent would be like. This is because an ideal agent would exhibit determination in any environment whatsoever by morphing their architecture in passing from one environment to another, in indefinitely many ways that no one could anticipate. But if so, what would the ‘mission’ of such an agent be? What would count as staying determined across arbitrary transformations of agential architecture? Since we cannot know what determination is like for an ideal agent, we should not try to understand real-world bounded agency in terms of ideal agency. Our goal, as Millgram recommends in closing, should rather be to investigate the different forms of bounded agency more “in the piecemeal manner of old-time botany and zoology.”

Agency and games In ‘Agency and games,’ C. Thi Nguyen shows that reflection on the nature of games can teach us much about the nature of agency, especially about its motivational and structural complexity. In addition, games—when considered as an art form—can be said to be the ‘art of agency’ because their artistic medium is nothing other than our own agency. The motivational structure of games is somewhat peculiar. In playing games, we temporarily take on alternate goals and an alternate, designed agency (with its own distinctive abilities). Whereas in ordinary enterprises our motivations are normally instrumental and transparent to our larger and enduring ends, when playing a game, we get temporarily absorbed in that game and care primarily about winning, in isolation from our enduring ends. However, we do not care about winning in an enduring way and we might play in the pursuit of other aims, such as having fun. But the best way to pursue these other aims is for us to try, during the game, sincerely to win, as if that were our only end. This is the characteristic motivational inversion of game play. Another interesting feature of games is that the bare state of affairs that we are trying to achieve is not necessarily what we value. According to Nguyen’s account (inspired by Bernard Suits), in playing games we add unnecessary obstacles on the way of pursuing the goal 25

Luca Ferrero

of the game; an addition that would not make sense in ordinary enterprises where we try to pursue our goals efficiently. But in gaming, the goal we care about is partially constituted by the inefficiencies imposed by that game. There are also two different motivational attitudes available when one engages in this obstacle-oriented play. In achievement play, one plays for the sake of the value of winning. In striving play, one temporarily takes an interest in winning for the sake of the value of the struggle (the two kinds of motivations might be present together in varying proportions). Striving play is the interesting kind given that, unlike ordinary activities, we don’t pursue means for the sake of the ends but we pick the ends for the sake of the means they force us to take. In striving play, we show our ability to create layers of agency: to manipulate ourselves to temporarily seal ourselves off from the enduring ends and be absorbed in the local instrumental struggle. These layers are what game designers can sculpt for us. Designers determine what we should care about, what abilities to use, and what obstacles to face. They do so to give us a particular experience of our own agency, which can be aesthetically valuable. Games can thus be an art form, where the artistic medium is agency itself. This goes together with another special function of games. Games, as Nguyen claims in closing, are a technology that allows us to record, learn, and communicate different modes of agency. Games provide a library of agencies, exposure to which can enhance our freedom and autonomy.

26

1 AGENCY AND CAUSATION Jesús H. Aguilar and Andrei A. Buckareff

In this chapter, we examine some foundational issues at the intersection of the metaphysics of agency and the metaphysics of causation. Regarding the metaphysics of agency, we favor a general conception of the power of agency as the power to settle whether a state of affairs obtains or not (Steward 2012 and Clancy 2013). We favor this way of thinking about agency over a common conception of agency in terms of manifesting the power to act (Schlosser 2015) for two reasons. First, we not only manifest the power of agency in acting but also in omitting to act (Sartorio 2009; Clarke 2014; Buckareff 2018). While we sometimes intentionally omit by acting, it is best to have an account of agency that can handle the metaphysical possibility of basic omissions. Second, we also exercise agency with respect to the intentional outcomes of our actions and omissions (e.g., winning a tennis match, relieving suffering). Understanding agency as settling affords us a more liberal conception of agency that accounts for more than just agency in acting. Assuming that agency is best understood as involving causation, having a serviceable metaphysics of causation is imperative. In what follows, we explore three broad issues concerning the metaphysics of causation and intentional agency. We first consider the best way to think about the relationship between exercising agency and causation. Specifically, is intentional agency best identified with a causal process or should we take intentional agency to be either the causal initiation of some outcome or the effect of a cause? Next, we examine the nature of the causal relation involved in exercising agency. We consider whether it is best to understand causation as productive or as mere difference-making. We finally consider where to locate the effective causes in agency. We consider three alternatives: events, substances, and powers manifesting in processes. In examining the alternatives, we present reasons for taking intentional agency to be a productive causal process involving constellations of manifesting reciprocal causal powers collectively generating an outcome.

1 Cause, outcome, or process? There is no consensus about how to think about the relationship between causation and intentional agency. Nevertheless, it is possible to distinguish three general approaches: agency as causal initiation, agency as causal outcome, and agency as causal process.

DOI: 10.4324/9780429202131-327

Jesús H. Aguilar and Andrei A. Buckareff

Agency as causal initiation. Many proponents of volitionism identify exercising intentional agency with basic mental action, such as an act of will or a trying. This basic mental action is then seen as causing other events in the form of mental activity or bodily movements (Hornsby 1980; Pietroski 2000; Lowe 2008). Accordingly, it is in the act of trying, willing, or choosing that we find the primary locus of agency. For some volitionists, the act itself is uncaused (Lowe 2008); for others, it has mental causes that motivate and explain it (Hornsby 1980; Pietroski 2000). Similarly, many agent-causalists identify the primary locus of an agent’s exercise of agency in the causing of an occurrence. On the account often ascribed to agent-causalists, agency is exercised when agents cause their actions (Chisholm 1966; Taylor 1966; Swinburne 2013). However, other agent-causalists locate agency in an agent’s manifesting powers enabling the agent qua substance to cause an intention that then triggers a causal chain of events that constitute a “wider observable action” (O’Connor 2009: 196; see also O’Connor 2000; O’Connor and Jacobs 2013). The agency-as-initiation approach is not without its liabilities. For instance, Harry Frankfurt (1978) forcefully argued for the centrality of agential guidance which requires that agents maintain their influence over their actions for as long as they are taking place. Agency-as-initiation approaches do not clearly afford their proponents an advantage in accounting for how agency can involve agents causally guiding their behavior given the emphasis on agential activity primarily occurring in the initiation of behavior.1 Agency as outcome. Another widely held view takes exercising agency to be the effect of mental causes that explain what an agent does. Given the widespread assumption that an exercise of agency is identical with the performance of an intentional action, it is easy to see why exercising agency might be regarded as a causal outcome. Many proponents of standard formulations of the causal theory of action (CTA) have held this view, identifying exercising agency with acting (Goldman 1970; Davidson 1980; Enç 2003). On the CTA, some behavior A (whether overt or mental) is an intentional action if and only if it is caused in the right way by some appropriate mental items that motivate and explain A (Aguilar and Buckareff 2010). A significant problem with identifying intentional agency with intentional actions and identifying actions with the effects of causes is that the agent is absent from the putative exercise of agency (Frankfurt 1978). By the moment an alleged exercise of agency is initiated, the key agential components that led to it in the form of some appropriate mental causes would have done their job, leaving the agent behind, as it were, estranged from their exercise of agency. The agent’s contribution would be over by the time the action commences. To cope with this type of problem, some CTA models introduce sustaining causal antecedents in the form of intentions and feedback mechanisms (Mele 1992; cf. Bishop 1989). However, any such move will still only get us so far since the causes of an action are external to the exercise of agency. Agency as causal process. A third view considers agency as involving a causal process that includes among its proper parts the causing items (e.g., an intention to A) and their outcome (the A-ing). This view finds support from proponents of variants of the CTA ( Thalberg 1977; Searle 1983; Dretske 1988; Bishop 1989; Stout 2002, 2007, 2012; Setiya 2007), agent-causalism (Alvarez and Hyman 1998), volitionism (Ginet 1990), and causal pluralism (Steward 2012). One way of understanding causal processes is to see them as essentially involving causal chains of events (Thalberg 1977; Dretske 1988). Rowland Stout attributes this sort of view to Bertrand Russell (1903). According to Stout, a Russellian process involves “a continuous series of states of affairs” that when causally connected constitute a causal process (2005: 88). 28

Agency and causation

It is reasonable to assume that “states of affairs” would include events in its extension. Thus, a Russellian causal process consists of discrete, diachronically ordered items that could stand in relations of counterfactual dependence to one another. In contrast to Russellian causal processes stands a neo-Aristotelian conception of causal processes.2 A causal process exhibits a unity and directedness inherited from the constellation of manifesting causal powers of its constituent components. Each causal power active in a causal process is directed at manifestations activated by interactions with other causal powers. A temporally extended causal process unfolds until it reaches the terminus toward which it is directed. It can be described in several ways (e.g., as fast, slow, impeded, accelerated, interrupted, completed, incomplete, or ongoing) owing to its constituent powers manifesting and how they may influence one another. Finally, the outcome of a causal process understood in this way can be simultaneous with the causing that is the manifestation of the powers of one or more of its components (Mumford and Anjum 2011). If exercising agency is a causal process (whether Russellian or neo-Aristotelian), then exercising agency commences with the initiation of a causal process and it would terminate with the intended outcome. This third view about agency raises significant questions with respect to the existence of basic actions and the precise boundaries of the relevant causal process. Nevertheless, a compelling advantage of seeing agency as a causal process is that exercising agency is by default agent-involving and eminently an activity. The other two approaches emphasize one of these two aspects of agency at the cost of losing the element de-emphasized, and hence they wind up with a diminished metaphysics of agency.

2 Production or mere difference-making? It is uncommon for those working in the philosophy of agency to specify whether they think of causation as productive or mere difference-making. With some exceptions (e.g., Taylor 1966; Davidson 1980; Gustafson 1986; Stout 2007; Lowe 2008; O’Connor 2009; Mayr 2011;O’Connor and Jacobs 2013; Swinburne 2013), the rule has been to keep questions about the metaphysics of causation at arm’s length from questions about the metaphysics of agency. In this section, we sketch the differences between productive and mere difference-making accounts of causation, offering reasons for why causation in agency is best understood as a productive relation.

2.1 Causation as mere difference-making The typical way of characterizing the non-productive causal framework is in terms of the nomic regularity view of causation associated with David Hume, which is then linked to counterfactual dependence views. Not only are these two views seen as non-productive accounts of causation, but the latter assumes the former. Canonical statements of non-productive theories of causation were articulated by J.L. Mackie (1974) and David Lewis (1986, 2000). There have since been increasingly sophisticated variants of this approach which share a core commitment to some basic ideas. Many proponents of such theories articulate their account of causation in terms of counterfactuals: an event E1 causes another event E 2, if E 2 would not have occurred without the occurrence of E1. This dependence relationship is also often formulated in terms of possible worlds: a given counterfactual is true, if in the closest possible worlds in which E 2 does not occur, E1 does not occur (Lewis 1986: 164–167). A world W1 is closer than a world W2 to the actual world WA, if W1 is more similar to WA than W2 is. And two worlds are more similar than some other worlds if they share particular matters of fact and natural laws. 29

Jesús H. Aguilar and Andrei A. Buckareff

The foregoing establishes a connection between causation as difference-making and laws of nature. According to the most prominent characterization of this connection, the so-called Ramsey-Lewis view: “laws are those generalizations which figure in the most economical true axiomatization of all the particular matters of fact that obtain” (Beebee 2000: 571). What allows for such descriptive contingencies to exhibit true laws is that they supervene on “the deepest, most pervasive, most explanatory regularities” over the course of the history of the universe (Beebee 2003: 273).3 So perhaps only the most fundamental regularities would provide the truth-makers for laws of nature. At this juncture we may wonder about the ontological commitments of accepting causation as mere difference-making, especially concerning the truth-makers for counterfactuals. For instance, it seems that modal realism has to be assumed as a first step to getting the required truth-makers. However, few, apart from Lewis, seem to fully appreciate this metaphysical consequence of a difference-making approach to causation.4 There is a further worry that we think is especially damning for difference-making approaches. Insofar as the relevant intentional antecedents of an action need to satisfy a spatiotemporal condition of locality and contiguity, agency requires real causal connectedness (Kim 2010). That such a condition is required seems clear when we look at some fundamental features of agency. Consider the unique epistemic authority agents have about their exercising agency. This type of non-inferential direct acquaintance concerning agential authorship is hard to explain unless the causes giving rise to an outcome in agency are so closely related that having epistemic access to one immediately provides epistemic access to the other (Ginet 1990). There is also the sustaining capacity of agential control required to causally maintain the necessary influence over an action for as long as this influence is required to complete it. Again, this type of sustaining relationship is hard to understand without constant causal connectedness (Bishop 1989; Stout 2007). Furthermore, agency is highly context-sensitive. The causal influence of the environmental local conditions in which an action occurs is often a determining contributing factor in its execution. Action guidance, for instance, involves monitoring the action as it unfolds, often adapting the execution to the dynamic conditions surrounding it. These are only some of the reasons supporting the view that agency requires constant causal connectedness. Difference-making theories cannot account for constant causal connectedness for at least two reasons (Kim 2010). First, no locality condition in the form of temporal and spatial contiguity is required. For counterfactual dependency to obtain it is not necessary that the relevant antecedent conditions need to be anchored in a time and place connected to an outcome. Imagine a causal scenario involving a bomb explosion captured by a counterfactual statement like: “Had agent A not placed the bomb at time t at location x, the explosion E would not have happened.” If one learns that A was a considerable distance from x when the explosion took place or perhaps that A was long dead by t, nothing about the truth of the counterfactual would change. The idea that an agent has to be present and in charge of exercising agency is foreign or irrelevant from the perspective of the counterfactual analysis. Second, difference-making causal theories are ultimately grounded on the nomic-regularity analysis of causation. The only plausible way to account for the distinction between possible worlds presupposed by a counterfactual analysis depends in turn on how similar those possible worlds are to the actual world. This similarity corresponds to nomic regularities. Once we reach such basic nomic regularities we are at a significant metaphysical distance from the type of connection that agents have with their exercises of agency that involve singular causation divorced from any relevant regularities. 30

Agency and causation

2.2 Causation as production or generation An alternative singularist view emphasizes the productive or generative aspects of causation and seems better suited to the explanatory demands of agency than its difference-making competitor. Jaegwon Kim (2010: 257) has insisted on the need to assume a productive account of causation in order to explain the connectedness between mental causes involved in our exercises of intentional agency and their outcomes. Without this view of causation, we are at a loss in making sense of the actual physical processes involved in, say, a limb moving as the result of the physical transfer of energy or momentum from some other item in the world associated with its mental causes. Kim traces this way of understanding causation back to the work of Elizabeth Anscombe who stresses the pervasiveness of this type of causation to the point of invisibility: There is something to observe here, that lies under our noses. It is little attended to, and yet still so obvious as to seem trite. It is this: causality consists in the derivativeness of an effect from its cause. This is the core, the common feature, of causality in its various kinds. Effects derive from, arise out of, come of, their causes. (Anscombe 1993: 91–92) It is this causal-productive dimension which, according to Kim, evades the neo-Humean model in both its counterfactual and nomic-regularity versions. The productive approach to causation encompasses several realist theories of causation that take causings to be the result of actual features of objects rather than counterfactual features.5 Among them is the causal power approach that we consider ideal for the purposes of accounting for the type of causal connectedness exhibited by exercises of agency.6 If causation involves the manifestation of some executive capacity of an agent in collaboration with other powers, then the agent is in a privileged epistemic position in the case of agency to have direct access to the manifestations of such a causal power without the need to corroborate or confirm its exercise by identifying its effect. Further, intentional agency understood as a process involving the activation of causal powers involves a constellation of elements working together in producing an outcome. The mutual activation of these powers in response to one another is what we earlier identified as affording intentional agency its unity as a causal process whose elements exhibit a level of connectedness that we do not find on mere difference-making accounts. Finally, on this sort of approach, in cases of synchronic agency, we have the outcome generated at the same time that we have the manifestations of causal powers in the process. Similarly, in cases of diachronic agency where the outcome is unfolding, the outcome is being produced at the same time that we have the activated powers manifesting. So we get an account of causation in agency on which the agent is causally connected to the outcome in the exercise of agency in a way sufficient for them to exercise guidance over what they are doing in real time.

3 Events, substances, or powers-in-process? Assuming it is best to understand intentional agency as a causal process of producing an outcome, we are left with the central metaphysical question about what is doing the productive work in this causal process. If we accept a Russellian approach to causal processes, then the answer is that events are doing the work. Alternatively, if we accept a neo-Aristotelian approach, such as the one we favor, then the answer is more complex involving powers 31

Jesús H. Aguilar and Andrei A. Buckareff

manifesting in a unified teleological process. This in turn splits the neo-Aristotelian approach into two camps: those who take causal powers to enable a substance to cause an outcome, and those, like ourselves, who take a constellation of manifesting powers to be the collective cause of an outcome. In considering these alternatives we make an indirect case for the second neo-Aristotelian approach.

3.1 Events as causes Proponents of Russellian causal processes regard them as constituted by chains of events. On this view, an exercise of agency would correspond to a series of causally related events typically involving the acquisition of a present-directed intention, some behavior, and some further intended outcome caused by such behavior. For example, a striker on a football team forms the intention to score a goal; this causes the bodily movements corresponding to kicking the ball, which, in turn, causes the ball to move toward its intended goal (Thalberg 1977; Dretske 1988). All the events involved in the chain interact diachronically to produce an outcome, and hence they make isolable causal contributions to its production. Furthermore, the sole unifier of the process is the executive mental cause, the conditions for satisfaction of which are provided by its content. So, according to this view, agential causal processes are essentially complex events. However, assuming that events are changes involving aspects of objects that produce outcomes in causal processes (see, e.g., Lombard 1979; cf. Kim 1976 and Davidson 1980), it is natural to ask in virtue of what they are causally productive. In response, we typically get a “just-so” story or perhaps an appeal to nomic subsumption of events under a law identified with a second-order relation between universals (Armstrong 1983). Postulating such second-order universals may seem promising, but it is not satisfactory since it fails to explain why those universals stand in such relations to one another. Furthermore, given that such explanations make use of higher-order relations between universals, a problematic regress looms in the background since these higher-order relations also need to be explained. This lack of explanation brings us to a second problem. As Anjan Chakravartty has argued, focusing on events “has the unfortunate consequence of obscuring the role played by those properties of things that we take to explain their causal behaviours” (2005: 19). Specifically, they obscure the role of an object’s first-order capacity to make a causal difference.

3.2 Substances as causes The appeal to substances as causal sources in the philosophy of agency is associated with the debate between proponents of event-causation and proponents of agent-causation, where agent-causation is understood as involving agents qua substances causing outcomes. The most sophisticated versions of agent causation propose that agents qua substances are enabled by their powers to produce outcomes. For example, an exercise of agency can be seen as the manifestation of an agent’s causal powers enabling them to do something (Mayr 2011; O’Connor and Jacobs 2013; Brent 2017). As an explanatory strategy that traces the causal sources of agential events, this is surely an improvement over any “just-so” story. Here we have a neo-Aristotelian account that explains why an agent did something in terms of the activation of the agent’s relevant causal powers enabling the agent qua substance to cause an outcome. The powers-enabling-substances-to-cause-outcomes view faces two significant challenges related to the way in which the participation of the agent and their powers should be understood. First, the agent-causal strategy focuses on the agent qua substance as the cause 32

Agency and causation

rather than on the ways the agent qua substance is at the time causation takes place. But presumably it is because a substance exhibits some features rather than others at a time that it causes something rather than something else or anything at all. Moreover, this view obscures the fact that it is not just the causal powers of the agent qua substance that are relevant in agential processes. In the case of overt movements, there are the causal powers of other objects in the agent’s environment that also contribute to produce an outcome. To use a simple case, consider an agent who is hungry and perceives a pear and reaches to grab it. In this case, the causal powers of the pear would have to interact with the causal powers of the agent’s nervous system to initiate and guide the action until the agent successfully grabs the pear. So, it seems that the correct picture of a causal exercise of agency by a substance includes not only various different ways the agent is, but ways that other substances are. The second challenge to the view that proposes substances as causes is related to the first. If one were to enrich the model by focusing on the different causal powers activated in producing an action, it appears that adding the substance as the efficient cause of an outcome results in a sort of double-counting. It is true that a substance is causally relevant in a causal process given that the actual occurrence of the causal process depends on the substance being the bearer of the causal powers that are active. But it is only because of the ways the substance is at the time that it can claim to be causally active. This only begs the question as to what the causally productive contribution of the substance would be apart from the supporting role of bearing the subset of causal powers that are causally productive at the time of exercising agency.7

3.3 Powers-in-process as causes From the foregoing criticisms of both event-causal and substance-causal views, an indirect case has already been offered for accepting a powers-in-process view. Unlike the eventcausal view, we have an account that has more explanatory power. The individual powers constitutive of any constellation of interacting mutually manifesting powers at a time are each directed at a multitude of different manifestations with various powers. The directionality of these powers explains why they do what they do when they interact with one another. We get an explanatorily richer account of intentional agency as a result. And, unlike the substance causal view, our focus is shifted away from the substance as a whole to specific relevant ways a substance is. Finally, we don’t have double-counting worries.

4 Conclusion The difference we make in the world as agents is causal. We have examined three fundamental problems related to the metaphysics of causation and intentional agency. In trying to address the challenges posed by these issues, we offered reasons in support of understanding intentional agency as a productive causal process involving interacting causal powers collectively generating an outcome. This sort of account has been employed to provide responses to problems raised for causalist approaches to theorizing about agency, including the nature of agents (Aguilar and Buckareff 2015), agential guidance (Stout 2012), basic causal deviance (Stout 2007), and intentional omissions (Buckareff 2018). This sort of account continues to be a minority view among philosophers of agency.8 And, while there are exceptions, any clear articulation of an author’s assumptions about the metaphysics of causation are mostly lacking in the philosophy of agency. We hope this will change, given that many proposed solutions to problems about agency are only as tenable as the assumptions about the nature of causation held by their proponents. 33

Jesús H. Aguilar and Andrei A. Buckareff

Related topics Agency, events, and processes; Agency, functions, and teleology; The aim of agency; Negative agency; Intentional agency; Agency, powers, and skills.

Notes 1 See Aguilar (2020) for a discussion of the significance of Frankfurt’s treatment of guidance for thinking about agency and what it reveals about the limitations of some causalist strategies. 2 The view being advanced may not be Aristotle’s but it is influenced by Book V of his Physics (1996) and Metaphysics Θ (2016). 3 See also Beebee (2006). 4 See Heil (2015) for more on these issues. 5 Realist theories include accounts that emphasize the transference of a conserved quantity (e.g., Dowe 2000 and Fair 1979), persistence of a property (e.g., Ehring 1997), nomic subsumption theories (e.g., Armstrong 1983), primitivism (e.g., Anscombe 1993), and causal power theories (see note 6). 6 See, e.g., Chakravarrty (2005), Heil (2012), Marmodoro (2017), and Mumford and Anjum (2011). 7 See Buckareff (2017) for a general critique of substance causation along these lines (see Kuykendall 2019 for a critical response). For a critique that is more specific to the context of exercising agency, see Buckareff (2011) (see Kuykendall 2021 for a critique). 8 Hyman (2015) approximates the view, but he sticks to the framework of events, which strikes us as ill-advised.

Further reading Buckareff, A. 2018. ‘I’m just sitting around doing nothing: on exercising intentional agency in omitting to act’. Synthese, 195: 4617–4635. A causal process theory of agency is presented on which intentional agency involves an agent’s causal powers manifesting. The account is applied to the problem of omissions. Kim, J. 2010. ‘Causation and mental causation’. Chap. 12 in Essays in the metaphysics of mind. New York: Oxford University Press: 243–262. An understanding of causation as a productive relation is presented as required for theorizing about mental causation and agency. Lowe, E.J. 2008. Personal agency: the metaphysics of mind and action. New York: Oxford University Press. Event and substance causation are compared in chapters 6–8, with a case made for the priority of substance causation over event causation. Menzies, P. 2003. ‘The causal efficacy of mental states’. In S. Walter and H. Heckmann, eds., Physicalism and mental causation: the metaphysics of mind and action. Exeter: Imprint: 195–223. A sophisticated variant of a difference-making account of causation is developed as part of a non-reductive physicalist account of mental causation. A response to the problem of basic causal deviance for causalism assuming this account of causation is given. Stout, R. 2007. ‘Two ways to understand causality in agency’. In A. Leist, ed., Action in context. New York: de Gruyter: 137–153. Causal chain theories are critiqued. A neo-Aristotelian alternative is defended on which exercises of intentional agency are identified with causal processes involving the manifestation of an agent’s power of intentional agency.

References Aguilar, J. 2020. ‘The standard story of action and the problem of agential guidance’. Crítica: Revista Hispanoamericana de Filosofia, 52: 3–25. Aguilar, J. and Buckareff, A. 2010. Causing human actions: new perspectives on the causal theory of action. Cambridge, MA: Bradford/The MIT Press. Aguilar, J. and Buckareff, A. 2015. ‘A gradualist metaphysics of agency’. In A. Buckareff, C. Moya, and S. Rosell, eds., Agency, freedom, and moral responsibility. New York: Palgrave-Macmillan: 30–43.

34

Agency and causation Alvarez, M. and Hyman, J. 1998. ‘Agents and their actions’. Philosophy, 73: 219–245. Anscombe, G. E. M. 1993. ‘Causality and determination’. In E. Sosa and M. Tooley, eds., Causation. New York: Oxford University Press: 88–104. Aristotle. 1996. Physics. New York: Oxford University Press. Aristotle. 2016. Metaphysics. Indianapolis: Hackett. Armstrong, D. M. 1983. What is a law of nature? New York: Cambridge University Press. Beebee, H. 2000. ‘The non-governing conception of laws of nature’. Philosophy and Phenomenological Research, 61: 571–594. Beebee, H. 2003. ‘Local miracle compatibilism’. Noûs, 37: 258–277. Beebee, H. 2006. ‘Does anything hold the universe together?’ Synthese, 149: 509–533. Bishop, J. 1989. Natural agency: an essay on the causal theory of action. New York: Cambridge University Press. Brent, M. 2017. ‘Agent causation as a solution to the problem of action’. Canadian Journal of Philosophy, 47: 656–673. Buckareff, A. 2011. ‘How does agent-causal power work?’ The Modern Schoolman, 88: 105–21. Buckareff, A. 2017. ‘A critique of substance causation’. Philosophia, 45: 1019–1026. Buckareff, A. 2018. ‘I’m just sitting around doing nothing: on exercising intentional agency in omitting to act’. Synthese, 195: 4617–4635. Chakravarrty, A. 2005. ‘Causal realism: events and processes’. Erkenntnis, 63: 7–31. Chisholm, R. 1966. ‘Freedom and action’. In K. Lehrer, ed., Freedom and determinism. Atlantic Highlands, NJ: Random House: 11–44. Clancy, S. 2013. A strong compatibilist account of settling. Inquiry, 56: 653–665. Clarke, R. 2014. Omissions: agency, metaphysics, and responsibility. New York: Oxford University Press. Davidson, D. 1980. Essays on actions and events. New York: Oxford University Press. Dowe, P. 2000. Physical causation. New York: Cambridge University Press. Dretske, F. 1988. Explaining behavior: reasons in a world of causes. Cambridge, MA: Bradford/The MIT Press. Ehring, D. 1997. Causation and persistence. New York: Oxford University Press. Enç, B. 2003. How we act: causes, reasons, and intentions. New York: Oxford University Press. Fair, D. 1979. ‘Causation and the flow of energy’. Erkenntnis, 14: 219–250. Frankfurt, H. 1978. ‘The problem of action’. American Philosophical Quarterly, 15: 157–162. Ginet, C. 1990. On action. New York: Cambridge University Press. Goldman, A. 1970. A theory of human action. Englewood Cliffs, NJ: Prentice-Hall. Gustafson, D. 1986. Intention and agency. Dordrecht: D. Reidel. Heil, J. 2012. The universe as we find it. New York: Oxford University Press. Heil, J. 2015. Aristotelian supervenience. Proceedings of the Aristotelian Society, 115: 41–56. Hornsby, J. 1980. Actions. New York: Routledge and Kegan Paul. Hyman, J. 2015. Action, knowledge, and will. New York: Oxford University Press. Kim, J. 1976. ‘Events as property exemplifications’. In M. Brand and D. Walton, eds., Action theory. Dordrecht: D. Reidel: 310–326. Kim, J. 2010. ‘Causation and mental causation’. Chap. 12 in Essays in the metaphysics of mind. New York: Oxford University Press: 243–262. Kuykendall, D. 2019. ‘Powerful substances because of powerless powers’. Journal of the American Philosophical Association, 5: 339–356. Kuykendall, D. 2021. ‘Agent causation, realist metaphysics of powers, and the reducibility objection’. Philosophia, 49: 1563–1581. Lewis, D. 1986. Philosophical papers, vol. II. New York: Oxford University Press. Lewis, D. 2000. Causation as influence. Journal of Philosophy, 97: 182–197. Lombard, L. 1979. ‘Events’. Canadian Journal of Philosophy, 9: 425–460. Lowe, E.J. 2008. Personal agency: the metaphysics of mind and action. New York: Oxford University Press. Mackie, J. L. 1974. The cement of the universe. New York: Oxford University Press. Marmodoro, A. 2017. ‘Aristotelian powers at work: reciprocity without symmetry in causation’. In J. Jacobs, ed., Causal powers. New York: Oxford University Press: 57–76. Mayr, E. 2011. Understanding human agency. New York: Oxford University Press. Mele, A. 1992. Springs of action: understanding intentional behavior. New York: Oxford University Press. Mumford, S. and Anjum, R. 2011. Getting causes from powers. New York: Oxford University Press. O’Connor, T. 2000. Persons and causes: the metaphysics of free will. New York: Oxford University Press.

35

Jesús H. Aguilar and Andrei A. Buckareff O’Connor, T. 2009. ‘Agent-causal power’. In T. Handfield, ed., Dispositions and causes. New York: Oxford University Press: 189–214. O’Connor, T. and Jacobs, J. 2013. ‘Agent-causation in a neo-Aristotelian metaphysics’. In S. Gibb, R. Ingthorsson, and Lowe, E.J., eds., Mental causation and ontology. New York: Oxford University Press: 173–192. Pietroski, P. 2000. Causing actions. New York: Oxford University Press. Russell, B. 1903. Principles of mathematics. New York: Cambridge University Press. Sartorio, C. 2009. Omissions and causalism. Noûs, 43: 513–530. Schlosser, M. 2015. Agency. In E. Zalta, ed., The Stanford encyclopedia of philosophy, Fall 2015 edition,

Searle, J. 1983. Intentionality. New York: Cambridge University Press. Setiya, K. 2007. Reasons without rationalism. Princeton: Princeton University Press. Steward, H. 2012. A metaphysics for freedom. New York: Oxford University Press. Stout, R. 2002. ‘The right structure for a causal theory of action’. Facta Philosophica, 4: 11–24. Stout, R. 2007. ‘Two ways to understand causality in agency’. In A. Leist, ed., Action in context. New York: de Gruyter: 137–153. Stout, R. 2012. ‘Mechanisms that respond to reasons: an Aristotelian approach to agency’. In F. O’Rourke, ed., Human destinies: philosophical essays in memory of Gerald Hanratty. Notre Dame: University of Notre Dame Press: 81–97. Swinburne, R. 2013. Mind, brain, and free will. New York: Oxford University Press. Thalberg, I. 1977. Perception, action, and emotion: a component approach. Oxford: Basil Blackwell. Taylor, R. 1966. Action and purpose. Englewood Cliffs, NJ: Prentice-Hall.

36

2 AGENCY, FUNCTION, AND TELEOLOGY Scott Sehon

1 Introduction At the heart of the concept of agency is a fundamental distinction: some events are actions performed by agents, and other events are what we might call mere happenings, even if they likewise involve the bodies of human beings. For example, if I decide to donate money to the Democratic Party, and if I then proceed to write a check, then that is an action. While I sign the check, the hair on my head continues to grow, but this is something that merely happens within my body. There is a causal sense in which my body is responsible for the growth of my hair, but as an agent, I deserve no credit for it. Actions are the things that we do: events in our lives that happened on purpose or for a reason; events that are goal-directed. This way of putting the point suggests a close connection between agency and teleology, for a teleological explanation is one that explains an event by citing its purpose, the state of affairs towards which it was aimed or directed. In the abstract, teleological explanations have the following form: Agent A did B in order to C. For example, we might say: I pulled out my chequebook in order to donate money to the Democratic party. By contrast, bodily happenings with no purpose are not actions and are not explicable teleologically. So it seems natural to conclude that actions, as opposed to mere happenings, are precisely those events that can be explained teleologically. However, the connection to teleology threatens to make the phenomenon of agency seem rather puzzling. We do not use teleological explanations when explaining motions of rocks, planets, or elementary particles. If teleological explanations apply in the realm of human agency but not to other natural events in the world, then this suggests that people and their actions stand mysteriously outside the natural order. On the other hand, if teleological explanations reduce to something within the normal sphere of naturalistic science, the problem would dissolve: acting for a reason would be no more mysterious than other phenomena of nature.

DOI: 10.4324/9780429202131-437

Scott Sehon

Two strategies for reduction suggest themselves. According to the first, teleological explanations are really just straightforward causal explanations in disguise. On the second, teleological explanations are a form of the functional or selectional explanation that we see in biology. This entry examines these strategies, concluding that both face substantial obstacles.

2 Causal reduction and deviant causal chains One might propose the following simple reduction of teleological explanation to causal explanation: (1) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B On this proposal, actions (i.e., behaviours done for a purpose) are simply those behaviours caused by appropriate mental states. The causalist might not want to commit to the details of the right side of the biconditional. Perhaps intentions cause the behaviour, rather than desires and beliefs; or perhaps we should speak of the neural correlates of the appropriate mental states as the causes (cf. Mele 2013). Irrespective of such details, the analysis proposed in (1) has a well-known defect: If the mental states cause the behaviour but do so in an abnormal or deviant way, then the right side of the biconditional might be met even when it seems intuitively that the agent did not act purposively. Here is an example of this sort from Donald Davidson: A climber might want to rid himself of the weight and danger of holding another man on a rope, and he might know that by loosening his hold on the rope he could rid himself of the weight and danger. This belief and want might so unnerve him as to cause him to loosen his hold, and yet it might be the case that he never chose to loosen his hold, nor did he do it intentionally. (2001, 79) As applied to (1), A is the climber, B is loosening his hold on the rope, and C is ridding himself of the weight and danger of holding another man on the rope. Davidson’s example then provides us with a case in which the right side of the biconditional of (1) is apparently true but the left side is false. So the right side fails as a sufficient condition. At first glance, this might appear to be just a small technical matter. The oddity about Davidson’s example seems to be that the desire caused the behaviour in a strange or abnormal way. Perhaps we need merely amend the analysis: (2) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B in the right way. However, if the causalist merely adds the phrase ‘in the right way’ without spelling out what this means in causal terms, then the analysis becomes empty. (2) effectively attempts to stipulate that there is a causal analysis of teleology without actually providing it. Perhaps the causalist can attempt to rule out the sorts of states that make the causal chains seem deviant. In Davidson’s example, the climber’s belief and desire unnerved him—that is to say, the mental states caused a state of nervousness or of being emotionally upset; the belief and desire did not cause the behaviour directly. This is a common theme in many other 38

Agency, function, and teleology

examples of deviant causal chains since Davidson. Accordingly, one might try to rule out such cases: (3) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B without any intervening states. However, this will clearly not work. Given the nature of the human body, mental states will never cause bodily motions immediately and without intervening states, as the mental states will work via nerve signals from the brain to the body. This means that there will be many cases in which the left side of the biconditional is true, but the right side is false and fails as a necessary condition for goal-direction. One might try to work around this by ruling out certain sorts of intervening states, namely psychological states like nervousness or distress: (4) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B without any intervening psychological states like nervousness. However, (4) would also fail as a necessary condition. As George Wilson (1989, 252) points out, in some cases performing an action might positively require an intervening state of nervousness. For example, suppose a teenager at a swimming pool hopes to impress a love interest by jumping off the high dive. He climbs the steps, but when he looks down at the water far below he is petrified with fear. However, his thought that he could impress the girl by jumping off the diving board triggers a rush of nervous excitement that gets him moving again, and he takes the plunge. The left side of the biconditional comes out true: he jumped off the diving board in order to impress the girl; but the right side will be false. (For a subtler attempt at an analysis along these lines, see Schlosser 2007; for a reply to it, see Sehon 2016, 98–102.) Rather than ruling out intervening states, the causalist might suggest that the crucial factor is the manner in which the causal interaction takes place: the bodily behaviour must be guided to the desired result. (Mele 1992, 182; 2007, 345, suggests the importance of the notion of guidance. However, Mele also makes clear that he does not intend to be putting forward an analysis of teleology, so I do not mean to be attributing any such analysis to him.) The idea faces two sorts of dangers. Suppose, at one extreme, the proposal were this: (5) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B, and A guided the behaviour B to the goal of C. However, stipulating that the agent guided the behaviour to that goal essentially adds the condition that the agent A did B in order to C; the analysis has become circular and does not provide an explanation teleology in purely causal terms. On the other hand, the causalist might try to spell out the relevant notion of guidance in terms of some sort of feedback loop. (6) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B, and there is a systematic feedback loop from the bodily motions back to the desire and belief, such that if the bodily motions are off course from doing B, the desire and belief will cause corrections. 39

Scott Sehon

However, this fails as a necessary condition: there could be cases in which A did B in order to C but where this more stringent condition is not met. Suppose a basketball player steals the ball and sees that the court is clear for her to execute a right-handed layup at the other end. She desires to score a basket, and she believes she can do so by running down the court and laying the ball in the hoop. Along the way her behaviour will be intelligent: she will make minor adjustments in her steps to arrange to be jumping off of her left foot in her final step; if a defender catches up with and hits her arm she might shoot more quickly than previously planned (so as to have been fouled in the act of shooting), and so forth. However, none of this intelligent goal-direction requires that the original belief and desire themselves were engaged in a causal feedback loop. Perhaps the brain states that realized the original belief and desire handed off operations to other brain states; this would mean that the right side of the biconditional in (6) is false, although we would still say that the player ran down the court in order to score a basket. Perhaps the causalist could work something out between the extremes of (5) and (6), but it is not easy to see how this would go (for a more detailed discussion of this sort of proposal, see Sehon 2016, 93–98). Neil McDonnell (2015) suggests a rather different approach, one that I will consider in more detail. He claims that typical examples of deviant causal chains ‘violate a proportionality constraint on causation’ (2015, 162). He describes the notion of a proportionate cause in the following way: …A cause is proportional to its effect if and only if there is no more general specification of the event that would have sufficed for the effect (no more determinable specification), and there is no more precise specification of the cause (no more determinate specification) that is required. (2015, 162) To illustrate the concept, McDonnell draws on an example from Yablo (1992). Suppose a pigeon has been trained to peck at red things. Now suppose the pigeon sees a scarlet triangle and pecks at it. ‘The pigeon pecked because it was a scarlet triangle’ would not be a proportionate description of the cause, for a more general specification of the event would have sufficed: we could have said, ‘the pigeon pecked because it was a red triangle.’ Or, from the other direction, if the pigeon had been trained to peck at only scarlet triangles, then describing the cause as a red object would not be proportionate; a more precise specification of the cause would have been appropriate, since the pigeon would not have pecked at any red object but only at scarlet triangles. McDonnell suggests that to avoid the deviant cases, ‘defenders of a causal theory need only add a clause which states that the causal connections in question be proportional’ (2015, 168). So the revised analysis would look like this: (7) A did B in order to C ≡ A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire were proportionate causes of her doing of B. In Davidson’s example of the climber we had the following: A: the climber B: loosening his hold on the rope C: ridding himself of the weight and danger of holding another man on the rope

40

Agency, function, and teleology

McDonnell claims that this fails as a counterexample to (7), for the climber’s desire for C and his belief that he can achieve C by doing B were not proportionate causes of his doing B. McDonnell says that the desire and belief were causes, but he claims that there was a ‘more general specification of that event that would have caused the same state’ (2015, 165). Given the context, the content of the climber’s belief and desire ‘belongs to a broader type of thought, of wilfully harming someone else, and the consequences of such’ (2015, 165). With the A, B, and C described above, the climber did have a desire for C, but the description of his belief (that he could achieve C by doing B) was not the proportionate one, because it was the thought of wilfully harming someone else that led to the nervousness. The proportionate cause would be the belief under this description: ‘I can rid myself of the weight and danger of holding another man by wilfully harming someone else.’ Thus, instead of B above, we have: B1: wilfully harming someone else. In sum, McDonnell claims that there is a more general specification of the climber’s belief that he can achieve C by doing B, namely, that he can achieve C by doing B1. This means that with A, B, and C as above, the right side of the biconditional is false, because the belief as specified was not a proportionate cause. But if we were to consider Davidson’s example with A, B 1, and C, it would also fail, for it is false that the climber has a desire to wilfully harm someone else. So, either way, Davidson’s case fails as a counterexample to (7). McDonnell’s key claim is that we have two specifications of the same belief, and that the first is more general: But it is unclear why we must see these as different specifications of the same belief. We can imagine the climber thinking to himself, ‘if I loosen my hold on the rope, I can rid myself of the weight and danger of holding another man’, and then subsequently having disturbing mental images of the other man falling to his death. His initial belief was simply about ridding himself of the weight in danger by loosening his hold, and this thought in turn triggered disturbing mental images which caused him to involuntarily loosen his grip. Seen this way, the climber’s belief is simply as in (ii), and there is no more general specification of that very event that would have sufficed for the effect. So the climber’s belief, as given in (ii), is a proportionate cause; the case still makes the right side of (7) true but the left side false. That is to say, McDonnell’s condition fails as a sufficient condition for goal direction. McDonnell’s proposal may be subject to counterexamples in the other direction as well. Suppose that Kerry leaves the house in order to go Christmas shopping for her mother’s present. Given (7), this entails that she desired to go Christmas shopping for her mother’s present, that she believed she could do so by leaving the house, and that these were proportionate causes of her behaviour, meaning that ‘there is no more general specification of the event that would have sufficed for the effect’ (2015, 162). However, one might think that there was a more general specification of Kerry’s desire: perhaps a simple desire to go Christmas shopping would have sufficed. But on (7), this means that is false to say that Kerry left the house in order to go Christmas shopping for her mother. On the likes of (7), the only correct explanation of the form ‘A did B in order to C’ is the one that involves the proportionate

41

Scott Sehon

cause—the cause that is neither too general nor too specific. Common sense usage is not so exacting: it seems that we could say either that Kerry left the house in order to go Christmas shopping or that Kerry left the house in order to go Christmas shopping for her mother. So, McDonnell’s proposal would also fail as a necessary condition.

3 Selectional explanations and teleology If straightforward causal reduction of teleological explanation does not work, then we might turn to functional or selectional explanations. The beauty of Darwin’s theory of evolution was that he showed how apparent design or purpose could be fully accounted for by variation and natural selection. Prior to Darwin, most people thought it obvious that teleological explanations were appropriate and irreducible when applied to things like bird wings or human hearts. Darwin showed us otherwise, and we might think the same lesson could apply to teleological explanations of human action. We could start by saying that the proper function of something is given by the following: It is a/the proper function of an item (X) of an organism (O) to do that which items of X’s type did to contribute to the inclusive fitness of O’s ancestors, and which caused the genotype, of which X is the phenotypic expression, to be selected by natural selection. (Neander 2018) This formulation is about features of an organism, rather than behaviours, so it is not of immediate help in attempting to analyse teleological explanations of action. However, one might plausibly think that it can be adapted to behaviours as well, at least in certain sorts of cases. Consider: The cat arched its back in order to appear larger and frighten away the dog. The cat’s behaviour is not a feature of the cat, but we might speak instead of the cat’s disposition to behave in this way (namely, to arch its back in circumstances like this), and we might conclude that the proper function of this disposition is to frighten away predators. That would be true if instances of this disposition in the cat’s ancestors indeed frightened away some predators and if this contributed to their inclusive fitness, and this seems reasonably plausible. The strategy here would need to be modified for other behaviours. Consider Jerry Fodor’s description of the behaviour of his own ‘strikingly intelligent cat’: In the morning, at his usual feeding time, Greycat prowls the area of the kitchen near his food bowl. When breakfast appears, he positions himself with respect to the bowl in a manner that facilitates ingestion. (1987, ix) Greycat’s evolutionary ancestors did not have a disposition to prowl around metal bowls in kitchens. However, a broader evolutionary strategy still suggests itself here. Greycat’s ancestors did have a general disposition to seek food when hungry, as well as a general ability to learn from previous experiences in which food was obtained. More generally, we might say that evolution instilled in Greycat a disposition to seek certain ends (states that would have been good for the reproductive success of Greycat’s evolutionary ancestors) as well as 42

Agency, function, and teleology

a certain ability to do means-end reasoning, mostly based on previous experience. We could put this in the language of beliefs and desires: evolution instilled in Greycat a disposition to desire to eat when hungry, and having this disposition would have generally been good for the reproductive success of Greycat’s ancestors; evolution also endowed Greycat with enough learning ability and means-end reasoning to believe that being near the food bowl in the kitchen in the morning was a good way of fulfilling this desire. Putting these ideas together, one might propose something like this general account of teleological explanations: (8) A did B in order to C iff: (i) A had a desire for C and a belief that she could achieve C by doing B, and this belief and desire caused her doing of B; (ii) A’s disposition to desire C in this sort of circumstance contributed to the inclusive fitness of A’s ancestors, and caused the genotype of which the disposition is the phenotypic expression, to be selected by natural selection; and (iii) A’s belief that she could achieve C by doing B resulted from an ability to learn and do means-end reasoning, and these abilities likewise contributed to the inclusive fitness of A’s ancestors. However, two things stand out about this sort of proposed analysis. First, the analysis is dubious when applied to many routine actions that are further removed from reproductive success. For example, suppose that a tenured philosopher walks to her office in order to work on a book manuscript. Condition (ii) would require that a disposition to desire to write a book contributed to the fitness of the philosopher’s ancestors. This is, of course, silly, since books are a relatively recent innovation. One might reply by suggesting that the philosopher has some further desire—that the desire to work on a book ultimately serves some other end that the philosopher desires, and that the disposition to have this ultimate desire was good for the reproductive success of the philosopher’s ancestors. For such an account to get off the ground, we would need to find a candidate desire, D, such that: (a) the disposition to have the desire for D was good for the reproductive success of the philosopher’s ancestors; (b) the philosopher has the desire for D; (c) the philosopher reasons that she can achieve D by working on a book. Here is a candidate for the ultimate desire D: desiring to feel cultural approval. This might plausibly fulfil condition (a), insofar as cultural approval might increase one’s chances of finding a suitable mate. However, (b) and (c) are more difficult. The philosopher might report having no such desire, and the philosopher might insist that even if she has such a desire, it was not the reason she walked to her office. A second issue for (8): it runs right back into the deviant causal chains problem. Thus, (8) is just like the original proposed analysis in (1) except that it adds conditions (ii) and (iii). However, these conditions add nothing that helps with the deviant cases. Consider a variant of Davidson’s case, one we can call the killing climber. The killing climber quite deliberately loosens his grip on the rope, not being the slightest bit unnerved or disturbed by the thought of causing another person’s death. If something along the lines of (8) works as an analysis of teleological explanation, then it will work for the killing climber, 43

Scott Sehon

perhaps requiring the postulation of some ultimate desire, such as the desire to be free of a present danger. But if this works for the killing climber, then the conditions in (ii) and (iii) will likewise be met for Davidson’s original climber, whom we can stipulate to have had the same desires and general dispositions as the killing climber. This means that the entire right side of the biconditional will be met for the original climber, since, by hypothesis in the original case, the climber’s belief and desire indeed cause the loosening of his grip, albeit by making the climber so unnerved that he involuntarily loosened his grip—he didn’t do it purposively.1

4 Conclusion If the arguments of this entry are correct, then the naturalistic project of reducing teleology to causal or functional explanation has serious problems. This conclusion might be troubling, insofar as it seems to place human action outside the realm of naturalistic science. One might even think that the irreducibility claim requires that human action violates the known laws of physics: if all explanation in physics is causal explanation and if teleological explanation of human action is not reducible to causal explanation, then human actions will be events that cannot be predicted or explained by physics. However, this conclusion does not quite follow. Consider the action of our philosopher, who walked to her office to work on a book. Her walking was a physical event, and we explain it teleologically by saying that it was directed at the aim of working on her book. However, we need not deny that her physical motion will also have a causal explanation, presumably involving brain states and nerve signals; one need not say that the teleological explanation replaces the causal explanation. We can consistently maintain that the teleological explanation answers a different question about the event. The physicist can answer the question of the cause of the motion of the philosopher’s body, but if we want to know the purpose (if any) of the motion, then this is simply a different question, one not reducible to questions about the causes of the behaviour. This is not to say that all mysteries have been resolved. If teleology, and the phenomenon of agency more broadly, do not reduce to the sorts of explanation employed in physics or even biology, then it seems that we have two choices. First, we might say that agents in some way interfere with physical events to make it the case that teleological explanations come out as true. This would mean that physicists and neuroscientists will one day observe events involving human beings that are contrary to what one would expect by simply applying the laws of physics. Or, second, we might say that the physical events of the world—all causal chains—just happen to line up such that some of them have true teleological explanations. Physics would still be closed, but, as if through some cosmic coincidence, there will also be true teleological explanations, despite the fact that teleology does not reduce to the explanations of physics. Many philosophers will find both of these choices quite unpalatable, which suggests that there is indeed substantial motivation for seeking a naturalistic reduction of agency in general and teleological explanation in particular. However, if no reduction is forthcoming, then we must either become eliminativists about agency or conclude that the world is rather more mysterious than we might have thought or hoped.

Related topics Agency and causation; Agency, events, and processes; The aim of agency. 44

Agency, function, and teleology

Note 1 This entry has not explicitly engaged with the philosophers who have worked on teleological theories of mental content, or with teleosemantic theories (see, for example, Millikan 1984, 2004; Neander 2013, 2018; Papineau 1987; Sterelny 1990). Such theories attempt to explain the content of mental representations in terms of the notion of biological function. Whatever one thinks of the prospects of success of such an endeavour, it is only indirectly related to the immediate question of giving an account of agency or action in terms of biological function, for even if one can explain the content of beliefs and desires by appeal to biological function, the problem of analysing teleological explanations of behaviour remains, and the objections explored in this section would still apply.

Further reading Aguilar, J. and Buckareff, A. (eds.) (2010) Causing Human Actions: New Perspectives on the Causal Theory of Action. Cambridge, MA: MIT Press. This is an edited collection of articles defending the causal theory of action, including a number of articles which touch on the problem of deviant causal chains. Allen, C. (2009) ‘Teleological Notions in Biology’, The Stanford Encyclopedia of Philosophy (Winter 2009 Edition), Edward N. Zalta (ed.), URL = . (Accessed 1 January 2019) This article is an overview of the role of seemingly teleological concepts (like function and design) in biology. Stout, R. (2010) ‘Deviant Causal Chains’ in O’Connor, T. and Sandis C. (eds.) A Companion to the Philosophy of Action. Oxford: Wiley Blackwell, pp. 159–165 This article gives an overview of the problem of deviant causal chains and tentatively proposes a suggested solution involving which sort of mechanism is involved in the event. Wilson, G. (1989) The Intentionality of Human Action. Stanford, CA: Stanford University Press. Wilson’s book contains a sustained argument against the causal theory of action and in favor of the irreducibility of teleological explanations of human action, with substantial discussion of the problem of deviant causal chains. Wright, L. (1976) Teleological Explanations. Berkeley, CA: University of California Press. This is the classic account of teleological explanation as a form of functional explanation.

References Davidson, D. (2001) Essays on Actions and Events. 2nd edn. Oxford: Clarendon Press. Fodor, J. (1987) Psychosemantics: The Problem of Meaning in the Philosophy of Mind. Cambridge, MA: MIT Press. McDonnell, N. (2015) ‘The Deviance in Deviant Causal Chains’, Thought: A Journal of Philosophy, 4: 162–170. Mele, A. (1992) Springs of Action: Understanding Intentional Behaviour. New York: Oxford University Press. Mele, A. (2007) ‘Action’ in Jackson, F. and Smith, M. (eds.) The Oxford Handbook of Contemporary Philosophy, Oxford: Oxford University Press: 334–357. Mele, A. (2013) ‘Actions, Explanations, and Causes’, in D’Oro, G. and Sandis, C. (eds.) Reasons and Causes: Causalism and Anti-Causalism in the Philosophy of Action, Basingstoke: Palgrave Macmillan: 160–174. Millikan, R. (1984) Language, Thought and Other Biological Categories. Cambridge, MA: MIT Press. Millikan, R. (2004) Varieties of Meaning. Cambridge, MA: MIT Press. Neander, K. (2013) ‘Toward an Informational Teleosemantics’, in Kingsbury, J., Ryder, D., and Williford, K. (eds.) Millikan and Her Critics, Oxford: Blackwell. Neander, K. (2018) ‘Teleological Theories of Mental Content’, The Stanford Encyclopedia of Philosophy (Spring 2018 Edition), Edward N. Zalta (ed.), URL = . (Accessed 1 January 2019) Papineau, D. (1987) Reality and Representation. Oxford: Basil Blackwell.

45

Scott Sehon Schlosser, M. (2007) ‘Basic Deviance Reconsidered’, Analysis, 67(3): 186–194 Sehon, S. (2016) Free Will and Action Explanation: A Non-Causal, Compatibilist Account. Oxford: Oxford University Press. Sterelny, K. (1990) The Representational Theory of Mind: An Introduction. Cambridge, MA: Blackwell Wilson, G. (1989) The Intentionality of Human Action. Stanford, CA: Stanford University Press. Yablo, S. (1992) ‘Mental Causation’, Philosophical Review, 101(2): 245–280.

46

3 AGENCY, EVENTS, AND PROCESSES Matthias Haase

1 Progression and completion God said, ‘Let there be light’, and there was light. That is not how it is for us.1 Nowadays, turning on the light is a thing a person might do. Given the right kind of contraption, saying those words can in fact be a way of making it so. Still, it is nothing like it, precisely because the making depends on all those other things. We don’t make the world; we act in the world. Usually, illuminating the room involves getting up, walking over, raising one’s arm, and turning the switch. For this to be a possibility, someone had to build the house, run the cables, and install the switches. And the people who did those things needed training, tools, and material to work on. We can’t create things from nothing. In our case, the making is a changing of given things or stuff. For this reason, it takes time and it takes place in a world where other forces are at work that might interfere before one gets it done. In English, this feature of the human condition is exhibited in the grammar of ordinary action sentences. Verb phrases like ‘to build a house’, ‘to turn on the light’, or ‘to get up’ are what the linguist call telic action verbs: they define a goal, a proper endpoint or terminus of the acting so characterized. When a telic action verb ‘ϕ’ is used to ascribe action on a particular occasion rather than a person’s general habits or policies, it is predicated of the respective subject S under the aspectual contrast. The imperfective ‘S is ϕ-ing’ ascribes progression towards the respective terminus; the perfective ‘S ϕ-ed’ ascribes the corresponding completion: the attainment of the goal internal to the concept of the act. The finitude of our agency is marked by the following two inference patterns. Completion requires progression: to get any of those things done, one must be doing them for some time. For any telic action verb ‘ϕ’, ‘S ϕ-ed’ implies ‘S was ϕ-ing’. The reverse doesn’t hold: where ‘ϕ’ is a telic action verb, ‘S was ϕ-ing’ doesn’t entail ‘S ϕ-ed’. It can be true that one was getting up, standing up for the light, but never got up, stood up, let alone turned on the light. Something might have interfered. Until it is true to say that one got it done, interruption always remains a possibility. Progression leaves completion open. Yet, we sometimes get things done; and the only way to do that is by doing them. Understanding the efficacy of human agency thus requires an account of the unity of progression and completion. Recently, it has been argued that it is impossible to accommodate the temporality of human agency within the so-called standard story according to which actions

DOI: 10.4324/9780429202131-547

Matthias Haase

are events. To fill the lacuna, it has been proposed to introduce processes into ontology. But the accounts of what processes are and how they are related to events vary. On closer inspection, it is not clear whether any of them can meet the above requirement. In what follows, I discuss the main contenders in the recent literature and challenge an assumption they appear to share and seem to inherit from the conceptual framework in which the standard story couched. As Donald Davidson famously has it, intentional actions are events with a special etiology: events caused (“in the right way”) by the relevant mental states: belief-desire pairs according to the initial proposal, intentions according to the considered view.2 Central to the doctrine is a strict distinction between two ways of talking of action or what a person does. A phrase like ‘street crossing’ can be used in a sentence to specify the kind of action a person performs on an occasion, and it can be deployed to refer to the person’s performance on that occasion. In Davidson’s terms: the “act-type” and the “particular action” (Davidson 1978, 77). The formal or intentional objects of wanting or intending are supposed be act-types. The events caused by them are said to be particulars that can be described in different ways and thus conceived as instances or tokens of those types. In the seminal paper that introduces the view, the focus is exclusively on actions that are over or done with. In fact, the completeness of the deed seems to play a crucial role in the articulation of the distinction just rehearsed. Davidson writes: “If I turned on the light, then I must have done it at a precise moment, in a particular way – every detail is fixed” (Davidson 1963, 6). That is what it means for there to be a “particular action”, and that is why it can’t figure as the formal object of wanting or intending. The want, Davidson explains, is not “directed to an action performed at any one moment or done in some unique manner”; it is something general insofar as “any one of an indefinitely large number of actions would satisfy [it]” (Davidson 1963, 6). This suggests that when Davidson speaks of an event, what he has in mind is the sort of thing that exists when “every detail is fixed”: something with fully determinate spatiotemporal properties. Officially, events are “unrepeatable particulars (‘concrete individuals’)” (Davidson 1970, 181). Their individuation is an intricate matter on which he had a change of mind.3 The same goes for the question whether action sentences like ‘I turned on the light’ contain singular reference to an event, as he originally suggests, or whether they should rather be conceived as existential quantifications over events (Davidson 1967). Finally, there is the famous renouncement of the original ambition to provide a non-circular or reductive explanation of the concept of intentional action (Davidson 1973, 80). But what all these changes leave untouched is the association between events and action sentences in the simple past. On the face of it, this appears to have a peculiar implication: for concrete action to be is for it be over and done with. It would seem that when every detail is fixed such that there is the event, the relevant acting isn’t going on anymore. But then how is it ever for the agent? In the paper devoted to intending, it is crucial to the dialectic that the central example is presented in the present progressive (Davidson 1978, 96). However, it is anything but obvious how such statements are to be understood in the framework of the proposed analysis of the logical form of action sentences. Given the theses, first, that descriptions of concrete acting are existential quantifications over events and, secondly, that events are particulars with fully determinate spatiotemporal properties, it threatens to follow that action can never be interrupted. Accordingly, ‘S was ϕ-ing’ would entail ‘S ϕ-ed’. In the literature on aspect, this is often called the imperfective paradox: the failure of a theory to explain why progression doesn’t imply completion.

48

Agency, events, and processes

Various proposals have been made for how to account for the temporality of human agency within the conceptual framework of the causal theory of action. However, it tends to be treated as a secondary phenomenon that can be explained by appeal to higher-order attitudes.4 A radical alternative where the temporality of human agency enters on the ground level seems to be suggested by G.E.M. Anscombe’s Intention. Here, actions also figure under the heading of events (Anscombe 1957, §16). But the focus is on the present progressive rather than the simple past. Anscombe’s thesis that an action is intentional under a description only if it is known to the agent under that description is always stated in terms of knowing what one is doing (Anscombe 1957, §17 and §48). The analogous requirement with respect to what one did would be patent nonsense. Obviously, it is possible to get something done and intentionally so without knowing of one’s success. Not surprisingly, Anscombe notes in this connection that concerning “a process or enterprise which it takes time to complete”, it is possible that the agent “was doing it, but did not do it” (Anscombe 1957, §23). How the latter remark stands to the talk of actions as events is not obvious.5 In his development of the Anscombian approach, Michael Thompson argues for a rigorous solution in which the reflection on the aspectual contrast takes centre stage.6 His point of departure is Anscombe’s observation that in ordinary parlance there is not “much to choose between ‘She is making tea’ and […] ‘She is going to make tea’” where the latter statement figures as the ascription of an intention (Anscombe 1957, §23). In Thompson’s final account, the main metaphysical contrast is not between what is allegedly merely in the mind and what is out there in the world, but rather between imperfection and perfection: the pursuit of an end and its achievement. Accordingly, it is a mistake to conceive wanting and intending as mental states. Properly conceived, ascriptions of planning, intending, wanting, or trying belong together with progressive action sentences to the “forms of the imperfective” and stand opposed to the perfective: the description of the done deed. However, the rigorous solution threatens to give rise to another puzzle. Alluding to Davidson’s doctrine that wanting and intending are general, whereas concrete action is particular, Thompson writes: “‘I am doing A’ is no more, or less, ‘general’ than ‘I intend to do A’ is; the transition to a genuine particular arises only with ‘I did A’” (Thompson 2008, 137). Accordingly, the progressive ‘S is/was ϕ-ing’ does not entail the existence of an unrepeatable instance or concrete individual of the respective act type; only the perfective ‘S ϕ-ed’ has such existential import. This can seem quite perplexing. Treating ascriptions of intention and action sentences in the progressive as equally “general” threatens to obscure what would seem to be a crucial metaphysical difference between those two “forms of the imperfective”.7 Intuitively, the former describe mere tendencies for movement or change, whereas the latter describe their manifestations in actual movement or change. While both may be subject to interference by other forces, only the second amount to progress on the ground. Unless the logical notation has the resources to capture this difference, it will be unintelligible how one can ever get anything done. After all, the intending alone surely won’t get you there. By analogy to the familiar talk of the imperfective paradox, the threatening implication may be called the progressive paradox: the failure of a theory to explain how progression can be sufficient for completion. One way to react to the difficulty is to insist that the distinction between abstract act-types and concrete instantiations or tokenings must also find application on the side of the progressive rather than being reserved for perfective, as Thompson seems to suggest. In other words, what we describe as ongoing or in progress must somehow have “actual presence in time and space”, as Jennifer Hornsby puts the point (Hornsby 2012, 3). Given that assumption, there are broadly speaking two options: either one takes the progressive and the correlated

49

Matthias Haase

perfective to entail the existence of one and the same underlying particular in different stages of development, or one holds that the progressive introduces the existence of another kind of temporal entity. In the former case, the contrast between progression and completion appears as the contrast between two properties. In the latter case, it appears as the contrast between two distinct kinds of entities. In the next two sections, I argue that each view leads into paradox. Then I venture a diagnosis of the common source of the troubles.

2 Two properties The two properties view (TPV) can be introduced in the following way. Davidson was right about that much: if an action sentence with perfective aspect like ‘Mary crossed the street’ is true, then there is an unrepeatable particular or concrete individual action. But it is a mistake to tie its existence to the truth of the perfective such that every detail is fixed. The truth of the progressive ‘Mary is crossing the street’ already introduces the existence of concrete action. Moreover, if we are ever to be in the position to say that Mary got done with what she was doing, then it must be one and the same concrete individual or unrepeatable particular action in both cases. After all, it won’t count as completion with respect to the one occasion, when Mary has success on some another occasion. Accordingly, the contrast between progression and completion is to be understood as the contrast between two mutually exclusive properties that the underlying particular can bear. For any telic action verb ‘ϕ’, the progressive ‘S is/was ϕ-ing’ characterizes the relevant underlying particular as in progress at the respective time; the corresponding perfective ‘S ϕ-ed’ characterizes it as complete or culminated. In the recent literature, TPV often figures in the context of an attack on the Davidsonian event ontology. Helen Steward and David Charles argue that making space for the thought that one and the same particular action can be in progress and eventually run to completion requires the introduction of the category of individual processes into ontology.8 But TPV has also been put forward in the framework of orthodox event ontology. Terence Parsons presents a version of the view as a straightforward extension of the Davidsonian analysis of the logical form of action sentences. In order to treat aspect within this framework, Parsons introduces two primitive predicates that apply to underlying events “at a time”. In his notation, they figure as first-order functions that take events and times as arguments: HOLD (e, t) signifies that “the event e is in progress (in its development portion) at t”; CUL (e, t) signifies that “the event e culminates at t”.9 Here, “events” appear in the place where Steward and Charles have “individual processes”. Accordingly, one can distinguish between two versions of the approach: TPV P and TPV E. Then again, with respect to our topic they might come to same. Whatever title one might prefer, the question is how one has to conceive of the respective underlying particular so that it can figure in the role assigned to it. Steward and Charles reserve the word ‘event’ for occurrents understood in the following way: they have a determinate extension in time and can be divided into its temporal parts, the sub-events of which they consist. Accordingly, the event of Mary’s street crossing is an unrepeatable particular with fully determinate spatiotemporal properties. Whether or not Mary reached the other side, every detail is fixed: the exact period it occupies in time, the exact places where Mary’s feet touched the street, the exact manner in which she moved them, and so forth. So conceived, an event can’t be interrupted or run to completion. For, it exists only if all its temporal parts exist. Given this definition of events, TPV will force the introduction of another kind of temporal entity to bear the relevant properties through which we are supposed to understand the contrast and relation between progression and completion.

50

Agency, events, and processes

As Steward has it, a concrete individual process is an unrepeatable particular: “a genuinely individual, token entity” (Steward 2013, 806). But of an irreducibly distinct kind. It differs from particulars in the order continuants (e.g., substances) in that it has temporal parts. It differs from perdurants (“events” in her terminology) in that its identity is not defined by the set of temporal parts it actually has at a time. Whereas events have their temporal parts “essentially”, concrete individual processes are “modally robust” (Steward 2013, 807). There has to be an initial segment with a determinate spatiotemporal location. But once the process of Mary’s street crossing has come into existence, it can exist in a range of possible worlds that differ with respect to where Mary places her next steps, how far she gets, and how long it takes her. This thought, Steward suggests, supplies the conceptual resources needed to make sense of progression, interruption, and completion: “As long as [it] continues, […] the very same token process is going on” (Steward 2013, 807). Due to its modal robustness, the determinate set of temporal parts it currently consists in can be conceived as an accidental property it has at a time. Its acquiring further temporal parts as it unfolds consequently appears as a “real change in the process itself ” – as a change of its “properties” (Steward 2012, 384). Now, given that the process is of a type that introduces a goal or terminus beyond which it cannot continue, the contrast between running to completion and being interrupted can then be defined in terms of it acquiring or failing to acquire all the temporal parts required for the truth of the perfective ‘Mary crossed the street’. Steward insists that, insofar as we are to speak of quantification here, the things quantified over in “an action sentence which has perfective aspect […] are individual processes and not events” (Steward 2012, 384). Even when the individual process is complete, its modal robustness still sets it apart from the correlated event which only exists when the “whole set of temporal parts has occurred”. In the spirit of the idea that the ontologist should only list the fundamental elements of being, Charles considers claiming instead that “events are simply processes […] seen under their perfective aspect” so that there is “no need for a further tier of entities over and above processes” (Charles 2018, 39). By the same token, Parsons might insist that one can do away with those individual processes, since they are simply events seen under their progressive aspect – in his terminology: “events in their development portions”. Whatever names one may prefer, what is really doing the work here is the positing of an underlying particular that is then brought under two predicates associated with the progressive and the perfective. How the approach is spelt out exactly varies depending on the other commitments of the theory at hand. Parsons’ articulation of the thought is governed by the requirements of providing a formal semantics for natural language. Depending on philosophical proclivities, such technicalities may or may not interest the metaphysician. But the beauty of a rigorous notation is that it sometimes captures the problematic implication of the philosophical approach. One of the attractions of TPV is that it seems to make space for interruption of actual progression, since it introduces a concrete entity that can be presently going on but might remain incomplete. Steward and Charles both stress this point (Steward 2013, 807; Charles 2018, 21). Similarly, Parsons praises his analysis for being “immune to ‘paradoxes’ of the imperfective kind, since saying of an event that it holds at a given time does not imply that it culminates at that or any other time” (Parsons 1990, 171). But the very formulation that states the purported solution gives rise to another puzzle. On the assumption that completion can be defined through a predicate that applies to the relevant underlying particular “at a time”, it should be possible to predicate it in the present tense. Accordingly, there seems to be logical space for present tense action sentences with perfective aspect. That is precisely what the

51

Matthias Haase

notation predicts. Parsons claims to find ordinary statements that fit the bill: ‘Mary arrives on the other side, crosses the finish line, and wins the race’ would be an example for what he calls the “reportative use” of the simple present (Parsons 1990, 30). Charles and Steward don’t go into the details here. But on closer inspection TPVp turns out to have the same implication as TPV E. Charles insists that a concrete individual process “exists without qualification at each moment during its career”, apparently including the time when it “draws to a close” (Charles 2018, 34). That the latter be included as a “moment during its career” is a consequence of the proposed definition. Given the doctrine that a concrete individual process is an unrepeatable particular that can be presently going on, has temporal parts, and acquires more of them as it unfolds, it seems natural to say that at time t1 the individual process token has a certain set of temporal parts and at t2 a larger set. That is how Steward describes what she calls the “growth” of an individual process (Steward 2012, 384). It follows that if we are to speak of completion, there must be a time tn when the individual process token of street crossing is still present while having all the temporal parts required for it to count as a complete street crossing. As Parsons himself notes, the “reportative use” of the simple present seems somewhat “odd”; strictly speaking, it can only be truly said “at one specific instant – the time the event culminates” (Parsons 1990, 30). On reflection, it is not just odd: the aspectual contrast doesn’t leave space for a present perfective. When it comes to such things as crossing the street on a particular occasion, the present is, as Bernard Comrie puts it, “essentially imperfective”.10 As long as street crossing is still going on, Mary is not yet on the other side. Once she is across, no street crossing is going on anymore – unless Mary turned around and is doing it again. For any telic action verb ‘ϕ’ and a specific occasion, the present progressive ‘S is ‘ϕ-ing’ entails the negation of the corresponding perfective ‘S ϕ-ed’, and vice versa. Progression and completion exclude one another. In the framework of TPV, this was supposed to be captured in terms of a pair of contrasting properties the underlying particular can bear. Introducing his terminology, Parsons said that one has to “choose between saying that the eventuality holds at the time in question, or saying that it culminates then” (Parsons 1990, 25). But the implied idea of present culmination would seem to suggest that the relevant underlying particular is still going on when it runs to completion. In Parsons’ terminology: “the hold time of an event extends through its time of culmination” (Parsons 1990, 306 FN 12). In either case, TPV forces positing the existence of what Gilbert Ryle once called that “perplexingly undetectable bit of crossing” in which the “arriving” appears to consist.11 The point of Ryle’s remark was of course to deny that there is such a thing. There is the moment when Mary is not yet across and is still crossing, and there is the moment when she is across and thus isn’t crossing anymore. This may be a matter of vagueness. But, in any case, there is no time intervening between the two and thus no moment to catch the arrival as presently happening. The very operation that was supposed to provide the solution to the imperfective and the progressive paradox leads to what one might call the perfective paradox: the failure of a theory to explain why completion excludes current progression and therefore can’t be in the present.

3 Two entities The two entities view (TEV) seems to avoid the troubling implication. Just as TPV, it includes the following two theses. Firstly, if the perfective ‘Mary crossed the street’ is true, then there is an unrepeatable particular or concrete individual action. Secondly, the truth of the corresponding progressive ‘Mary was crossing the street’ also requires the existence of 52

Agency, events, and processes

concrete acting. However, according to TEV, it is not one and the same underlying particular in both cases. Assuming so neglects the deep metaphysical difference between what the two kinds of action sentences describe. Properly conceived, the contrast and relation between progression and completion is the contrast and relation between two temporal entities of irreducibly distinct kinds. For any telic action verb ‘ϕ’, ‘S is/was ϕ-ing’ represents the one, and the ‘S ϕ-ed’ the other. Usually, the beings in question are called ‘processes’ and ‘events’, respectively. As Roland Stout has it, “processes are associated with the imperfective aspect and events are associated with the perfective aspect” (Stout 1997, 19). These “entities”, Stout argues, belong to “metaphysically distinct categories” or, for that matter, “different logical categories” (Stout 1997, 22). The relevant strand in the literature splits into two main branches depending on how that metaphysical distinction is articulated. According to the one version of the view, it is marked by the contrast between mass nouns and count nouns. Here, the entities associated with the progressive are not particulars, but rather “activities” that occupy time like stuff occupies space. The talk of particular actions is reserved for the perfective. According to the other variety of the approach, one can speak of concrete individuals in both cases; it is just that they belong to different categories. The details of an account along the former lines have recently been spelt out by Thomas Crowther and Jennifer Hornsby building on considerations by Alexander Mourelatos.12 The latter approach is developed in the work of Stout, a series of papers by Anthony Galton, and in an essay by Charles preceding the one mentioned above.13 Let’s call these the mass-individual version and individual-individual version of the approach: TEVM-I and TEV I-I. Here, the subscripts clearly mark a substantive disagreement. Still, with respect to our topic it might ultimately come to the same. Abstracting from the indices makes it easier to present the coming attractions and repulsions. According to the proponents of the view, the imperfective paradox cannot arise, since the progressive is associated with a distinct element whose existence at a time is not affected by its ceasing at a later time.14 And since completeness is reserved for the other element, the approach also appears immune to paradoxes of the perfective kind. As Hornsby puts it, the respective “event is on the scene” when the relevant “stretch of ongoing activity is over” and “no longer present”.15 But the very separation that promises salvation just gives rise to another set of the questions: What is the connection between those basic elements of temporal being such that we can speak of progression and completion? Why does the existence of the one require the prior existence of the other? How is that, unless something interferes, the existence of the latter leads to the existence of the former? And why, for that matter, does it ever count as interruption, if it doesn’t? Without answering these questions, the troubles with arrival won’t go away. There is still the issue with the “temporal boundaries”, as Crowther calls them: the “starts and stops of processes”. In the spirit of separation, he holds that such “change […] is not a time at which the [relevant activity] is present” (Crowther 2011, 20). So what about the starting and the stopping – or, for that matter, the finishing or arriving? It could hardly count as liberation from that troubling implication of TPV, if the alternative amounted to the peculiar doubling of the deed that Ryle presents as just another guise of the same confusion. After all, to describe a person as “having journeyed and arrived” is not to talk about “two things”, two distinct “acts, exertions, operations, or performances”; rather, it is to describe the person as “having done one thing with a certain upshot” (Ryle, 1949, 150). That better not be how we are to take the talk of two entities. Everything turns on the thesis that the entity associated with the perfective is composed or comprised of the entity associated with the progressive. The accounts of composition vary 53

Matthias Haase

depending on what figures in our subscripts. According to TEVM-I, it is analogous to the relation between a material object in space and the stuff of which it is made: “A complete ϕ event is a temporal particular made of ϕ-ing”, says Crowther.16 In the framework of TEV I-I, the composition appears as a relation between the two kinds of temporal particulars: “An event consist of a certain process going on for a certain period of time”, says Galton.17 The difficulty is the same on both sides of the divide. The initial account of composition is developed while leaving the talk of interruption and perfection aside. Roughly speaking, when what has been going on for a while comes to a hold and is thus not going on anymore, then there is something with a determinate extension in time. The latter is an event, the former the process or activity of which it is composed. The ideas of completion and progression towards it are supposed to be introduced in a second step – through a sub-specification within the general category of processes: directed activities or processes are said to be those temporal entities that have, so to speak, a built-in terminus. But how are we to understand that idea while holding on to the thesis that such entities can never be complete? If the perfective is to be reserved for the talk of events, then it has to be denied that ongoing processes have temporal parts at times. Otherwise, it would seem to be possible for such an entity to eventually acquire the relevant set of parts required for it to count as complete. That was the central thought of TPV P. In all varieties of TEV, the rejection of that thought is based on the formula that processes, by contrast to the events, can be “wholly present”.18 The division marked by our subscripts comes down to the two possible ways of spelling out that slogan. On the face of it, neither leaves space for the idea of progression, the unfolding of movement in time. According to TEV I-I, processes are, just much as events, proper individuals such that one “can make identity claims concerning members of this domain of quantification” or “demonstratively identify elements of this domain” (Stout 2016, 51). The proposed distinction between them turns on the thesis that whereas an event is extended in time such that only one of its parts can be present at any moment during the interval it occupies, a process is a temporal individual that is present as a whole at any moment of its existence. In that way, processes are said to be like continuants or a distinct type of continuants: they are not extended in time, but persist in time.19 Officially, this is supposed to explain why processes, in contrast to events, can be experienced directly. But it is hard to see what it could mean for Mary’s walking across the street to be given in a moment. We might look at a photograph or perhaps a series of snapshots, but neither can capture the continuity of her motion. What the relevant progressive describes must somehow be inherently durative; otherwise, there won’t be any actual progression towards the other side. One might try saying that the “present moment” is not an extensionless point but more like the so-called “experiential present”. But nothing in the proposed ontology explains that idea. Moreover, it wouldn’t make a difference unless actual progress could be made in that extended present. That, however, would threaten to reintroduce the idea of a temporal part. After all, the impossibility for a process to be present at a duration-less instant is Steward’s main reason for the thesis that processes have temporal parts.20 Hornsby and Crowther make the same objection against TEV I-I.21 But they resist Steward’s conclusion that leads back to TPV. Instead, they deny that processes are individuals. According to TEVM-I, activities or processes can be said to be “wholly present at a moment”, because what is present in that way is a homogenous mass. To be present at all, an activity has to be present for some time; but since each stretch within it is the same activity, the activity is wholly present at any moment at which it is going on. Unless what is going on comes to a halt such that there is a detached stretch of activity, there are no temporal individuals to count and thus no particular parts 54

Agency, events, and processes

either. The question to ask is ‘How much?’, not ‘How many?’ So for any action verb ‘ϕ’, the progressive ‘S is ϕ-ing’ should be read as: ‘There is some ϕ-ing by S’. What takes the place of the analogy between processes and continuants is the analogy between activity and stuff.22 However, that comparison seems equally misleading. As long as one sticks to the bare notion introduced via the analogy with space-occupying stuff, what is currently present as “activity” doesn’t reach ahead to any further continuation. Just as nothing in the snow on the roof can give you any reason to expect that mass to grow, nothing in that time-occupying stuff seems to point to the coming of more of it. But if that’s the model, the same should hold for what exists already. Accordingly, any stretch of ϕ-ing will be accidental to the one that came before, just as what is on the left must be deemed as accidental to what is on the right, if all I tell you is that there is some black paint on the wall. The analogy with a homogenous mass that is supposed to explain what it means for activity to be “wholly present” obscures the very structure among its elements through which we understand progression or unfolding in time. What figures as the paradigm for the proposed treatment of the progressive are so-called atelic action sentences like ‘Mary is walking’ where whether she is going to walk some more seems to be left open at any point. In Mourelatos’ paper that provides inspiration for TEVM-I, the telic progressive is introduced via the claim that “the generic activity of [walking] can be further differentiated into species”: say, walking a mile or walking across the street. The suggestion is that such differentiation introduces a specific quality while retaining the masslike “character as an activity” (Mourelatos 1978, 420). But this presupposes that the concept of walking is intelligible independently of the idea of walking somewhere. Of course, there is such a thing as taking a stroll. However, if there is to be walking, then there has to be the taking of steps; and unless there is order in the sequence there won’t even be any strolling, but only spinning around in one place. For the same reason, one cannot think of any actual walking in complete abstraction from the whence and the whither. Apprehended through the whence and the whither, the ongoing walking doesn’t appear as a homogenous mass; it exhibits an internal structure. The apprehension of its “presence” is in fact a temporal synthesis that collects the past, the steps that have been taken, as much as reaches ahead to the next steps to come.

4 Realization In both varieties of TEV, the separation that promises to solve the imperfective and the perfective paradox leads back into a version of the progressive paradox: the proposed definitions of the way of being that is supposed to be distinctive of the entity associated with progression ultimately exclude conceiving of that entity as unfolding towards completion. Without the idea that what is present reaches ahead to what is yet to come, it is equally mysterious why its ceasing to be present should count either as interruption or as completion. Similarly, the puzzle about present perfection that plagues the different varieties of TPV also infects the other notions: in consequence of the failure to explain why completion excludes current progression, the possibility of interference can’t be understood as intrinsic to being in progress. These concepts can only be understood together. Confusion about any one of them inevitably spreads to the other two. Our three paradoxes are just different guises of one and the same puzzle arising from the failure to account for the dynamic presence of action. On closer inspection, there is nothing dynamic about the entities TEV associates with the progressive. The same holds for the properties that take centre stage in TPV. In Parsons’ version of the view, the predicate HOLD (e, t) signifying the event being “in progress” is 55

Matthias Haase

“treated semantically as a state verb” (Parsons 1990, 170). At the same juncture, Steward defines what she calls the “growth” of a process (i.e., its progressing) in terms of the change of its properties: the set of temporal parts it has at a time (Steward 2012, 384). What replaces the picture of a persisting state of motion is thus the image of a series of states that each holds for a moment. For the purposes of understanding the continuity of movement, the one seems just as inadequate as the other. A peculiar feature common to TPV and TEV is that they presuppose the concept of change. One would think that this notion belongs to the concepts we are trying to understand when inquiring into the metaphysics of events and processes. In both approaches, however, the relation between progress and completion is defined in terms of change: in the one case as change in the properties of an underlying particular, in other case as a change with respect to the entities that can be said to exist. What forces the choice between these two options is an assumption that is taken for granted in all varieties of those views–namely, that the metaphysics of action is to be approached through the binary contrast between type and token.23 So conceived, instantiation or realization is not itself a dynamic category. Consequently, the dynamic character of the phenomenon at hand has to be somehow introduced in a second step: either in terms of the changing properties of the item that is taken to exist right from beginning or in terms of the item’s relation to another entity that is taken to come into being at a later time. But why accept the assumption that makes those alternatives seem exhaustive? Intentional action is the realization of action concepts. Accordingly, it seems reasonable to wonder whether the relevant notion of realization or instantiation can be captured in the conceptual framework of the distinction between type and token, the abstract and its instances. On closer inspection, it turns out that the real target of Thompson’s rejection of Davidson’s way of contrasting intention and action as general and particular is precisely the underlying type-token model of instantiation. The clearest expression of the point is hidden in corresponding footnote where Thompson remarks that action concepts “do something more than classify individual events”.24 Another way of putting it is to say that the contrast and relation between progression and completion can’t be captured through those nominalization transcriptions introducing the type-token model and the correlated classifications of entities, whether they are called ‘events’ or ‘processes’. Rather, it is to be understood through our ordinary action sentences in which the agent figures as subject. Thompson’s puzzling remark in the main text suggesting that ongoing ϕ-ing is just as “general” as the intention to ϕ threatens to distort his own point. The question how this point is to be developed into a full positive account I leave to another occasion.25

Related topics Diachronic agency; Agency and causation.

Notes 1 The following is based on considerations from the manuscript “The Existence of Action” that Douglas Lavin and I have been working on for years. All the mistakes in this completed bit of work are mine. 2 See Davidson (1963, 1978). 3 According to the original account, the identity of events is defined in terms of their causal relations to other events (Davidson 1969, 179). Later he considers Quine’s suggestion to individuate events by their spatiotemporal location (Davidson 1985, 309).

56

Agency, events, and processes 4 According to Michael Bratman, for instance, understanding “temporally extended agency” requires the introduction of “plans”: higher-order attitudes through which “we achieve the organization and coordination of our activities over time” (Bratman 2000, 46). This presupposes the intelligibility of a primitive notion of realizing ends that doesn’t require the reflection on the temporal structure of action. In this connection, Bratman speaks of “time-slice agents” (Bratman 2000, 43). In other theories, the idea of “basic action” plays an analogous role. For a challenge of this kind of approach, see Lavin (2013). 5 In a later paper Anscombe agrees with Davidson that the talk of “under a description” requires conceiving of actions as “events” and that events are to be treated as “individuals”. At the same time, she rejects his “theory of event identity” and denies the need for such an account (Anscombe 1979, 219). But since that paper focuses on action sentences with perfective aspect, it seems to be a further question what this means for the imperfective. 6 See Thompson 2008. As he points out, the aspectual contrast played a central role in the classifications of action verbs that can be found in the work of Gilbert Ryle, Anthony Kenny, and Zeno Vendler. See Ryle (1949), Vendler (1957), Kenny (1963). For a general treatment of aspect, see Comrie (1976) and Galton (1984). 7 For the discussion of the implication for the account of practical knowledge, see Haase (2018). 8 Steward (2012; 2013; 2016) and Charles (2018). 9 Parsons (1990, 25). 10 Comrie (1976, 66). See also Galton (1984, 47) and Thompson (2008, 125). 11 Ryle (1949, 152). See also Galton (1984, 61) and Anscombe (1964, 194). 12 See Mourelatos (1978), Crowther (2011; 2018), Hornsby (2012; 2013). 13 See Stout (1997; 2003; 2016), Galton (2006; 2008), Charles (2015). 14 See Stout (1997, 21), Galton (2008, 337), Crowther (2011, 32), Hornsby (2012, 241). 15 Hornsby (2013, 9). For analogous remarks, see Galton (2006, 10), Stout (2016, 53). 16 Crowther (2011, 24). See also Hornsby (2012, 235). 17 Galton (2008, 327). See also Galton (2006, 6), Stout (1997, 22). 18 See Mourelatos (1978, 416), Hornsby (2013, 237), Stout (1997, 25), Galton (2008, 329). 19 See Stout (1997, 23), Galton (2006, 9), Charles (2015, 204). 20 See Steward (2013, 12). See also Charles (2018, 30). 21 See Hornsby (2012, 243 fn 6), Crowther (2018, 78). 22 See Mourelatos (1978, 424), Crowther (2011, 15), Hornsby (2012, 238). 23 See Stout (1997, 21); Galton (2008, 327); Hornsby (2013, 3); Crowthers (2018, 75). 24 Thompson (2008, fn 21, 137–138). 25 I am very grateful to Luca Ferrero, Douglas Lavin, Sarah Paul, and Sergio Tenenbaum for comments on an earlier draft of this paper.

Further reading Galton, Anthony (2017) “Dynamic Present”, in P. Hasle, P. Blackburn and P. Øhrstrøm (eds.), Logic and Philosophy of Time: Themes from Prior, Volume 1, Aalborg: Aalborg University Press: 167–187. The paper argues against so called “at-at” theories of change according to which this concept can be explained in terms of objects’ possessing different properties at different times. According to the proposed alternative, accounting for continuity of change and motion requires treating the present as an interval whose contents are inherently dynamic in nature. Rödl, Sebastian (2012) Categories of the Temporal: An Inquiry into the Forms of the Finite Intellect, Cambridge, MA: Harvard University Press. The book argues that, due to its Fregean heritage, analytic philosophy fails to account for the relation of judgement and truth to time. According to the proposed alternative, tempus and aspect are to be treated as logical categories that belong to the very idea judgement. Szabó, Zoltán Gendler (2008) “Things in Progress”, Philosophical Perspectives, 22 (1): 499–525. The paper argues against intensional analyses of the progressive and proposes an account of the unity of the progressive and the perfective that differs from the ones discussed above. Waterlow, Sarah (1984) Nature, Change, and Agency in Aristotle’s Physics: A Philosophical Study, Oxford: Oxford University Press.

57

Matthias Haase The book is a seminal study of Aristotle’s account of change and its relation to the power of agency. It argues that change, as Aristotle conceives it, possesses a unique metaphysical structure, and it suggests that this structure tends to be obscured in the logical frameworks of contemporary approaches to ontology.

References Anscombe, G.E.M. (1957) Intention. Oxford: Basil Blackwell. ——— (1964) “Before and After”, in Anscombe 1981: 180–195 ——— (1979) “Under a Description”, in Anscombe 1981: 208–219. ——— (1981) Metaphysics and the Philosophy of Mind: Collected Philosophical Papers Vol. III, Oxford: Basil Blackwell, 1981. Bratman, M. (2000) “Reflection, Planning, and Temporally Extended Agency”, The Philosophical Review, 109 (1): 35–61 Comrie, B. (1976) Aspect. Cambridge University Press. Charles, D. (2015) “Aristotle’s Processes”, in M. Leunissen (ed.), Aristotle’s Physics: A Critical Guide, Cambridge: Cambridge University Press: 186–205. ——— (2018) “Processes, Activities and Actions”, in R. Stout (ed.), Process, Action, and Experience, Oxford: Oxford University Press: 20–40. Crowther, T. (2011) “The Matter of Events”, The Review of Metaphysics. 65: 3–39. ——— (2018) “Processes as Continuants and Processes as Stuff”, in R. Stout (ed.), Process, Action, and Experience, Oxford: Oxford University Press: 58–81. Davidson, D. (1963) “Actions, Reasons, and Causes”, in Davidson 2001: 3–19. ——— (1967) “The Logical Form of Action Sentences”, in Davidson 2001: 105–122. ——— (1969) “The Individuation of Events”, in Davidson 2001: 163–180. ——— (1970) “Events as Particulars”, in Davidson 2001: 181–187. ——— (1978) “Intending”, in Davidson 2001: 83–102. ——— (1985) “Reply to Quine on Events”, in Davidson 2001: 305–311. ——— (2001) Essays on Actions and Events, Second Edition. Oxford: Clarendon Press. Galton, A. (1984) The Logic of Aspect, Oxford: Clarendon Press. ——— (2006) “On What Goes On: The Ontology of Processes and Events”, in Bennett, B. and C. Fellbaum, (eds.), Formal Ontology in Information Systems, Amsterdam: IOS Press: 4–11. ——— (2008) “Experience and History: Processes and Their Relation to Events”, Journal of Logic and Computation, 18: 323–340. Haase, M. (2018) “Knowing What I Have Done”, Manuscrito, 41 (4): 195–253. Hornsby, J. (2012) “Actions and Activity”, Philosophical Issues. A Supplement to Noûs. Action Theory: 233–245. ——— (2013) “Basic Activity”, Proceedings of the Aristotelian Society Supplementary Volume, 87: 1–18. Kenny, A. (1963) Action, Emotion and Will. London: Routledge. Lavin, D. (2013) “Must There Be Basic Action?” Noûs, 47 (2): 273–301. Mourelatos, A. (1978) “Events, Processes and States”, Linguistics in Philosophy, 2: 415–34. Parsons, T. (1990) Events in the Semantics of English: A Study of Subatomic Semantics. Cambridge, MA: MIT Press. Steward H. (2012) “Actions as Processes”, Philosophical Perspectives, 26: 373–388. ——— (2013) “Processes, Continuants, and Individuals”, Mind, 122: 781–812. ——— (2016) “Making the Agent Reappear. How Processes Might Help”, in R. Altshuler and M.J. Sigrist (eds.), Time and The Philosophy of Action, New York: Routledge, 67–83. Stout, R. (1997) “Processes”, Philosophy, 72: 9–27. ——— (2003) “The Life of a Process”, in Debrock, G. (ed.), Process Pragmatism: Essays on a Quiet Philosophical Revolution, Amsterdam: Rodopi: 145–57. ——— (2016) “The Category of Occurrent Continuants”, Mind, 125: 41–62. Ryle, G. (1949) The Concept of Mind, Chicago: University of Chicago Press. Thompson, M. (2008) Life and Action, Cambridge, MA: Harvard University Press. Vendler, Z. (1957) “Verbs and Times”, Philosophical Review, 66: 143–160.

58

4 NEGATIVE AGENCY Randolph Clarke

We perform actions of various kinds, and our doing so is an aspect of our agency. We also sometimes omit to act (to turn out the lights when we leave) or refrain from acting (from reaching for a chocolate), and these phenomena, too, are agential in important respects. Just as one can intentionally perform a certain action, acting for certain reasons when one does, so one can intentionally not do a certain thing, refraining for reasons from doing that thing. And just as one can be morally responsible for an action and for things that result from it, so one can be morally responsible for omitting or refraining and for things that result from it. A comprehensive theory of agency should encompass omitting and refraining as well as acting. But there is reason to doubt that a theory can uniformly cover all of these phenomena. To make the point, I focus here on widely held causal theories of action. Some of these take an action to count as such because of how and by what it is caused, others because the action is itself a causing of a certain kind. Some take the causation in question to be causation by events or states, others by agents.1 Despite the variety, causal theories commonly hold that what is caused, when one acts, is some event or state, usually including (if not entirely consisting of ) some motion or position of one’s body (or part of one’s body). For example, when I type a sentence, my action might be a certain movement by me of my fingers, consisting of certain motions caused in a certain way or consisting of a causing of such motions. The motions are said to be the action, or to partly constitute it, because of satisfaction of the causal requirement. But when I refrain from typing a sentence, my doing so is, it seems, partly a matter of there not existing any such movement by me. To apply a causal theory here, it appears, we would have to say that what is caused is an absence of a motion, the non-occurrence of a certain kind of event. And absences or non-occurrences, if they are things of any kind at all, seem to be rather different from bodily motions. Similarly, when I am responsible for results of my actions, commonly these are causal effects of my actions. In contrast, a result of my omitting to perform an action—to turn out the lights when I leave my office, say—seems to stem from an absence of any such action by me on that occasion. If we have causation here, it appears to be causation by a non-occurrence. The ontological category of the cause, it seems, differs from that of act particulars—it might be said to be that of merely possible things,2 or of abstract entities,3 or of beings that are in

DOI: 10.4324/9780429202131-659

Randolph Clarke

some way ontically negative4 —or we might hold that we have causation here even though there is nothing at all that is the cause (as Lewis 2004 suggests).5 Despite these appearances, some theorists maintain that a causal theory can uniformly apply to acting, omitting, and refraining. For, they claim, each omission and each instance of refraining can be identified with a garden-variety action. I argue here that this strategy often fails. We can nevertheless provide sufficient conditions for omitting or refraining without committing to exotic entities or questionable causal claims. The account of agency we are left with is less uniform than what might have been preferred, but diversity in the subject matter might preclude any more uniform theory.

1 Actions negatively described There are cases of refraining in which an agent’s not doing a certain thing is arguably identical with the agent’s performing an action, one that can be understood in terms of a causal theory. A mime stands on a sidewalk, frozen in mid-gesture, for many minutes refraining from noticeably moving. Maintaining the pose requires exerting effort, making fine adjustments to muscle tension in response to careful self-monitoring. The mime is active in doing this; with a causal theory assumed, we may take her act of holding still to be her continuing body posture, appropriately caused, or to be the causing of that posture. What is the mime’s intentionally not moving, her refraining from moving? There is a good case for saying that this is simply her act of holding the pose, now negatively described, in terms of something it is not.6 The refraining takes place when and where the act does, it is just as static and as difficult as the act is, and it has that act’s causal features—caused by all and only the things that cause the act and causing all and only the things that the act causes. In this case, commitment to a causal theory requires neither an admission of exotic kinds of being nor the acceptance of causation where there is nothing that is the causal effect. Causalists might hope that all omissions can be treated in this way, and some claim that they can. In defending ‘the standard story of action’—largely following Donald Davidson (1980a, 1980b)—Michael Smith maintains that each action is an event particular, a bodily movement, with a certain causal etiology. This claim ‘has to be interpreted so that any orientation of the body counts as a bodily movement’ (2010: 45–46); keeping still can count.7 Such a theory can provide ‘a story about actions in a quite general sense in which the distinction between actions and omissions is invisible’ (48). A causal theory of omissions, then, is a causal account of certain bodily motions that count as actions—and count as such for the same reason that any action counts as such (see also Payton 2018: 87).8 The appeal of identification as a general strategy is obvious. It promises a comprehensive theory of agency on which each action and omission is the same kind of thing—an event particular—something of a familiar ontological category and not ontically negative (although it may be described negatively). All such events count as agential for the same reason: they all satisfy, in the same way, the same causal requirement.

2 Unwitting omissions Before addressing Smith’s argument, consider cases of unintentional or unwitting omission. Sometimes one intends to do a certain thing at a certain time—stop and get milk on the way home, say—but forgets to do it. Or one has a policy of doing a certain kind of thing in certain circumstances—leave the pool pump running on nights when the temperature dips below freezing—but on some occasion one fails to follow it because one forgets or fails to 60

Negative agency

notice that the relevant circumstances obtain. Sometimes there is something that one ought to do—send a thank you note—but one does not so much as think to do it. In all such cases, during the time when the thing in question is meant to be done, or when one ought to be doing it, one isn’t aware that one is omitting to do it. These are ‘unwitting omissions.’ Unwitting omissions are failures to act. It is implausible that each such failure is itself an action. Further, at the time when the action was meant to be done, or ought to have been done, one might be performing no action at all; one might be sound asleep. Here, the strategy of identifying each omission with an action seems misguided. An identification theorist might reply by holding that each omission to act is (identical with) a bodily movement (broadly construed), without claiming further that each such movement is an action (Payton 2018). On this view, some unwitting omissions are said to be actions, and some not to be; it depends on what the agent is doing at the time at which she intended to be, or ought to be, doing what she omits. Indeed, some unwitting omissions will be said to have parts that are actions and parts that are not. This variant of the identification strategy fails to deliver what seemed most promising about the approach. Although unwitting omissions are identified with bodily movements, they are not agentive by satisfying a causal requirement, for they need not satisfy any such requirement. They are agentive only because they occur when the agents are not doing what they meant to do or ought to have done. Further, things that result from an omission need not be caused by bodily movements of the agent at the time when the omitted action was meant to be, or ought to have been, performed. Imagine a patient who died because a life-saving injection was not administered. The doctor who was supposed to give the injection was instead, at the crucial time, dining at a restaurant. There will be no counterfactual dependence of the death on the dining if, had the doctor’s act of dining not occurred, an act of her dancing would have. Indeed (to adapt a point from Dowe 2009: 32), it is possible for the doctor to have been sufficiently distant from the patient at the crucial time that the causal speed limit precludes causal influence on the death by her movements then. We might still count the death as a result of her omission. What are unwitting omissions, if we reject the identification strategy? We might hold that unwitting omissions are things of some sort, though not garden-variety bodily movements. They might be said to be mere possibilia, or abstracta, or ontically negative concrete entities. Alternatively, an unwitting omission might be said to be nothing at all (more carefully: although agents sometimes unwittingly omit to do this or that, there is no entity that is an unwitting omission). Compatibly with this last option, the following conditions appear to suffice for it being the case that an agent S unwittingly omits to A at time t: (i) S ought to A at t; (ii) S has the ability and opportunity to A at t; (iii) S believes that she can A at t; (iv) S does not A at t; and (v) S is not aware that she does not A at t. In this formulation, the relevant time t at which S omits to A is a time when S ought to A. It is not the time when something that is an omission by S occurs; there need be no such thing, and thus no time, at which such a thing occurs. Cases of unwitting omission are varied, and in some, satisfaction of different sets of conditions suffices. For instance, one can unwittingly omit to do what one intended to do but 61

Randolph Clarke

wasn’t required to do.9 Still, the conditions seem to be jointly sufficient. By disjoining several sets of jointly sufficient conditions, we might state a condition both necessary and sufficient for unwitting omission. Our account would then be disjunctive, as would a comprehensive theory of agency incorporating it. But it remains to be seen whether taking unwitting omissions to be entities of some kind can yield greater uniformity.

3 Refraining Even if the standard causal story is not applicable to unwitting omissions, it might be thought to be straightforwardly applicable to intentionally omitting or refraining. Smith, for one, argues so. Given a capacious notion on which keeping still can count as bodily movement, an intentional omission, he maintains, is a bodily movement with respect to which the agent has a certain know-how, and which is caused in the right way and rationalized by a certain desire-belief pair (2010: 45). For example, when an agent S decides not to take a chocolate and refrains from moving her arm toward the chocolate box, her act of refraining is ‘a bodily movement that she knows how to perform, where her knowledge how to perform that bodily movement is not explained by her knowledge how to do something else’ (48–49), which movement has the indicated causal etiology. Smith speaks of whatever actual movement of S’s arm occurs as the way in which S refrains; the agent, he says, ‘has no alternative but to refrain from moving her arm in one of [the available] ways…. For the only way that [S] can keep her arm from moving toward the chocolate box is by ensuring that it is somewhere else’ (2010: 49). Imagine that S’s arms were hanging freely at her sides before she made her decision, and they remain hanging there freely when she decides, and for some time thereafter. Likewise, the chocolates were sitting in the box before and remain there after the decision. Gravitational forces, with no help needed from S’s will, might suffice to keep both her arms and the chocolates in place. In both cases, all that S has to do is not do something (move her arms, move a chocolate). No act of keeping her arms from moving toward the chocolates is then required, any more than is an act of keeping the chocolates from moving toward her.10 Smith says that such an agent ‘clearly does exercise control over the way her body moves, because she makes sure that it doesn’t move toward the chocolate box’ (2010: 49). About a similar case, he says that the orientation of the agent’s body ‘is under [her] control in the sense of being sensitive to what [the agent] desire[s] and believe[s]’ (46). I agree with the latter claim. But the same can be said about the position of the chocolates. S can move them; their position is under her control, sensitive to what she desires and believes. But neither S nor any of her desires or beliefs are causing the chocolates’ remaining where they are. Their position is under her control even though she is not controlling it. The same can be said about motions of parts of our bodies. The rate of one’s breathing can be under one’s control, in that one can speed or slow it at will. The rate can be sensitive to whether one desires to alter it. But most of the time, one is not controlling the rate of one’s breathing. So it might be with the position of S’s arms when she refrains from taking a chocolate. It might be suggested that S’s decision not to reach for a chocolate is the action that is her refraining. But deciding not to do something isn’t refraining from doing it; having made the decision, one might change one’s mind or become distracted, failing to carry out the decision. And in any case, one can refrain from A-ing without having decided not to A. We make decisions when we face uncertainty about what to do; but intentions are sometimes acquired without any such uncertainty, as when one reaches for one’s keys on arriving at 62

Negative agency

one’s office door in the morning. Intentions not to do this or that can be similarly acquired, as when one sees a snake on the path ahead and immediately comes to intend not to take another step. One might then intentionally not take another step, without having had to decide not to do so. Some cases are amenable to Smith’s approach. In what Bruce Vermazen (1985: 103) calls ‘displacement refraining,’ an agent performs one action to prevent herself from performing another. If S had doubted her self-control, she might have shoved her hands into her pockets or grasped them tightly behind her back, to keep herself from reaching for a chocolate. It is crucial that in this kind of case, the agent conceives of the action she performs as something she does in order to prevent herself from performing the action from which she refrains. But intentionally not doing a certain thing, or refraining, does not always take this form. One does not always, when one refrains from A-ing, perform any action at all as a means to not A-ing. When one does not, although one exercises one’s capacity to do things intentionally, it is not clear that there need be anything at all that is an act of refraining. Suppose that on some occasion all of the following are true: S has the ability and opportunity to A at t; S believes that she can A at t; shortly before and at t, S intends not to A at t; S’s so intending prevents her A-ing at t without undermining her control over whether she A-s then; and (v) S is aware at t that she is not A-ing then.

(i) (ii) (iii) (iv)

Consider whether we have here sufficient conditions for it being the case that S refrains from A-ing at t, that she intentionally does not A then. Prevention can be understood as what Phil Dowe (2001) calls quasi-causation of a non-occurrence. We have it if one’s coming to intend not to A causally interacts with an ongoing process—deliberation, for example—that, were it not for this interaction, would have culminated in one’s A-ing, and one does not then A. Such prevention could undermine one’s control over whether one A-s, as it would if, for example, it immediately and unexpectedly caused an episode of narcolepsy. But the prevention need not (and would not ordinarily) do this. One’s control over whether one A-s is retained if one remains able to change one’s mind and A at the relevant time. Again, it should be recognized that refraining is varied, and in some cases satisfaction of different sets of conditions will suffice. Commonly when we refrain from doing something, there is no process underway that will culminate in our doing that thing unless we come to intend not to. In many of these cases, one’s intention still plays an inhibitory role, diminishing motivation, preventing thoughts about doing the thing in question, or turning one’s attention to alternatives or to other matters altogether.11 Further, there are cases of intentionally not A-ing in which the relevant intention of the agent is not an intention not to A. (Reaching a fork in a path, I might intend to take the shorter left path and intentionally do so, realizing when I do that I am foregoing the pleasures of the scenery from the right path. It might then correctly be said that I intentionally do not take the right path.) And it is not necessary for intentionally not A-ing at t that one is aware at t that one is not A-ing then. (One can intentionally not attend a conference held on certain dates without being aware on those dates that one is not at the conference. Having earlier foregone making any arrangements to go, having foregone traveling to the conference location, and so forth, there is no need for one to be aware at the time of the meeting that one is skipping it.) But the goal 63

Randolph Clarke

was to identify sufficient conditions. If we’ve done that, we’ve done so without commitment to there being, when one intentionally omits or refrains, an action—or, indeed, anything at all—that is one’s intentional omission or one’s refraining.12 It is worth reflecting briefly on a more complicated case. The instances of refraining considered so far are episodic: there is some relatively brief interval of time that is the relevant time during which one intentionally does not A. Other instances are periodic: one refrains for years, or one resolves never again to do a certain thing, and one carries out that intention. For example, I’ve refrained from smoking for the past 20 years. Although my resolution not to smoke, when first formed, might have played an inhibitory role, and might have done the same on occasion since then, there is little that it needs to do nowadays—I’m never tempted—and nothing at all is needed from it during many stretches of time (when I’m asleep, for example). Still, it isn’t that I’ve refrained only on certain occasions when, say, cigarettes were on offer. If credit is due, then I get it for the whole period. It seems to me even less plausible in this kind of case (in comparison with episodic refraining) that there exists some concrete particular—some spatiotemporally located thing—or some fusion or set of such things that is my refraining for the past 20 years. Further, I find it unclear just what kind of causal story, if any, must be told about this kind of case. If there is one, it seems unlikely to follow closely the lines of a causal theory of action.

4 Differences Smith maintains that ‘the standard story aims to explain agency quite generally’ (2010: 48), encompassing omission as well as action. Taking such an aim for granted, some critics argue, against the standard story, that cases of omitting or refraining present stumbling blocks to, or undermine, such a theory.13 I suggest that the starting point of this debate is mistaken. Agency is diverse. A causal thesis might apply in one way to some agential phenomena, in another way to others, and perhaps not at all to yet others. It is, in the first instance, a thesis about action. But there are agential phenomena that are not actions. In cases of unwitting omission, there need be no action that the agent performs that is her omission. This might be so even in some cases of intentional omission or refraining. While the mime’s holding still, or an instance of displacement refraining, is an action, it is not clear that there need be any action that is one’s refraining when one refrains from reaching for a chocolate. Nor does it seem that there must be some action (or string of actions) that one performs that is one’s refraining when one refrains from smoking for 20 years. The phenomena, though not uniform, nevertheless form a decently coherent subject matter. Reasons for action bear on omitting and refraining as well as on acting. An episode of any of these can reflect an agent’s practical identity. Each can be attributable to one as an agent, justifying ethical evaluation, reactive attitudes, and overt responses, as can their results. We have a coherent practical interest in this variety of things. There are ontological and causal differences among these phenomena. Certainly we should strive to achieve theoretical unity where we can, but we should also recognize difference where it exists. A theory lacking a simple uniformity can nevertheless be illuminating, and that may be all that we can reasonably hope for in a comprehensive theory of agency.14

Related topics Agency and causation; Agency, events, and processes; Intentional agency; Agency and responsibility. 64

Negative agency

Notes 1 Theorists who hold that the things that are actions count as such because of how and by what they are caused include Brand (1984), Davidson (1980a), Enç (2003), and Goldman (1970). Those taking actions to be causings (or bringings about) include Alvarez and Hyman (1998), Bach (1980), Dretske (1988: ch. 1), O’Connor (2000: ch. 3), and Steward (2012: 45). Although many in the latter group appeal to irreducible agent causation, Bach and Dretske do not. 2 Bernstein maintains that omissions are non-actual possible events, although, as she sometimes states the view, it is that ‘an omission is a tripartite metaphysical entity comprised of an event at a possible world, an event at the actual world, and a counterpart relation between them’ (2014: 6). 3 The cause might be said to be a fact—the fact that I didn’t turn out the lights—with facts construed as what are stated by true sentences, statements, or propositions. Mellor (1995) takes all causes and effects to be facts of this kind. 4 Omissions might be seen as concrete negative states of affairs, where these are construed as instantiations of negative properties, as anti-instantiations of properties (Barker and Jago 2012; Brownstein 1973; Hochberg 1969), or as states of affairs with negative valence (Beall 2000; Priest 2000). 5 A similar problem arises from the fact that actions (e.g., blocking relief convoys) can result in absences of things (food and medicine). Here, if we have causation, at first blush the effect (if it is anything at all) seems to be something other than an ordinary event. 6 Objects are commonly described in terms of properties that they lack; e.g., Eris may be said to be the most massive non-planet directly orbiting the Sun. So, the thought is, an action can be described as an instance of intentionally not moving. Here I agree with Davidson (1985), Schaffer (2012), and Varzi (2006, 2008), and disagree with Moore (2009: 438). 7 Davidson himself suggested that, for the purposes of his theory, the notion of a bodily movement must be generous enough ‘to encompass such “movements” as standing fast, and mental acts like deciding and computing’ (1980b: 49). 8 Davidson (1985) endorsed the identification of omissions with bodily movements in a limited range of cases, as have Clarke (2014: 21–28), Varzi (2006, 2008), and Vermazen (1985). Schaffer (2005) advances the identification strategy for absences more generally. 9 Condition (iii) might be more than is required. Perhaps lacking a belief that one cannot perform the action in question is enough. For discussion of a doxastic requirement on omission, see Clarke (2014: 94–95). 10 As the case may be imagined, it is strikingly different from that of the mime: no continuing effort by S, no ongoing adjustments to muscle tension in response to self-monitoring, produce the continuing position of her arms; none are needed. 11 Shepherd (2014: 22–23) discusses similar roles of intentions not to do certain things. 12 Sartorio (2009) suggests that it suffices for intentionally omitting to A that an omission to A be caused by an intentional omission to intend to A. An omission to intend to A is intentional, she says, if one ‘voluntarily failed to form that intention, after deliberating about whether to do so, after considering reasons for and against doing so, etc.’ (523). However, failing to intend to A after deliberating about whether to do so doesn’t suffice to make the omission to intend intentional (one  might simply have become distracted), and it is not explained what more might make an omission to intend voluntary. Further, as has been argued, one can intentionally omit to A without having deliberated about whether to A. For more on this point, see Clarke (2014: 69–70) and Shepherd (2014: 17–18). 13 See, e.g., Hornsby (2004, 2010) and Sartorio (2009). 14 For comments on a precursor to this chapter, thanks to Andrei Buckareff, Jonathan Payton, and participants at the workshop on Theories of Causation: Mental Causation, Negative Causation, and Other Challenges, Düsseldorf Center for Logic and Philosophy of Science, July 2017.

Further reading Bach, K. 2010. ‘Refraining, Omitting, and Negative Acts.’ In T. O’Connor and C. Sandis, eds., A Companion to the Philosophy of Action, 50–57. Oxford: Wiley-Blackwell. Bach surveys the phenomena of omitting and refraining and argues that there are no actions that are negative in any interesting sense.

65

Randolph Clarke Clarke, R. 2014. Omissions: Agency, Metaphysics, and Ethics. New York: Oxford University Press. This monograph on omitting and refraining examines these phenomena from the perspectives of action theory, metaphysics, and moral responsibility. Fischer, J. M. and Ravizza, M. 1998. Responsibility and Control: A Theory of Moral Responsibility. Cambridge: Cambridge University Press. Chapter 5 of this book advances a detailed account of responsibility for omissions, addressing as well some questions about the metaphysics of omission. Payton, J. D. 2021. Negative Actions: Events, Absences, and the Metaphysics of Agency. Cambridge: Cambridge University Press. Payton argues that intentional omissions and instances of refraining are token-identical with ordinary, ‘positive’ events. Sartorio, C. 2009. ‘Omissions and Causalism.’ Noûs 43: 513–30. Sartorio argues that a standard causal theory of action cannot be extended to cases of intentionally not doing a certain thing, for in these cases the absence of action is caused not by an intention to omit but by an omission to intend.

References Alvarez, Maria and John Hyman. 1998. ‘Agents and Their Actions.’ Philosophy 73: 219–245. Bach, Kent. 1980. ‘Actions Are Not Events.’ Mind 89: 114–120. Barker, Stephen and Mark Jago. 2012. ‘Being Positive about Negative Facts.’ Philosophy and Phenomenological Research 85: 117–138. Beall, J. C. 2000. ‘On Truthmakers for Negative Truths.’ Australasian Journal of Philosophy 78: 264–268. Bernstein, Sara. 2014. ‘Omissions as Possibilities.’ Philosophical Studies 167: 1–23. Brand, Myles. 1984. Intending and Acting: Toward a Naturalized Action Theory. Cambridge, MA: MIT Press. Brownstein, Donald. 1973. ‘Negative Exemplification.’ American Philosophical Quarterly 10: 43–50. Clarke, Randolph. 2014. Omissions: Agency, Metaphysics, and Responsibility. New York: Oxford University Press. Davidson, Donald. 1980a. ‘Actions, Reasons, and Causes.’ Chap. 1 in Essays on Actions and Events, 3–19. Oxford: Clarendon Press. ———. 1980b. ‘Agency.’ Chap. 3 in Essays on Actions and Events, 43–61. Oxford: Clarendon Press. ———. 1985. ‘Replies to Essays I–IX.’ In Essays on Davidson: Actions and Events, ed. Bruce Vermazen and Merrill B. Hintikka, 195–229. Oxford: Clarendon Press. Dowe, Phil. 2001. ‘A Counterfactual Theory of Prevention and “Causation” by Omission.’ Australasian Journal of Philosophy 79: 216–226. ———. 2009. ‘Absences, Possible Causation, and the Problem of Non-Locality.’ Monist 92: 23–40. Dretske, Fred. 1988. Explaining Behavior: Reasons in a World of Causes. Cambridge, MA: MIT Press. Enç, Berent. 2003. How We Act: Causes, Reasons, and Intentions. Oxford: Clarendon Press. Goldman, Alvin I. 1970. A Theory of Human Action. Englewood Cliffs, NJ: Prentice Hall. Hochberg, Herbert. 1969. ‘Negation and Generality.’ Noûs 3: 325–343. Hornsby, Jennifer. 2004. ‘Agency and Actions.’ In Agency and Action, ed. John Hyman and Helen Steward, 1–23. Cambridge: Cambridge University Press. ———. 2010. ‘The Standard Story of Action: An Exchange (2).’ In Causing Human Actions: New Perspectives on the Causal Theory of Action, ed. Jesús H. Aguilar and Andrei A. Buckareff, 57–68. Cambridge, MA: MIT Press. Lewis, David. 2004. ‘Void and Object.’ In Causation and Counterfactuals, ed. John Collins, Ned Hall, and L. A. Paul, 277–290. Cambridge, MA: MIT Press. Mellor, D. H. 1995. The Facts of Causation. London: Routledge. Moore, Michael S. 2009. Causation and Responsibility: An Essay in Law, Morals, and Metaphysics. Oxford: Oxford University Press. O’Connor, Timothy. 2000. Persons and Causes: The Metaphysics of Free Will. New York: Oxford University Press. Payton, Jonathan D. 2018. ‘How to Identify Negative Actions with Positive Events.’ Australasian Journal of Philosophy 96: 87–101. Priest, Graham. 2000. ‘Truth and Contradiction.’ Philosophical Quarterly 50: 305–319. Sartorio, Carolina. 2009. ‘Omissions and Causalism.’ Noûs 43: 513–530. Schaffer, Jonathan. 2005. ‘Contrastive Causation.’ Philosophical Review 114: 297–328. ———. 2012. ‘Disconnection and Responsibility.’ Legal Theory 18: 399–435.

66

Negative agency Shepherd, Joshua. 2014. ‘Causalism and Intentional Omission.’ American Philosophical Quarterly 51: 15–26. Smith, Michael. 2010. ‘The Standard Story of Action: An Exchange (1).’ In Causing Human Action: New Perspectives on the Causal Theory of Action, ed. Jesús H. Aguilar and Andrei A. Buckareff, 45–55. Cambridge, MA: MIT Press. Steward, Helen. 2012. A Metaphysics for Freedom. Oxford: Oxford University Press. Varzi, Achille C. 2006. ‘The Talk I Was Supposed to Give…’ In Modes of Existence: Papers in Ontology and Philosophical Logic, ed. Andrea Bottani and Richard Davies, 131–151. Frankfurt: Ontos Verlag. ———. 2008. ‘Failures, Omissions, and Negative Descriptions.’ In Meaning, Intentions, and Argumentation, ed. Kepa Korta and Joana Garmendia, 61–75. Stanford: CSLI Publications. Vermazen, Bruce. 1985. ‘Negative Acts.’ In Essays on Davidson: Actions and Events, ed. Bruce Vermazen and Merrill B. Hintikka, 93–104. Oxford: Clarendon Press.

67

5 BOUNDED AGENCY1 Elijah Millgram

Recent work on agency has been largely an attempt to characterize the ideal agent, that is to say, not necessarily an agent that is always successful, and not necessarily an agent that is always morally attractive, but in any case an agent that is always and 100% an agent. It is granted that real-world agency falls short of the ideal, but is either affirmed or presumed that the defective or incomplete agency we all-too-often encounter is to be understood by way of the ideal. To be sure, there is disagreement as to what the anchoring features of ideal agency are, with candidates such as full-fledged commitment to one’s actions, knowing what one is doing, and taking on challenges all in the mix.2 Here I want to recommend a different approach, one that takes the bounded-rationality research program as a model for investigating agency. Not only is all real-world agency, as I will explain, bounded, and not only should we try to make sense of the varied forms of bounded agency on their own terms, without seeing them as deviations from an ideal; we should not be trying to articulate a conception of ideal agency.

1 In the dialect of English encountered in analytic philosophy departments, “agent” has come to be just another word for “person”; here, we want to use it with a more tightly focused sense, which we need to first introduce. If I am standing on one side of the parking lot, and I kick a ball toward my car on the other side, it might get there… but even if my aim is good, it might not if, for instance, a gust of wind blows it off-course, or if a group of children get in the way and perhaps pick it up, or if a driver distracted by his search for an empty spot drives over it. Whereas if I am going to my car, I will compensate for the force the wind is exerting, detour around the children, and make sure to catch the eye of that absent-minded driver: in normal circumstances, I will get there anyway. Agents absorb various kinds of noise and buffeting from their environment; they stay on track; they exhibit, I will say, determination. Here I am borrowing ideas from recent work by Jenann Ismael (2016); briefly, in her view, what it is to be a self-governing system is to be constructed so as to see a course of action through in a way that buffers and absorbs physical noise: a self-governing system executes

68

DOI: 10.4324/9780429202131-7

Bounded agency

its plans resiliently. Extending that view, the buffeting an agent is able to absorb isn’t merely physical. A friend has decided that, in the interest of her sanity, going forward, an appropriate dosage of task-free leisure time has to be a priority; but every one of the tasks on her bottomless to-do list is accompanied by convincing reasons, and often enough, urgent reasons. If she were unable to resist the force of those reasons, she would be perpetually distracted, discombobulated, and swamped; agency, in her case, consists in part in protecting time in which to do, perhaps paradoxically, nothing. Perhaps the most memorable picture of agency as I am proposing to understand it comes from Nietzsche, describing how human beings have been brought to have the ability to make and keep promises: If we place ourselves at the end of this tremendous process, where the tree at last brings forth fruit… then we discover that the ripest fruit is the sovereign individual, like only to himself… [:] the man who has his own independent, protracted will… this mastery over himself also necessarily gives him mastery over circumstances, over nature, and over all more short-willed and unreliable creatures … those with the right to make promises… give their word as something that can be relied on because they know themselves strong enough to maintain it in the face of accidents, even “in the face of fate”… 3 So the extreme Nietzschean version of an agent, as I’m proposing to construe such a thing here, carries on not only in the face of gusts of wind and passing cars, but despite earthquakes and other disasters; and not only in the face of the sorts of urgent interruptions that impose themselves on my friend’s academic and domestic schedules, but even when presented with an offer that, as Mario Puzo’s Godfather phrased it, you can’t refuse. Leaving to one side Nietzsche’s apparent identification of determination with the will (if you look at the unabridged passage, with free will), we can make two further preliminary points. First, agency staying on track isn’t necessarily in the service of an objective, that is, a goal or an end. For a politician, say, to continue to live up to his commitment to serve his constituents and his party over the long term, he normally needs to absorb an ongoing stream of political exigencies, which he will do by changing out his goals, not to mention his principles, on a regular basis.4 This means that we need a more general way of talking about follow-through. Now, before “agent” became merely another synonym for “someone,” it meant someone whom you sent somewhere to do things on your behalf, as in the phrase “secret agent.” We can tip our hat to the former usage by saying that what an agent carries out is its mission. Second, what aspect of following through on a mission are we centrally after? Reasons for action are almost always defeasible: that is, although your conclusion—in the practical case, your evaluation or decision—really does follow from them, additional information or assessments can nonetheless be apparently compelling grounds from withdrawing it. My friend is right to conclude that she should set aside one day a week for gardening, trail running, and other leisure activities; but her decision can be defeated by, to start off what will prove to be an indefinitely long list, an injured colleague needing her to cover a class, or a hard drive crash, or the need to arrange accommodation for a disabled student, and so forth. An agent is not only able to compensate for gusts of wind and the like; it is good at absorbing prima facie defeaters—by which I mean now not just considerations that at first glance appear to be defeaters, but which, unless they are handled successfully, would be defeaters—and at not allowing its mission to be derailed by them.

69

Elijah Millgram

2 Now we need to fold the notion of bounded rationality into our discussion, and as I introduce it, I’m going to distinguish old-school and cutting-edge versions. Sometimes we understand the in-principle-correct way to figure something out, but it won’t actually solve our problem, because the procedure would take too long, or is otherwise too resource-intensive. For instance, we are told in decision theory classes to choose the option with the highest expected utility, but finding it would typically mean calculating the utility of every option, and satisficing can be a much faster alternative—that is, setting a threshold for what counts as good enough, and taking the first option that comes in over the threshold.5 Satisficing is an example of a heuristic, namely, a cost-effective and therefore feasible alternative to that in-principle-correct procedure, which is understood to give incorrect or suboptimal answers sometimes, but good-enough answers most of the time or on the most important occasions. Because we have finite minds and there are bounds to what we can do before a given deadline, we are boundedly rather than ideally rational—in the old-school way of construing that concept. The cutting edge of work in bounded rationality drops the assumption that there always (or even mostly) is an in-principle-correct way to figure things out; it avoids using ideal rationality as a reference point. It’s all very well to tell people to maximize their expected utility, but as it turns out, just about nobody satisfies the preconditions for having a utility function; for actual human beings, that method isn’t well defined. Or again, we’re supposed to compare the utilities of all the options, but if those lists of defeaters were genuinely openended, the option space is, again, not well defined.6 To be rational is to deploy heuristics that won’t always give us the right answer, but take into account the costs, computational and otherwise, of figuring things out, and will give us good-enough answers in a timely manner—even when we don’t have an ideal with which the heuristics are contrasted. Here is the observation we need as we turn from bounded rationality to bounded agency: because a heuristic trades off performance on some tasks for speed and cost improvements on others, heuristics perform well in some environments and not in others; there’s no such thing as a heuristic that is the right choice regardless of where it’s used. In domains where its repertoire of heuristics works well, a boundedly rational agent figures things out more or less rationally; in other domains, as its performance degrades, it is likely to look arational, irrational, or even just plain boneheaded.

3 An agent stays on mission, absorbing (as a character in a Bond thriller once put it) happenstance, coincidences, and enemy action. The resources needed to support determination are expensive, and so any real-world agent faces tradeoffs. And whether or not an agential resource supports determination is location-specific, in roughly the way that bounded rationality is: what keeps an agent on track in one kind of environment, and for one type of mission, will derail it in another. Laozi tells us: “A weapon that is too strong will not prove victorious; A tree that is too strong will break.” 7 We have already contrasted the unrelenting pursuit of an objective with the flexibility about their ends that keeps politicians in the game, and here are a handful of additional illustrations. In transparent institutional environments, it is the upstanding and incorruptible who are able to carry on with their missions, without being sidetracked by temptation and scandal. However, in an organization where functionaries have to be 70

Bounded agency

persuaded to do their jobs, and where they can only be paid off or threatened with bureaucratic decisions (not cash or other extra-institutional incentives), getting things done, and so, carrying through with one’s mission, requires that an agent be corruptible in the way that his colleagues are. (In the sort of environments I have in mind, if you don’t let others do you institutional favors and make bureaucratic concessions accordingly, you end up completely ineffectual.) In some environments, wealth supports agency; obstacles to one’s mission can be surmounted by spending money. But in other environments, the overhead involved in managing wealth turns out to be the impediment that keeps one from sticking with one’s plans. In adventure films, it is robust health, athletic ability, and quick reflexes that get the hero to the finish line; in periods of wartime mass conscription, it is the sickly and disabled whose lives manage to stay on track, and the fit whose best-laid plans get left behind in the trenches. So, in general, an agential resource will underwrite determination in some environments and not in others, and a class of special cases of special interest to philosophers is that of different architectures of agency. Some of these correspond roughly to various theories of practical reasoning, and I’ll gesture at just a handful of them.8 Perhaps the most widely recognized such architecture makes desires—or goals or ends, taken to be the objects of the desires—into the sole initiators of action. An instrumentalist agent of this kind will lapse into apathy or flailing in circumstances where its desires are moot, or irrelevant, or when they trigger only background processes: the ancien régime social climber, whose sole concern is his standing among the nobility, is left stranded when a revolution does away with the aristocracy; the peasant mobilized entirely by hunger does not know what to do with himself in a post-scarcity economy.9 Next, Michael Bratman has blueprinted an agential architecture built out of plans and policies. This sort of agent does not merely pursue whatever goals seem important to him at the moment; once he has adopted a plan, he sticks with it, unless special circumstances arise that would cause him to reconsider, and you would think that this particular design had determination as its very raison d’être. Nonetheless, as Candace Vogler has noticed, planning is a successful approach in a managerialist society, where planning is normal, the plans of other people and institutions provide a relatively stable background for one’s own plans, and agents have enough in the way of a resource buffer to proceed with their plans. If a would-be Bratmanian agent is so impoverished that trivial unexpected expenses regularly prevent him from stepping through his agenda, or if he lives in the ongoing chaos of a failed state, in which planning is a futile endeavor, the planning approach to life will fail to display as agency, as we are now construing it.10 Finally for the moment, agents that are feedback-driven, rather than goal- or plan-driven, can do better in some of the environments in which instrumentalist or Bratmanian agents break down. Such an agent registers when things are, say, going well and going badly, and is disposed to do, respectively, more of the same, or less; so it hill-climbs in the welfare space. An agent of this kind selects new goals on the basis of its affective feedback, and so can continue to operate when its former goals have slipped into irrelevance.11 Nonetheless, bounded agents of this variety perform well only in some sorts of environments; for instance, a feedback-driven agent can be trapped into addictions. Suppose you’re built to do more of whatever feels good, and you come across something that operates directly on the affective signal (nicotine, heroin, fentanyl, cocaine, or even just alcohol); you can end up as a junkie or alcoholic, someone who cares only about that something. But this particular vulnerability is one to which agents that are merely plan- or goal- driven are immune.12 We could survey further architectures of agency, but let’s recap. The crucial aspect of agency we are emphasizing here is determination: the propensity to carry on with one’s 71

Elijah Millgram

mission, despite being buffeted by variation in one’s physical, social, and other circumstances, and especially to absorb the reasons to deviate from it that will normally impinge on any extended course of action. The agential architecture which supports that propensity in one sort of environment will fail to do so in others. And that more than suggests that all agency is bounded.13

4 Is it really? Christine Korsgaard suggests thinking of agential architectures as the personal analogs of the constitutions of political states; France is on its fifth republic, and I won’t even try to count the other forms of government it has traversed; some such revolutions—by which I now mean transitions from one form of government to another, whether violent upheavals or not—were in retrospect inevitable. And here what goes for states goes for people. But even if we allow that any definite constitution will need to be discarded in suitably changed circumstances, what is to say that in principle we could not have a polymorphic agent, one that managed to be an ideal agent precisely by shifting its agential architecture in response to changes in its environment? We do not want to dismiss agential polymorphism. When human beings shape themselves to fit the highly differentiated slots produced by our division of labor, they adopt agential architectures suitable for one or another expertise. It is not only that the internalized standards and priorities that control their activities shift; deliberation is reshaped to match the requirements of one or another such social niche. For instance, if someone is in middle management, he’s probably oriented toward metrics, and he devises strategies for moving them; if someone is a researcher, he is—or ought to be—in the much more tentative business of exploring a terrain; that is, the two ought to decide what to do very differently. Moreover, people are able to move from one such highly specialized mode of agency to another, as when that researcher becomes a dean. So we do see agential polymorphism; our question about bounded agency proves to be whether to make sense of the polymorphism we exhibit by constructing a model of an ideally polymorphic agent. In Woody Allen’s mockumentary, Zelig, we are asked to imagine a human chameleon, someone who fits in anywhere.14 Now, and this is one way of taking the joke, what Zelig does is not possible: when he is talking to Frenchmen, he speaks French; in Germany, Zelig seems to be able to speak a language he has never learned, or to fake it well enough to fool native speakers. Given basic facts about our limitations—here, it’s not humanly possible to speak languages you haven’t learned—we are always boundedly conformist; for similar reasons, we always exhibit bounded agency. Pursuing the analogy with old-school bounded rationality would lead us to an argument that one can’t reshape one’s constitution to suit the demands of any and all environments whatsoever; the train of thought is straightforward enough not to detain us here. Let’s develop the analogy to cutting-edge work on bounded rationality. We opt for heuristics, you will remember, without necessarily having a conception of ideal reasoning with which to contrast them. Often, ideal rationality as promoted by decision theorists is not merely infeasible but moot, because the preconditions for the ideal being well-defined so rarely obtain in the real world: if your preferences don’t induce a utility function, there’s no such thing as maximizing your expected utility, and heuristics are thus not always fallbacks for when ideal decision-making is too costly or time-consuming. The more we think about an endlessly polymorphic agent, the less we understand what we are thinking about, in something like the way that we don’t understand what someone’s utility function would be, when his preferences are as all over the place as, say, mine. An 72

Bounded agency

agent stays on-mission because he has a stake in it, and an agent’s stake in his mission is tied to features of his agential architecture; for instance, a goal-driven agent exhibits determination in pursuing his goals, fielding potential defeaters for the steps he is taking toward them, and so on. A polymorphic agent will change out his agential architecture as appropriate in changed circumstances; but when the features of his former agential architecture are no longer present, how are we to understand his stake in his mission being sustained? (Not because he still has his goals.) Adapting our warm-up into an illustration, why does Zelig try to fit in everywhere? To really fit in, you have to be motivated in different ways in different places: in environments where discrimination makes “passing” a reasonable defensive strategy, conformism is motivated by fear; in others—think of the early kibbutzim—it is part of the enthusiastic pursuit of an ideal. If Zelig moves from that hostile society to the idealistic environment of a newly founded kibbutz and keeps fitting in, in a way that genuinely suits now one, now the other, how can his stake in his mission be the same stake? And if we don’t understand how it could be, is it the same mission? That we can’t give a general account of how one’s stake in a mission stays the same, when it has to undergo these sorts of substantive changes, tells us that we don’t have a coherent conception of determination for endlessly polymorphic agents. Just as you would not try to understand human language learning and social adaptation by trying to construct a theory of the inner workings of a Zelig, you would not try to explain the agential polymorphism that human beings in fact exhibit by attempting to construct a model of the ideally polymorphic agent. Ideal agency would exhibit determination in any environment whatsoever; sometimes it would have to do so through Zelig-like polymorphism, so we should not be using ideal agency as our reference point when we are trying to understand the bounded agency we encounter in real life.15

5 We’ve been considering an exotic (im)possibility, and it’s time to take a few steps back, rehearse how we got here, and say what we’ve learned. We haven’t given anything like a definition of agency, but we did firm up our conception of it in one direction in particular: it involves determination. And we pointed out that a given agential architecture (or deliberative constitution) will manage follow-through in some environments, rather than in others. Thus, if an ideal agent were one that (among whatever other things) exhibited stick-to-itiveness in any environment whatsoever, it would morph its own agential architecture, as it found itself passing from one such range of environments to another. And in principle, it would have to be ready, willing, and able to do so in any of indefinitely many ways, ways that it (and we) can’t anticipate. We then argued that we don’t really understand what it would be to be stay determined, across such arbitrary transformations of agential architecture. (That is, however, compatible with recognizing that it has been managed in one or another case.) Just as steam engines, long ago, had governors, different agential architectures rely on different controlling elements; we mentioned desires, plans, and affective feedback loops, but there must be indefinitely many types of controlling element doing their jobs in the indefinitely many arbitrarily different deliberative constitutions needed to keep an agent functional in the endless and as-yet-unimagined circumstances it might yet encounter. Without knowing what those controlling elements are and how they work, we don’t have the wherewithal to make sense of an endlessly polymorphic agent staying on point. If we can’t know what determination looks like in the endlessly polymorphic agent, we had better not insist on understanding ideal 73

Elijah Millgram

agency before trying to make philosophical sense of the bounded agents we are and meet on the street. Evidently, the next stage of a successful investigation of agency will proceed in the piecemeal manner of old-time botany and zoology. We have almost nothing in the way of a philosophical understanding of life in general; we suffice with understanding the forms and workings of life as we find it in some or another range of conditions.16 And likewise, given that all agency is bounded, we can anatomize first this type of bounded agent, and then that one, in each case looking at how it is that it remains on track, doesn’t give up, and completes its mission in the environments for which it is suited. As with bounded rationality, where the question is not what rationality tout court is, but what it can and ought to be for us, the question at hand is: What are the various forms of bounded agency that will allow you to, as the marines recruiting slogan has it, be all that you can be?

Related topics Material agency; The aim of agency; Planning agency; Minimal agency; Agency and games.

Notes 1 I’m grateful to Christoph Fehige, Luca Ferrero, Svantje Guinebert, C. Thi Nguyen, Madeleine Parkinson, Constantine Sandis, and Aubrey Spivey for comments on drafts; many thanks to the Hebrew University for a Lady Davis Fellowship, and to the University of Utah for support through a Sterling M. McMurrin Esteemed Faculty Award. 2 See, e.g., Katsafanas (2013), Korsgaard (2009), and Velleman (2015). 3 The passage is drawn from On the Genealogy of Morals, II.2; Kaufmann’s English rendering can be found at Nietzsche (2000, 494f ). 4 But then—I am used to hearing—isn’t serving his constituents and so on the goal? As I’m understanding goals here, they set finish lines one can reach, and they support means-end reasoning: you figure out steps that would reach the goal which, one after the other, you are then to take. (For an account of ends with roughly this shape, see Vogler, 2002.) The psychological correlate of goals or ends is desires, as philosophers nowadays mostly construe them, that is, as built around a representation characterized by one of two directions of fit: if the world isn’t as your representation has it, you change the world to make it match your representation, and when it does match, you stop (e.g., Searle 1983, 7f ). Now a responsible politician will always be updating his somewhat indefinite conception of what it is to serve his constituents and party, in the course of ongoing interaction with them—in something like the way that a responsible teacher continually clarifies to himself what he is attempting to do, or a responsible philosopher clarifies to himself and modifies his conception of philosophy, as he philosophizes. That is, there is nothing like a crisp, stable representation that the world is being brought to match. And there is no finish line, where the constituents have been completely served, and he can go home. 5 For discussion of the early bounded rationality tradition, with a focus on satisficing, see Bendor (2003). 6 For early work debunking the psychological realism of the conditions for having a von Neumann-Morgenstern utility function, see Kahneman et al. (1982); Millgram (2005, ch. 10) gives an argument to the effect that if you have a utility function, something is wrong with you. Wimsatt (2007) is an example of work on that cutting edge. 7 Daodejing, ch. 76., trans. P. J. Ivanhoe, in Ivanhoe and van Norden (2001, 200). I’m grateful to Eric Hutton for the reference. 8 To be sure, not all do: the fictional robots made famous by Isaac Asimov, which act on the instructions they’re given, subject to indefeasible side-constraints, instantiate an agential architecture that falls outside discussions of rationality. For treatment of a related personality structure, see Guinebert (2018).

74

Bounded agency 9 10 11 12 13

14 15

16

For some discussion of the problem, see Millgram (1997, ch. 5). Bratman (2007; 2018); Vogler (2002, 106f ). For agency of this kind, see Millgram (2005, ch. 2). When different architectures of agency generate actions, those actions are likely to themselves be variously structured, in ways which bear the respective stamps of their producers; for an overview of some of the variations, see Millgram, 2010. Haven’t we just characterized agency as essentially extended over time, and so excluded by stipulation agency that only lasts for a moment? Surely not all agency involves persevering. Although the conceptual home of agency is diachronic, it does seem to me to have a synchronic limit case. Determination exhibited in a course of action that only occupies a ‘specious present’—a moment that is long enough to be noticed, but not longer than that—is a matter not of continuing on with what you are doing, but rather has to do with managing distractions, and in particular the always available second thoughts about the merits of what one is about to do, in that very moment. The determined agent proceeds with his very brief mission, rather than letting the window of opportunity pass. Allen (1983). Would we have a clearer understanding of the special case of a polymorphic agent whose primary mission was preserving its own determination? (This would be the analog of a familiar philosophical move, as when Kantians make the agent’s own autonomy into what he must most value, or when Nietzscheans treat your “will to power” as what must be at bottom the only thing that matters to you.) It’s hard to see how this could help: in the polymorphic agent, the meta-mission of determination must appear in one formally and substantively different guise after another, and given that these guises are elements of agential architectures we have no way of describing ahead of time, we are not in a position to say anything useful about their common denominator. Surely we can sometimes make it out that an agent stays determined over alteration in its deliberative architecture; why isn’t making sense of a polymorphic agent just more of the same? Making out the mission of an agent, before and after such a transformation, to be the same, is something like solving a riddle, as that’s described by Cora Diamond (1991): that is, you only understand what would count as an answer to the question once you have answered it. Just as there can’t be a generic procedure for solving riddles of the form, Why is a like a ?, so there can’t be a general and antecedently available understanding of what makes the many versions of the mission of a polymorphic agent versions of the same mission. The most we seem to have along the former lines is something like a logical grammar which we bring to bear on living beings (Thompson 2008, Pt. I)—that is, logical form which any philosopher of biology will inform you fits life as we know it poorly indeed.

Further reading John Conlisk, Why Bounded Rationality? Journal of Economic Literature 34 (1996): 669–700. A useful survey of the first few decades of work on bounded rationality, with special emphasis on satisficing. Elijah Millgram, The Great Endarkenment. Oxford: Oxford University Press, 2015. Examines different forms of agency shaped by disciplinary specialization. C. Thi Nguyen, Games and Agency. Oxford: Oxford University Press, 2020. Argues that games can be a testbed for exploring different forms of agency. Nassim Taleb, Antifragile. New York: Random House, 2012. Explores the conceptual space needed to describe agency that is robust in the face of a changing environment. The Fall 2016 issue of Social Philosophy and Policy or, alternatively, David Enoch, Against Utopianism, Philosopher’s Imprint 18 (16), Sept. 2018, are recent entry points into the debate over (non-)ideal theory in political philosophy, for readers who would like to explore parallels with the move to bounded agency.

References Allen, W., 1983. Zelig. MGM, Santa Monica. Produced by Jack Rollins, Charles Joffe and Robert Greenhut. Bendor, J., 2003. Herbert A. Simon: Political Scientist. Annual Review of Political Science, 6, 433–471. Bratman, M., 2007. Structures of Agency. Oxford University Press, Oxford.

75

Elijah Millgram Bratman, M., 2018. Planning, Time, and Self-Governance. Oxford University Press, Oxford. Diamond, C., 1991. Riddles and Anslem’s Riddle. In The Realistic Spirit, pages 267–289, MIT Press, Cambridge, MA. Guinebert, S., 2018. Hörigkeit als Selbstboykott. Mentis, Paderborn. Ismael, J. T., 2016. How Physics Makes Us Free. Oxford University Press, Oxford. Ivanhoe, P. J. and van Norden, B., editors, 2001. Readings in Classical Chinese Philosophy. Hackett, Indianapolis, 2nd edition. Kahneman, D., Slovic, P., and Tversky, A., 1982. Judgment under Uncertainty. Cambridge University Press, Cambridge. Katsafanas, P., 2013. Agency and the Foundations of Ethics. Oxford University Press, New York. Korsgaard, C., 2009. Self-Constitution. Oxford University Press, New York. Millgram, E., 1997. Practical Induction. Harvard University Press, Cambridge, MA. Millgram, E., 2005. Ethics Done Right: Practical Reasoning as a Foundation for Moral Theory. Cambridge University Press, Cambridge. Millgram, E., 2010. Pluralism about action. In O`Connor, T. and Sandis, C., editors, A Companion to the Philosophy of Action, pages 90–96, Wiley-Blackwell, Oxford. Nietzsche, F., 2000. Basic Writings of Nietzsche. Random House, New York. Edited and translated by Walter Kaufmann. Searle, J., 1983. Intentionality. Cambridge University Press, Cambridge. Thompson, M., 2008. Life and Action. Harvard University Press, Cambridge. Velleman, J. D., 2015. The Possibility of Practical Reason. Maize Books, Ann Arbor, 2nd edition. Vogler, C., 2002. Reasonably Vicious. Harvard University Press, Cambridge, MA. Wimsatt, W., 2007. Re-Engineering Philosophy for Limited Beings. Harvard University Press, Cambridge.

76

6 AGENCY AND GAMES C. Thi Nguyen

Games can teach us an enormous amount about the nature of agency. Games are the formalization of play, and play is fertile territory for exploring some of the odder corners of our agency. But, for the most part, philosophical approaches have analyzed how our agency functions in more serious domains. These conversations about agency have largely centered on the analysis of deliberation and action as they proceed from our core and enduring values. They have worried about cases in which our agency might be permanently undermined. Thinking about games and play, on the other hand, teaches us about the nature of agency during less serious endeavors—when we hold our aims more lightly. This peculiar lightness is clearest in our relationship to the ends of a game. Something quite peculiar can happen when we play a game. We can adopt new roles and new modes of interaction; we can pursue goals that we don’t normally care about. We can bring ourselves to care desperately about trying to get a ball through a little hoop, even when that act means nothing outside of the game. Strangers can become teammates; friends can become opponents. And then, when the game is done, we can step back and let it all go. The ends, commitments, and relationships we take on in a game can be curiously fleeting. My own work recently has been devoted to showing what the philosophical analysis of games can show us about the nature of our own agency. In my terms, playing games involves taking on an alternate motivational structure and an alternate, designed agency. Very little has been written explicitly on this topic in philosophy, so the following summary will, perforce, draw heavily on my own work. I think that the relationship is natural, however, and that many of these insights can be found, or at least hinted at, in the non-philosophical discussion of games. So this will be something of a translation project, in which we attempt to mine a rich vein of thinking, by game designers, game-players, and humanists, to shed some light on our own agential complexity.

1 On games and roles In Homo Ludens, one of the foundational texts on games, Johan Huizinga offers an account of the peculiar nature of play. “Play is the direct opposite of seriousness,” says Huizinga (Huizinga 1955, 5). First, play is voluntary. By its nature, if you’re forced to do it, it isn’t play. Second, play actions have different meanings from ordinary or real life actions. It is, in an DOI: 10.4324/9780429202131-877

C. Thi Nguyen

important way, a form of pretend. And third, play is secluded from ordinary life, set apart in its own space and time (7–9). But play, far from being utterly anarchic, is deeply concerned with rules. Says Huizinga: All play has its rules. They determine what “holds” in the temporary world circumscribed by play. The rules of a game are absolutely binding and allow no doubt… Indeed, as soon as the rules are transgressed the whole play-world collapses. The game is over. The umpires whistle breaks the spell and sets “real” life going again. (11). Play involves a certain looseness from the normal rules of everyday life, but also a willingness to plunge into alternate sets of rules. Game-playing, says Huizinga, is part of a larger family of human practices, including theater and sacred rituals, where we enter a special space and adopt alternative roles. Games take place in the “magic circle” of play. Game designers Katie, Salen and Erich Zimmerman (2004) took this analysis and made it the central concept in their influential game-design textbook, Rules of Play. Their take is much harder-edged than Huizinga’s original. According to them, all play takes place in: (1) A bounded space for play, formally separated from everyday life (2) Precisely defined in space and time (3) Players crossing this boundary enter an alternate world, such that new rules have authority, and actions and objects acquire new meanings (95–97). Their account has since been subject to much criticism, with much attention paid to whether the space of play is really so absolutely separated from ordinary life. Much of the criticism has been devoted to the claim, attributed to Salen and Zimmerman, that the magic circle is utterly impermeable—that no action inside the magic circle can have any effect, consequence, or moral implications for life outside the magic circle. Scholars have pointed, for example, that gambling-play yields real profit and loss, and that winning athletes can earn status in the community (Malaby 2007; Taylor 2007, 2009).1 Surely, the magic circle of play isn’t perfectly impermeable. I can’t murder my opponent in a game of chess and then tell the jury to absolve me because it was only a game. But even if the most extreme version of the claim is untrue, talk of the magic circle approaches an important truth about game-play, whose familiarity may mask its oddity.2 Playing a game does involve adopting different and temporary relationships and aims. Game actions do not have all of their usual meaning and consequence. When we are on opposing basketball teams and I block your pass, it would be very strange if you were to stay sullenly angry with me after the game for my betrayal. There have been many attempts to explain the odd motivational states involved in games. For example, Robert Simon (2014) has argued that, in games, we are only apparently competing. We are, he says, actually cooperating in order to help each other achieve some sort of excellence. In a similar vein, Steven Weimer (2012) suggests that a game is a kind of social contract. In a boxing match, he says, we have entered into a contract to try to punch each other. We have a reason to enter into the contract: to develop our own excellences. And once we have entered into that contract, we have a reason to try to punch the other person: to fulfill our contractual duties.3 These accounts attempt to explain away the strangeness of game-play. But they are inattentive to the basic phenomenology of game-playing. When I am playing a serious game of 78

Agency and games

chess against you, I am not trying to help you improve. I am, for the duration of the game, simply trying to win. There is a very different mode of engaging with a game in which I would aim at improving my opponent. This is the state of teaching, or coaching, or being a tutor. A coach devotes their mental energy to trying to figure out what their student needs, and providing just the right level of opposition to foster improvement—and that aim drives the selection of their every move. But the coaching mindset is an entirely different one from the competitive playing mindset. Similarly, if our goals for boxing were simply to develop our own excellences, then we should not particularly care about winning and losing each particular match. We should care about whether we improved—which is not particularly correlated with winning or losing. And if our reasons for striking our opponent were merely that it was our contractual duty, we would be interested only in if we had fulfilled our obligations to provide them with a basis for their self-improvement—and not in, say, actually winning. Neither Simon’s coaching mindset nor Weimar’s contractual mindset can explain an elemental part of the experience of playing games. During the game, we care about winning. It grips us. We are terrified when threatened with defeat; we are elated when we pull off a clever stratagem. Explanations like Simon’s and Weimar’s attempt to assimilate game-playing to ordinary practical life and, in doing so, iron over some of the exquisite oddity that Huizinga was trying to expose. To put Huizinga’s insight into the terms of contemporary philosophy: games seem to involve a peculiar kind of agency-switching. When we play a game, we adopt different motivations—connected to alternate sets of abilities. In normal life, I am a person with hands and feet and the ability to use both. But when I play soccer, I temporarily make myself into a modified agent—one who cannot use their hands, and who has an overriding goal of getting this ball into that goal. And in multiplayer games, those temporary goals take on a social dimension. This is clearest in games with teams. In basketball, my interests are united with my teammates. Our score is tabulated collectively and not individually; insofar as we care about winning (rather than our own career or stats), then our interests are temporarily merged. To make sense of the agential fluidity of play, we need to come to grips with the complex and tangled motivational structure of game-playing. In ordinary life, the goals we pursue in a local activity lead directly to our larger and enduring goals. But there is, in gaming life, a fascinating cleavage between local goals and larger purposes. When I play a game, often, I need to try to win. But I don’t actually care about winning in an enduring way. What I really care about is, say, making sure everybody has fun. But the best way to get that fun is if we players try, during the game, sincerely to win. If we are all playing a party game of Charades for fun, the fun only arises when we are desperately trying to communicate under the bizarre constraints of Charades. The game is fun because it is a bit thrilling and frustrating, and it is only thrilling and frustrating if we adopt the goal of success at communication. But after the game, our interest in success lifts. After all, if I lost at Charades, I wouldn’t think that I’d wasted my evening. Unless I am a very bad sport, my interest in winning should evaporate when the game is through, and I should evaluate the evening’s success in terms of how much fun everybody had, and not in terms of whether or not I won. The goal of game-playing can, then, be oddly disconnected from my larger purpose of play. The problem with accounts like Weimer’s and Simon’s lies in their attempt to explain the motivational structure of play in a linear fashion. If we insist in treating player’s in-games motivations as consistent with their enduring motivations, then we will either have to think that all competitive players are some kind of sociopaths, or we will have to attribute very strange and kindly motivations to the players in-play. Neither approach, I think, helps us 79

C. Thi Nguyen

make sense of the place of games and play in our social lives. To understand games, we will have to understand the odd motivational transformation that we can undergo when we play.

2 Suits on the nature of games To understand this oddity, let’s turn to Bernard Suits’ illuminating account of games. Suits offers us two versions of his account: a technical and complex official version, and a “portable” version. For brevity’s sake, I will rely on the portable version.4 The portable version is: Playing a game is the voluntary attempt to overcome unnecessary obstacles. (Suits 2014[1978], 43) As Suits explains, a difference between ordinary activities and gaming activities is that, in ordinary activities, we take the most efficient pathway to our goals. But in gaming life, we take on extra constraints. We try to cross the finish line of a race, but we forbid ourselves from getting on a bicycle or calling a cab. We try to solve the fiendish puzzles in a computer game, but we refuse to look up the answers online. In games, we not only pursue a goal, but we add unnecessary obstacles in our path to that goal. This, says Suits, reveals something very interesting. The bare state of affairs that we’re trying to achieve within the game is not, by itself, what we value. If we just cared about getting to the finish line—to that particular location in space—we would take the most efficient means of transport. If we actually cared just about getting the ball through the hoop, we would just go to an empty basketball court with a stepladder and pass the ball through the hoop to our heart’s content. The goal of the game isn’t to bring some state of affairs into being, but to achieve that state of affairs inside certain constraints. What “making a basket” actually is, in basketball, is not simply getting the ball through the hoop. It is making the basket while facing opponents and obeying all the rules of basketball, like the rule requiring you to dribble the ball while you move. In other words, the goal of basketball—making baskets—is partially constituted by certain constraints of how we achieve it (24–43). In ordinary life, says Suits, we are often engaged in technical activity. We pursue ends that are valuable on their own. A steaming, crispy, fresh-made falafel wrap is valuable in itself, because it is utterly delicious. And it is instrumentally valuable, because it is full of protein. We can specify these values independently of any particular means of obtaining the falafel wrap. On the other hand, the value of making a basket is inextricably tied up with the means by which I made the basket. If I made the basket with a stepladder and no opponents, then it isn’t valuable at all. It’s only valuable insofar as I made it under certain constraints. Many have read Suits as saying that game-play must be done solely for the intrinsic value of the struggle itself. This would be a radically restrictive view of game-playing. For one thing, it would mean that you could never play games professionally, since you’d be playing as a means to get money. But Suits rejects this view quite explicitly. Imagine an amateur mountain climber, who is trying to climb K2 for the joy of the struggle. Then imagine a professional mountain climber, who is trying to climb K2 for the money and glory that will follow. For Suits, both amateur and professional count as playing a game because the goal they are aiming at—having climbed the mountain—depends on having done it under certain restrictions. Suppose a helicopter passes by and offers them a lift to the top. Neither will take it. The amateur won’t take it because it will eliminate the joys of the struggle. The professional doesn’t take it because then they won’t have climbed the mountain and won’t get any of the accolades that follow 80

Agency and games

from that achievement. As Suits says, the professional wants to win for the proceeds of winning, but to get those proceeds they must win at the game. The constraints of ascending under their own power help constitute the activity and the goal at which they aim. Now compare the professional and the amateur mountain climber to an herbalist climbing K2 in search of a rare medicinal herb that grows at the top. The herbalist doesn’t care about the struggle; they are simply enduring it in order to get the goods. If somebody offers them an easier route—like a free helicopter ride to the top, or an online shop for that herb— the herbalist will take that instead. In technical activity, we simply want some end-state, for which we will take the most efficient possible path. In gaming activity, the goal we care about is partially constituted by certain inefficiencies.5

3 Striving and achievement play Suits’ analysis, as I’ve argued elsewhere, reveals that there are actually two motivational attitudes available in such obstacle-oriented play (Nguyen 2019, 2020, 1–51). First, there is achievement play: playing a game for the sake of the value of winning. Second, there is striving play: temporarily taking on an interest in winning, for the sake of the value of the struggle. The achievement player genuinely wants to win. The striving player’s interest in winning is merely instrumental to their true purpose: having a good struggle. We can detect the motivational attitude of the player from how they would react toward losing. If an achievement player loses, then the game-playing event was mostly valueless to them. They have significantly wasted their time. The striving player’s interest in winning, on the other hand, is merely a temporary construct. Once the game is through, their evaluation of the worthwhileness of their activity hinges on some other quality. Perhaps they care that the game was fun, absorbing, interesting, or relaxing—or that it gave them a workout. The game can be entirely worthwhile, even if they lost, so long as the struggle to win was satisfying. The motivational structure of striving play is refracted. What we aim at during the game is not what we, in fact, value. It is easy to confuse the distinction between striving play and achievement play with the distinction between intrinsic and extrinsic value. But these distinctions track very different dimensions. The intrinsic/extrinsic value distinction concerns whether the game is valuable in itself or as a means to something else. The striving/achievement distinction concerns whether the value of the game adheres to the process of play or to the victory. And these dimensions come apart. You can have intrinsic achievement players, who care about the win for its own sake. But you can also have extrinsic achievement players, who care about victory for the sake of what follows from it, like fame, money, or status. You can have intrinsic striving players, who care about the value of struggling itself. Or you can have extrinsic striving players, who care about the goods that follow from the struggle, like, say, fitness or stress-relief. Notice that if your interest in running marathons is to become healthier, you are an extrinsic striving player. And we can tell because it doesn’t really matter if you actually win the marathon. What matters is that the attempt to win gets you involved in an activity that helps make you healthier. And one can have both sorts of motivation, in varying proportions. As a matter of fact, many players value striving and achievement, and do so both intrinsically and extrinsically. Many professional rock climbers care about the aesthetic glory of the movement itself and about the health benefits they get from performing them. At the same time, they also care about achieving victory over a particular climb in and of itself and about the various endorsements and sponsorships that will support them, which are earned through a track record of victories. 81

C. Thi Nguyen

Of course, one might doubt the reality of striving play. Why think that we are capable of such a peculiar and tangled motivational state? Why not think the much simpler thing: that all players are achievement players, and pursue the win simply because they care about winning? For one thing, though we may pursue the win with all our might during the game itself, we will often take actions outside of the game that make winning less likely. For example: suppose my Tuesday night gaming group has discovered a new board game in which we’re all equally skilled at. Our games are exquisitely balanced struggles. Suppose that, late one Monday night, I discover a strategy guide for that game while surfing the web. I have a choice: I can either read the guide or avoid it. (And I know nobody else will have the time to read the guide between now and then.) If I was an achievement player, then the only sensible action for me is to read that strategy guide. After all, what I care about is winning, and that’s the best way to win. But I, myself, wouldn’t read the strategy guide, because I think it would wreck the exquisitely balanced struggle that we had managed to find. And I do not think my behavior is irrational. Reading the strategy guide would make winning easy, which would deprive the struggle of its complexity and savor. And this decision only makes sense if I am a striving player, and what I truly value is the interesting struggle, and not the win. For another, consider the phenomenon of what we might call “stupid games.” A stupid game is a game where: (1) The fun part is failing. (2) You can only have fun if you’re actually trying to win. Examples of stupid games include Twister, the children’s game of Telephone, and many drinking games. The funniest part of Twister is when you collapse into a heap. But if you actually aim directly at collapsing into a heap, it won’t be funny at all. It’s only funny as a genuine failure, and it’s only a genuine failure if you were actually trying to win. So if we can actually play stupid games, then striving play is a real motivational possibility. In stupid games, we aim at success, but success is not what we actually value. We aim at success—and become temporarily absorbed in really trying to get it—to get at our real, enduring interest, which is the experience of comic failure. Stupid games illustrate our capacity for a particular sort of complex psychological self-manipulation. We can bring ourselves to become absorbed in the pursuit of an end that we don’t actually care about—that is, in this case, essentially the opposite of what we care about. Striving play also helps us see how games have been so often misunderstood and misvalued. A common criticism of games—usually made by those who don’t play them—is that game-playing is stupid, because you’re doing all this struggle for something utterly arbitrary and meaningless. Getting to the top of the cliff, getting the ball in the hoop, getting more experience points in Dragon Age—these goals have no real value. But that criticism, we can now see, misunderstands the motivational structure of striving play. In striving play, we shouldn’t look to the end-goal, or what follows from it, to really understand the value of the activity. The pursuit of those end-goals is justified by how they shape the activity of their pursuit. In ordinary life, we pursue the means for the sake of the ends. But in striving play, we pick our ends for the sake of the means they force us through.6

4 Layered agencies Thinking about stupid games highlights a particular capacity we have in our agential self-manipulation. We can take up a local goal for a larger purpose. We can then temporarily seal ourselves off from the larger purpose, to some extent, and put it mostly out of mind. 82

Agency and games

A friend told me the following story: he was playing Monopoly with his ten-year-old son. For the first time ever, the son was beating the father, and was utterly delighted by the experience. The father would be on the verge of losing and then look down and find he had slightly more money than he had expected. It turned out that the son was actually distracting his father and then slipping a little extra money in his father’s pile, to elongate the experience of beating his father. This story is funny precisely because of what the son doesn’t understand about game-playing. But what the son doesn’t understand is actually quite complicated. The son is clearly a striving player. He is playing, in significant part for some delight in the process, rather than in the victory itself. Since he knows he is playing for the sake of the struggle and winning would end the struggle, he tries to forestall the win when it comes close. When described this way, the son’s actions may seem like utterly rational behavior—though the game-player will intuitively recognize that something has gone amiss. What the son misunderstands is that what’s good about the struggle is precisely that it is a struggle to win. To put it another way: we cannot keep perfectly transparent the way our larger purpose supports our local goal, if we are to achieve the typical goods of striving play. Suppose that a striving player kept constantly before their mind the fact that they had adopted their local goal—winning—merely as an instrument to achieve their larger purpose—being absorbed in a struggle. In that case, when the win was in their grasp, it would be reasonable for them to make a bad move and throw the game, so that the struggle could go on. But this would create a perpetually anxious secondary consciousness. They would need to try to win, but also need to try to undercut the win if it approached. This makes some of the key pleasures of game-playing impossible. What many striving players want is the experience of total absorption in a singular instrumental struggle. So it must be that striving players can create layers of agency. They can set up an alternate sub-agency and then submerge themselves within it. They can momentarily occlude their larger purposes and bury themselves in an alternate agency with alternate goals. But that submersion is, in most cases, limited and tenuous. All kinds of things can break through. If somebody has a heart attack, or breaks down into tears because of a nasty text from an ex, or if it becomes obvious that the game is just too boring and miserable for everybody involved—we can snap out of the inner layer and recall our larger purpose (Nguyen 2020, 52–73).

5 The art of games Thinking about what games reveal about our agency, we can also see something crucial about the nature of games. What is special about games, as an art form? Many defenders of games have tried to argue for the cultural importance of games by assimilating them to some other art form with a longer history of cultural recognition. Scholars have tried to argue that games are art because they are special new kinds of fiction (Tavinor 2009); because they are a new digital and interactive variant of cinema (Gaut 2010); because they are a kind of socially transformative conceptual art (Flanagan 2013); or because they can present rhetorical arguments (Bogost 2010). But the anxiety to assimilate games to more respected art forms can, I think, obscure their real potential as a novel and distinctive form of art. What is the game designer’s real art? Reiner Knizia, the great German board game designer, says that the most important tool in his game-design toolbox is the points (quoted in Chalkey 2008). The points set the motivation for the players—they tell the player what matters, what they are to pursue. Using the analysis I’ve offered, we can put Knizia’s insight in the following way: the artistic medium of games is agency itself. The game designer sculpts 83

C. Thi Nguyen

who we will be during the game. They tell us what we will care about, what abilities we will use to try to attain it, and what practical obstacles we will face. And they can do so, not just to tell stories or make arguments but to give us an experience of our own shaped, beautiful action (Nguyen 2020b). This reveals one of the special functions of games. Games let us record modes of agency, and then communicate them to others. Each game encodes a different form of agency—a mode of being practical. Chess encodes the mode of tight, sharp, rigorous tactical calculation. Diplomacy encodes Machiavellian manipulation and deceit. Spyfall, a game of discovering the traitor in your midst, encodes the concentrated interplay of deceit and deceit-detection. Tetris encodes the mental mode of imagining rotations and fittings. Portal encodes thinking about the world through fourth-dimensional manipulations. And we can pick these up from games. I myself learned the sharp, tactical frame of mind required for analytic philosophy from chess. Games can enhance our freedom and autonomy by giving us exposure to different mindsets. Martha Nussbaum (1992) suggests that narratives are how we pick up different finely attuned emotional perspectives—how we find out what it is like to live life from another emotional perspective. I am suggesting that games do something analogous, but for practical styles rather than emotional perspectives. Games constitute a library of agencies.

Related topics The aim of agency; Aesthetic agency; Agency and normativity.

Notes 1 Zimmerman (2012) has since denied that he and Salen had intended that the “magic circle” concept include such a demand for radical impermeability, calling it a straw man uncharitably imposed on their text. For a further discussion, see Nguyen (2017b, 2020, 177–180). 2 For some useful recent attempts to revive the magic circle view in particular, see Stenros (2012) and Waern (2012). 3 For a more elaborate analysis of these particular views, see Nguyen (2017a). 4 Using the portable version here will conceal certain complexities. For a full discussion of the official version, see Nguyen (2020, 27–51). 5 The comparison between the amateur and the professional is directly adapted from Suits (2014 [1978], 90–92, 154–160). The herbalist is my own addition, to clarify what I find to be the most common misreading of Suits (Nguyen, 2020, 30–32). 6 The argument for the possibility of striving play is significantly elaborated in Nguyen (2020, 27–51).

Further reading Bradford, Gwen. 2015. Achievement. Oxford: Oxford University Press. -Recent central work on the notion of achievement; offers the most contemporary version of a perfectionist, excellence-oriented explanation of the value of playing games Huizinga, Johan. 1955. Homo Ludens: A Study of the Play-element in Cult. Boston: Beacon Press. -The foundational text for the study of games; establishes the idea that play spaces are an alternative, secluded space where we take on alternate roles and alternate rules apply. Lugones, Maria. 1987. Playfulness, “world”-travelling, and loving perception. Hypatia 2 (2): 3–19. -Important feminist analysis of “playfulness” as moving between normative worlds, and as holding onto those worlds lightly. Nguyen, C. Thi. 2020. Games: Agency as Art. Oxford: Oxford University Press.

84

Agency and games -The chapter author’s account of games as the art form working in the medium of agency; stresses the way that games sculpt agency in order to sculpt aesthetic experiences of action. Suits, Bernard. 2014[1978]. The Grasshopper: Games, Life, and Utopia. 3rd ed. Peterborough, CAN: Broadview Press. -The definitive work in analytic philosophy on the concept of a game; argues that to play a game is to voluntarily take on unnecessary obstacles to make possible the activity of struggling to overcome them.

References Bogost, Ian. 2010. Persuasive Games: The Expressive Power of Videogames. Cambridge: MIT Press. Chalkey, Dave. 2008. Reiner Knizia: “Creation of a Successful Game”. Critical Hits. http://www. critical-hits.com/blog/2008/07/03/reiner-knizia-creation-of-a-successful-game/ Flanagan, Mary. 2013. Critical Play: Radical Game Design. Cambridge: MIT Press. Gaut, Berys. 2010. A Philosophy of Cinematic Art. Cambridge: Cambridge University Press. Huizinga, Johan. 1955. Homo Ludens: A Study of the Play-element in Cult. Boston: Beacon Press. Malaby, T. M. 2007. Beyond play: A new approach to games. Games and Culture 2 (2): 95–113. Nguyen, C. Thi. 2020. Games: Agency as Art. Oxford: Oxford University Press. ———. 2020b. The arts of action. Philosopher’s Imprint 20 (14): 1–27. ———. 2019. Games and the art of agency. Philosophical Review 128 (4): 423–462. ———. 2017a. Competition as cooperation. Journal of the Philosophy of Sport 44 (1): 123–137. ———. 2017b. Philosophy of games. Philosophy Compass 12 (8). Nussbaum, Martha. 1992. Love’s Knowledge: Essays on Philosophy and Literature. Oxford: Oxford University Press. Salen, Katie and Eric Zimmerman. 2004. Rules of Play: Game Design Fundamentals. Cambridge: MIT Press. Simon, Robert L. 2014. Fair Play: The Ethics of Sport. Boulder: Westview Press. Stenros, Jaakko. 2012. In defence of a magic circle: The social and mental boundaries of play. In Proceedings of DiGRA Nordic 2012 Conference: Local and GLobal – Games in Culture and Society. http:// www.digra.org/wp-content/uploads/digital-library/12168.43543.pdf. Suits, Bernard. 2014[1978]. The Grasshopper: Games, Life, and Utopia. 3rd ed. Peterborough, CAN: Broadview Press. Tavinor, Grant. 2009. The Art of Videogames. Oxford: Wiley-Blackwell. Taylor, T. L. 2009. Play between Worlds: Exploring Online Game Culture. Cambridge: MIT Press. ———. 2007. Pushing the borders: Player participation and game culture. In Structures of Participation in Digital Culture. Ed. Joe Karaganis. New York: Social Science Research Council. Waern, Annika. 2012. Framing games. Proceedings of Nordic DiGRA 2012. Weimer, Steven. 2012. Consent and right action in sport. Journal of the Philosophy of Sport 39 (1): 11–31. Zimmerman, Eric. 2012. Jerked around by the magic circle: Clearing the air ten years later. Gamasutra, February 7th. https://www.gamasutra.com/view/feature/135063/jerked_around_by_the_magic_ circle_.php

85

PART 2

Kinds of agency Introduction to Part 2 Luca Ferrero

The philosophy of agency is primarily concerned with the full-blooded agency of adult human beings. This agency goes beyond self-maintenance and goal-directed purposiveness, and it has been variously characterized as intentional, rational, reflective, self-conscious, and autonomous. This is the kind of agency usually taken to be both necessary and sufficient for being held responsible and subject to the demands of morality. But is this the only kind of agency that can be found in the world? First, could there be full-blooded agents who are not human beings? As far as we know, there might be other beings in the universe who enjoy full-blooded agency. And if there are none, this might simply be because of some accidents in the history of the universe. Humanity as a biological category might not be essential to full-blooded agency. It also seems possible to conceive of supernatural beings, like angels, demons, and gods, who might enjoy agency of the full-blooded kind, possibly in a form that is better than ours (if not even perfect) along such dimensions as intentionality, rationality, reflection, self-consciousness, and autonomy. Second, what about beings who do not have the full set of powers and properties distinctive of full-blooded agency? Consider non-human mammals, for instance. They seemingly engage in behaviors that have at least some of the following features: they are purposive or goal-directed in a complex way, rather than being merely reactive or reflexive; they are usually directed at their own self-maintenance; they seem to originate in the mammals themselves rather than being simply the direct effect of fully external causes; they are responsive, often in opportunistic and clever ways, to the mammal’s specific needs and desires in light of its view of its individual circumstances and in such a way as to make its conduct intelligible, at least within the confines of the mammal’s biological nature and psychology. All of these features are also present in our own case, even in the absence of more sophisticated features due to reflection and self-consciousness. This suggests that, at least for some organisms, attribution of some kind of genuine agency might be in order, even if an agency does not include all of the dimensions and aspects of our own agency. But why stop there? Could it be that there are simpler kinds of agency possessed by even simpler organisms? Several philosophical questions arise at this point: what are the minimal conditions for the attribution of genuine, even if simpler, agency? Do all organisms exhibit at least some kind of agency? Does agency, even if only in a minimal form, extend at least as far as biological life? DOI: 10.4324/9780429202131-987

Luca Ferrero

Or does agency only come into existence with some more complex kinds of life, especially those that exhibit some kind of mentality or psychology (such as representational mental states)? What are the implications for attributions of responsibility, moral status, and the imposition of moral demands, if some kind of simpler agency can be genuinely attributed to non-human organisms? These and related questions are discussed in the chapters on ‘Minimal agency’ and ‘Animal agency.’ If we then look back at our own agency, we should ask about what differences such features as intentionality and rationality make. What do they contribute to agency? Are these additions simply incremental or radically transformative? These are the topics of the chapters on ‘Intentional agency’ and ‘Rational agency.’ (Related questions are also raised when we consider the kind of distinctive knowledge that we might have about our own intentional agency—see later ‘Agency and practical knowledge’—and the temporal structure of our agency—see ‘Diachronic agency’ and ‘Planning agency.’)

Minimal agency In ‘Minimal agency,’ Hans van Hateren presents and evaluates various theories that have been proposed to characterize the minimal agency of biological organisms. He begins by discussing the ‘explanatory role’ theory, according to which agency can be ascribed whenever it is explanatory useful in accounting for an organism’s conduct. The problem with this view is that it does not offer an actual account of the causal role performed by agency. A better alternative seems to be the neo-Aristotelian ‘intrinsic power’ theory, according to which agency is an intrinsic causal power of organisms. The trouble with this view is that it seems to fail to accord with the naturalistic explanations offered by current scientific views. An alternative view, ‘evolutionary etiology,’ rests its explanation on an etiological understanding of biological function, which is, however, unable to explain the present causal efficacy of biological functions as opposed to their historical provenance. This problem does not seem to arise for the ’organizational theory,’ which characterizes agency as a central element in adaptive mechanisms that maintain the structural integrity of organisms. This is a promising view, which accounts for several features of agency. Unfortunately, it still cannot explain the kind of freedom enjoyed by organisms and relies on some groundless norms. Finally, Hateren presents the ‘fitness estimation’ approach, which explains agency as the product of an internal estimate of the organism’s own evolutionary fitness. This approach is promising because it can, in principle, account for the emergence of minimal agency with all of its features. This is a relatively novel proposal, which is promising but is still awaiting adequate empirical validation. The chapter concludes by showing how an account of minimal agency can help explain several features of full-blooded agency, since many of them (such as teleology, normativity, freedom, individuality, representation, and causal efficacy) are already present in a simpler form in minimal agency. Other aspects—such as consciousness, symbolic representation, and social agency—are not present as such, but we might envisage ways to explain their emergence (and demystify them) starting out from simpler kinds of agency.

Animal agency As Helen Steward argues in ‘Animal agency,’ the topic of animal agency has been unjustly neglected in the mainstream philosophy of agency. This neglect is doubly problematic: first, it has prevented us from having a clear view of the nature of animal conduct; second, it has 88

Kinds of agency 

obscured our view of the metaphysics of human agency as well, given that we are, first and foremost, animals. Hence, she argues, any account of our agency that ignores our animality is doomed to failure. In this respect, a problem might arise for a causal theory of action, according to which a purposive bodily movement counts as an action only if it has been caused by the appropriate sorts of mental states or events. The worry is that, although animals exhibit some kind of mentality (e.g., they see, want, and try to get certain things), they might lack more sophisticated mental states, such as beliefs and intentions. But even if one is willing to allow for less sophisticated kinds of mentality to play the required causal role, there is still the worry that such causal theory might suffer from a dualistic heritage. This heritage encourages us to think about the agential production of effects in the world as though it were equivalent to the production of effects by minds alone. But a less mentalistic way of thinking might be in order: one that focuses on seeing the agent itself rather than their mind as the source of action. A second issue that arises in connection to animal agency concerns the range of organisms to which agency can be properly attributed. On the one hand, one might be inclined to rule out cases such as a paramecium, whose behavior is just the product of a set of simple mechanistic creature-environment interactions. On the other hand, it is unclear how a mere increase in the complexity of basically mechanistic interactions would make such a dramatic difference in the kind of agency that could be attributed to a creature. Steward then considers whether rationality is key to the distinction between animal and human agency. The matter is difficult to adjudicate, given that the participants in the debate might simply be talking past one another. For some philosophers, there is a basic continuity in nature in the manifestations of rationality, so attributions of rationality vary in degree. But for others, being rational amounts to having a different kind of life, one that is ‘transformed’ by the acquisition of rationality, that is, of the ability to respond to reasons as such, to justify one’s actions, and to act on principles. In closing, Steward considers the ethical implications of acknowledging that animals are genuine agents. These attributions of genuine agency do not necessarily entail that animals are rational and/or morally responsible agents. But this does not mean that we owe no duties to animals. We still owe respect to animal agency, even if it fails to be full-blooded rational and to support moral responsibility. Steward concludes that we still have pro tanto obligations, within our own powers, to allow animals to lead lives in as natural a way as possible, unobstructed by our interference.

Intentional agency It is often argued that a distinctive feature of our full-blooded agency is its ‘intentional’ character. However, as Lilian O’Brien shows in ‘Intentional agency,’ there might be a meaningful general category of intentional agency that includes creatures such as cats, children, and sophisticated adult human agents (including those acting under duress). A necessary but not sufficient condition is the capacity for purposeful behavior. In addition, (a) intentional agents have ‘intentional states,’ that is, states that represent, or are about, or directed upon things and states of affairs beyond themselves (e.g., beliefs and desires); and (b) these states play a role in generating and guiding purposeful behavior. Because of this role, the behavior of intentional agents is subject to a distinctive kind of explanation, one that appeals to the guiding role of their intentional states. In addition, intentional agents are capable of behavior that is flexible and is not the direct causal effect of their environment or of externally determined motivational states. That is, intentional agents show some independence from their 89

Luca Ferrero

past, their make-up, and their environment. Finally, the conduct of an intentional agent has a source in (and can be attributed to) the agent herself as a whole, rather than just to parts of the agent. Intentional actions are ‘personal’ rather than ‘sub-personal’ phenomena. All of these features, O’Brien argues, are common to intentional agents at large. There is something special, however, about the intentional agency of adult human beings: we both (i) act out of decisions to take a course of action from among two or more courses of action that seem to be open to us, and (ii) act out of our capacity to recognize and respond to reasons as such. Our capacity for normative and reflexive thought, however, does not completely separate us from other kinds of intentional agents. O’Brien argues that there might be important similarities between the first-person perspectives of intentional agents at large: there might be something that it is like for intentional agents to navigate the world and these phenomenal aspects play a role in generating and guiding their actions. These phenomenal aspects might be sufficiently similar across different kinds of intentional agents to make the actions of these agents (such as non-human animals or children) to be at least partially intelligible to us, even if they lack the distinctive rational and reflective character of our specific kind of intentional agency.

Rational agency In ‘Rational agency,’ Eric Marcus presents a ‘transformative’ account of the role of rationality in shaping human agency. Marcus argues that our practical and theoretical capacities constitute a distinctive kind of agency. In exercising these capacities, we bring about a distinct kind of change. Plant agency, for instance, is distinctive in bringing about growth; animal agency is distinctive in bringing about self-movement. What is distinctive of human agency, Marcus argues, is that it brings about a change that is essentially tied to our rational sensitivity to reason. We have the capacity to make up our mind about what to do on the basis of considerations about what we ought to do or to believe something on the basis of considerations about what we ought to believe. We do not simply (and passively) find ourselves acting or believing, as non-human animals do. The exercise of this capacity constitutes the causal connection between the action (or the belief) and the reason on which the action (or the belief) rests. This is what Marcus calls ‘rational causation.’ The causal connection between a reason and the action (or belief) that rests on this reason is nothing over and above the very exercise of our rational powers (the power to make up our minds on the basis of normative consideration about how we ought to make up our minds). Our distinctive kind of agency, then, is rational agency: the power to decide normative questions in such a way as to constitute the causes (or causal explanations) of certain actions (or beliefs). This kind of agency explains the essential tie between the paradigmatic exercises of rational powers and certain causal phenomena, such as the performance of an action for reasons or the adoption of a belief for reasons. In the paradigmatic cases, for a rational agent to do x is nothing over and above this agent’s viewing that action as to be done (likewise, for a rational subject to believe that p is nothing over and above this subject’s viewing that belief as to be had). The subject’s judgment that such-and-such is to be done constitutes the fact that that subject is doing ‘such-and-such.’ The action of a rational agent is a manifestation of (as opposed to a mere effect of ) her practical rationality, that is, a manifestation of the agent’s answer to the question ‘What should I do?’ (which, in turn, explains the distinctive kind of epistemic access that we have our own actions and beliefs). According to Marcus, this kind of agency is ‘transformative.’ When we acquire it, our suite of cognitive abilities, our perceptions, our thoughts, and our actions are changed into qualitatively different capacities. This is what makes our agency a genuinely distinctive kind of agency in the mode of ‘rationality.’ 90

7 MINIMAL AGENCY Hans van Hateren

1 Introduction Agency is the capacity to act, and to act is to generate behaviour according to goals that comply with norms. Full-blown agency in humans can involve complex cognitive functions, such as reasoning and forming intentions to act in a certain way. But many other species behave in ways that suggest that they have agency too, even if this is not the full-blown variant (Frankfurt 1978). The question is then whether there is a minimal version of agency, with a minimal set of features, that can still be regarded as a proper form of agency. Such minimal agency (Barandiaran et al. 2009) is also called primitive agency (Burge 2009), natural agency (Di Paolo 2005), biological agency (Kauffman and Clayton 2006), elementary form of agency (van Hateren 2015), and proto-agency (Fulda 2017). This chapter focusses on the agency of living organisms, but agency is often ascribed to artefacts as well, such as to robots (see the chapter on ‘Artificial and machine agency’). What are the conditions to be met for the attribution of minimal agency? Ideally, such conditions should indicate which species have agency and which behaviours are acts rather than something else (such as a reflex or some other kind of automatic behaviour, such as sneezing or shivering). Conditions may specify which properties agency must have. Alternatively, they may specify how agency is produced, such as through an underlying mechanism of some sort. These two kinds of conditions are not independent, because one’s understanding of possible mechanisms constrains which properties one is willing to accept as realistic, and how agency is produced determines which properties arise. Nevertheless, it is useful to start with an overview of properties of minimal agency that are typically thought to be important. The next section then discusses theories that aim to explain how agency and its properties are realized. Burge (2009) notes that agency requires that a significant part of an organism’s own movement is not produced by the environment but originates from within the organism instead. Moreover, agency should involve coordinated behaviour by the whole organism, issuing from its central behavioural capacities, not purely from subsystems (a heartbeat, for instance, is not an action). Finally, the coordinated behaviour must fulfil a function for the organism. Thus, agency involves authorship and behavioural freedom (‘originates from within’), individuality (‘whole organism’ and ‘central behavioural capacities’), and goal-directedness

DOI: 10.4324/9780429202131-1091

Hans van Hateren

(‘fulfils a function’) that complies with a norm (‘for the organism’). The latter is a minimal norm—taken as a measure of performance that gauges how adequate a function or purpose is fulfilled—because fulfilling a function is here supposed to be beneficial to the organism (the more beneficial, the better). Similar notions can be found in Kauffman and Clayton (2006), who define an autonomous agent as a system that is able to act on its own behalf. Autonomy here does not mean independence, but merely the capacity to influence the conditions under which the system exists. An autonomous system may still strongly interact with its environment (including other autonomous systems), but it is not just pushed around. Autonomy can then be regarded as a combination of individuality and behavioural freedom. At least some behavioural freedom is regarded by van Hateren (2015) as a key feature of agency, resulting from a capacity to initiate behaviour and to direct what form behaviour takes. According to Barandiaran et al. (2009), agency requires a system that is distinguishable from its environment (individuality), that is doing something by itself in that environment (autonomy), and that does so according to a certain goal or norm (normativity). Dretske (1999) and Moreno and Mossio (2015) note that agency must include the capacity to generate causal effects. Moreover, agency must be based on internal representations of its environment according to Dretske (1999) and Sterelny (2001: Ch. 9). Having representations with a causal role then distinguishes acts from behaviour that is a mere reaction to the environment. Theories of full-blown agency are indeed often based on representations, such as beliefs about the state of the world and representations of goals and norms. However, the term representation is closely associated with the symbolic representation that one finds in human language and thought. This association becomes problematic when one considers the agency of creatures that lack the capacity to use symbols. An alternative term, intentionality, is ambiguous, because it can refer either to intentions (in the sense of purposes or plans) or to aboutness (which conveys the idea that a mental process can be about—pointing to and being directed at—something other than itself ). In order to avoid ambiguity as well as associations with the use of symbols, I will simply use the term ‘directedness’ below to refer to the minimal form of intentionality discussed here. If a (possibly non-mental) process has directedness, it is pointing to something beyond that process. The process can then be viewed as representing that something in a minimal, non-symbolic way. In summary, commonly proposed features of minimal agency are goals, norms, freedom, individuality, directedness, and causal efficacy. The next section explains the main theories of minimal agency and evaluates how they fare with respect to these features.

2 Theories of minimal agency Here is a taxonomy of the main theories of minimal agency (in the spirit of this introductory chapter, only a subset of the studies could be cited, and the views are only presented in outline).

2.1 Explanatory role Dennett (1996: Ch. 2) views agency as a property that one can ascribe to a system when its behaviour is best described and explained by assuming that it has such a property. How a system is described depends on one’s stance. Simple physical systems are best described mechanistically in terms of physical components and processes, from a physical stance. However, biological systems and artefacts can be so complex that they have no intelligible physical 92

Minimal agency

description and explanation. A better explanation is available by switching either to the design stance (by viewing a system as, or as if, designed for a purpose) or the intentional stance (by viewing a system as exhibiting intentionality in the sense of aboutness). The latter stance allows one to ascribe directedness, goals, and norms to a system having agency. Individuality can be ascribed if the system appears to be seeking its own good. Importantly, the properties attributed by assuming the intentional stance are explanatory tools that are ultimately only in the eye of the beholder, as the system remains fully mechanistic. But as long as such properties provide good explanations, they are real enough to be taken seriously. One problem with Dennett’s approach is that it is rather lax, as it ascribes agency to more systems than seems plausible. Fulda (2017) proposes a stricter approach by granting an explanatory role to agency only if such a role is indispensable. As in Dennett’s theory, agency is believed to be realized mechanistically, where agency refers to dynamical patterns of behaviour rather than to the mechanisms that underpin those patterns. A similar adjustment of Dennett’s theory is proposed (but without endorsing the explanatory role approach) by Steward (2012: Ch. 4.6), with agency granted only when it is necessary to take at least a teleological stance. A teleological explanation gives an explanatory role to goal-directedness. A necessary teleological stance then means that such a stance is required for ascribing goal-directedness to agency. There are major problems with the ‘explanatory role’ approach to agency. Having an indispensable role in theories does not imply having an indispensable role in Nature. According to the approach, agency has no independent causal role in the world, because all of its causal power is assumed to derive from that of its micro-parts and their interactions. This means that agency has a rather weak ontological status. Agency, goals, norms, individuality, and directedness exist as theoretical constructs, but they are, strictly speaking, absent from Nature itself. The corresponding patterns are real, but they play no role—causal or otherwise—in Nature at all.

2.2 Intrinsic power Neo-Aristotelian approaches to agency build on an ontology of intrinsic causal powers inspired by Aristotle. The main idea is that the causal capacities of biological agents cannot be fully explained by chains or networks of regular physical causation. In addition, such agents must have a qualitatively different form of causation that originates within the agent (immanent causation, see Oderberg 2013) and that has a built-in goal-directedness (an immanent teleology, which is related to Aristotle’s final cause). According to Aristotle, such goal-directedness serves the norm of an increased flourishing of the biological agent itself. The approach produces major properties of agency: goal-directedness, normativity, individuality, freedom, and causal efficacy. The theory has been applied mainly to human agency but also to animal agency (Breidenbach 2018). When the approach is applied to minimal agency, a fundamental weakness is revealed. It assumes a special causal power distinct from the causation produced by the physical and chemical processes that work in biological cells and organisms. A special causal power not produced by physico-chemical processes would, in effect, correspond to a special life force, an élan vital. Such a life force was a feature of the doctrine of (neo-)vitalism in the early 20th century. Vitalism became untenable when research in molecular and cellular biology during the 20th century showed that there is no evidence of extraordinary forces in biological organisms. Thus, the ontology of agency would be utterly mysterious if agency were indeed a distinct causal power in the Aristotelian sense. 93

Hans van Hateren

Agency as an Aristotelian causal power is not compatible with naturalism as the idea that everything in reality is produced by the entities, interactions, and events that are studied by the basic natural sciences (in particular physics and chemistry). ‘Produced’ is an ontological notion, which refers to reality as it is and as it is being realized, not necessarily to reality as it is described, explained, and represented by various theories. Thus, naturalism does not imply epistemological reducibility, that is, it does not imply that representations of complex entities and interactions can be formulated in terms of representations of more basic entities and interactions. Even ontological reducibility is not guaranteed (see the 5th paragraph of the section ‘Fitness estimation’). Moreover, naturalism does not require that all of reality be fully produced, in a determinate way, according to laws. All it demands is that all entities and causal powers that one allows in one’s ontology be embedded in reality in a way that is consistent with—and not disconnected from—the existing body of basic scientific knowledge. Agency as an Aristotelian causal power is too disconnected from that body of knowledge to qualify as naturalistic.

2.3 Evolutionary etiology A major property of agency is its goal-directedness. An act is not a random or imposed behaviour, but it is performed in pursuit of an end generated within the agent. Goal-directedness does not exist in abiotic entities, other than as metaphorically ascribed. Water tends to flow from higher to lower places, but that is not a sign of goal-directedness. It is merely something that happens. Although there appears to be a preferred final state—all the water being collected at the lower place—this is produced purely by momentary, mechanistically driven changes of state. The end state is not foreseen, expected, or anticipated by the water. There is no teleology, because a goal is not required as part of the explanation of why things happen as they do. This is different from the acting of a biological agent. For example, a cheetah clearly anticipates, in one way or another, the preferred final state when it starts to chase a gazelle. The goal explains the chase and it presumably participates in the process that causes the chase. One way to capture the goal-directedness of biological agency is through the goaldirectedness one can ascribe to biological functions. Specific traits of an organism can have a function for the organism. For example, the function of the heart is to pump blood. There is a goal, pumping blood, as well as a norm, because pumping blood benefits the organism’s viability. That pumping blood is beneficial was established by past evolution: organisms with poorly functioning hearts (because of inherited properties) were less likely to survive and reproduce than organisms with better hearts. Hence, current organisms are most likely descendants of the latter group, and therefore usually have well-functioning hearts. A malfunctioning heart can then be said to fail, because it does not do what hearts were selected for in past evolution. ‘Selected for’ is a metaphor that summarizes the result of differing propensities to survive and to reproduce (as produced, in this example, by differing hearts). This approach to understanding the teleology of functions is called the etiological theory of function (see Millikan 1989; Neander 1991). Specifically, the theory depends on an evolutionary etiology (i.e., prior causal history). Goals are associated with functions, where functions and their normativity are produced by past evolution. Just like the heart, agency can be regarded as a biological trait. According to etiological theory, agency has a function for an organism if having agency increased the chances that the organism’s ancestors survived and reproduced in the past. Agency is a capacity, which can only be causally efficacious through the acts that it produces. Therefore, agency could 94

Minimal agency

only be selected in past evolution if the acts that it produced increased the chances of surviving and reproducing, at least on average. The latter condition, ‘on average,’ is important. Some acts are beneficial, others neutral, and still others detrimental. The key point here is that the overall effect of acting is required to have been positive in a statistical sense. If the overall effect had not been positive, agency could not have evolved. As for the heart, the goal-directedness and normativity of agency are produced by past evolution. Theories based on evolutionary etiology consider causation in the past, but they cannot deal well with causation in the present. This becomes clear when we consider the acts that are produced by agency in the present. Such acts have no evolutionary history. Agency has a history, but not the specific acts that it produces in the here and now. How can such acts then be goal-directed and normative? One might postulate that the teleology and normativity of agency is automatically transferred to its acts. That would be defensible if agency would always produce the same stereotyped behaviours. But a key property of agency is that it produces at least some behavioural freedom. Acts are not stereotyped, at least not completely. Therefore, etiological theories cannot fully explain the goal-directedness and normativity of acting. More generally, the fundamental problem of etiology is that a history of causation is an epistemic account, which cannot produce, in an ontic way, the causal efficacy of agency in the present. Causal efficacy depends only on the current state of the organism, not on knowledge of the history of states of its ancestors. Thus, etiologically explained agency lacks actuality, that is, it has no ontological status in the present. Of course, one could ascribe such an ontology ad hoc, but then the theory collapses into an ‘explanatory role’ theory.

2.4 Organizational theory The organizational approach emphasizes that agency is a property of the whole organism, requiring an analysis of how the organism’s structures are organized. Much of the approach derives from the theory of autopoiesis (‘self-production’) of living organisms (Varela et al. 1974). A characteristic property of living organisms is that they maintain themselves, by collecting the energy and materials they need and by continually rebuilding and repairing their own structures. Importantly, an organism has an overall structure that is characterized by operational closure, that is, the organism operates in a way that produces and maintains the conditions for its own viability (Varela et al. 1974; Barandiaran et al. 2009; Moreno and Mossio 2015). The organism has to do so in a resilient, adaptive way (Di Paolo 2005), because environmental conditions usually vary over its lifetime. Adaptive behaviour that regulates how the organism engages with its environment can then be regarded as acts produced by agency. Organizational theories can explain several properties of agency. The organism has goals that serve self-maintenance, and maintaining its own viability conditions is a norm (Barandiaran et al. 2009). Operational closure distinguishes the organism from its environment, thus producing individuality. And the self-directed structure of the system produces an ontology that is self-generated. However, organizational theories have weaknesses too. One is that they do not generate behavioural freedom. Both non-adaptive and adaptive autopoietic systems are fully deterministic. Hence, it might be argued that they cannot initiate and direct behaviour with the kind of freedom required by agency. They just have to proceed automatically with the many chains of causes and effects that run through the causal network formed by organism and environment. An even bigger problem is that the norm is left groundless. The norm of organizational theories is based on the fact that the systems are structured such that they continue 95

Hans van Hateren

their own existence. They fail when they cease to exist in a preventable way. In other words, the basic norm that is implicitly assumed here is that existing is good. But this claim appears to be groundless. If it is based on the observation that the system’s organization explicitly produces continued existence, then one tries to derive a norm from what a system automatically does, which is a non-starter. Alternative, one might assume that existing is good in itself. However, existing is neutral in abiotic nature. Existing or not existing is irrelevant to a rock. In contrast, existing is clearly not neutral to humans, and presumably it is neither to other life forms. But that requires explanation, not an implicit assumption.

2.5 Fitness estimation The above theories are all designed for a reality that is fully governed by deterministic laws, or at least a reality where only such laws are relevant for causation. But physics tells us that there is indeterminacy in the world at small scales, where quantum events are due to chance. Moreover, many systems have inherent instabilities by which even infinitesimal indeterminacies—such as those arising from quantum effects or external disturbances—can produce large, random effects (examples are the weather and the solar system, Laskar and Gastineau 2009). Importantly, biological systems are known to display significant amounts of randomness—also called noise—at all scales (from the molecular up to the neural and behavioural levels, Faisal et al. 2008). Reality is neither fully determinate nor fully random, but a mixture of the two. Random events can produce significant causal effects, for example when a random mutation is produced in DNA when it is hit by a high-energy photon. Although the specific outcome of a random event has no sufficient cause, it can start a fresh chain of causes and effects. Randomness is generally believed to be incapable of being fully embedded in causal mechanisms, in the sense of not only being a sufficient cause but also being sufficiently caused itself. Yet, there is a major exception to this rule, a specific non-deterministic causal mechanism that can produce agency (van Hateren 2015). This is briefly explained next, omitting details and caveats. Random is not the same as arbitrary. A random variable, such as the potential outcome of a die throw, is still constrained by its probability distribution (which constrains the possible outcomes, 1 to 6 for a die, as well as their probabilities, 1/6 each if the die is fair). The distribution is fixed for a die, but there are physical processes where the distribution is variable and controllable. As a simple (artificial) example, one might systematically vary the distance between a radioactive source and a series of plants, each exposed for a fixed time. Then the width of the probability distribution of mutational change varies as well: the higher the dose, the more variation and the broader the distribution. More relevant to the present topic is that also regular molecular and neural processes within organisms can produce random structural variation (because such processes are partly noisy), with distributions that can be controlled to some extent. Thus, even though the outcome of individual random events is uncaused, this is not true of probability distributions. This allows randomness to become fully embedded in a causal mechanism, as follows. Suppose that the probability distribution of an organism’s behavioural dispositions is controlled by a variable produced within the organism. This variable, say x, is assumed to be a quantitative measurement of a behaviour-dependent property of the world (more on that later). Suppose further that small x causes large variability of behavioural dispositions and large x causes small variability (with variability depending in a similar way on intermediate x). This means that when x is large, only small changes of the organism’s behavioural state 96

Minimal agency

occur (because large x means small variability of dispositions and subsequent behaviour). Then x is likely to remain close to the value it has, which is large. In contrast, when x is small, the resulting large variability of behavioural dispositions tends to produce large changes of the organism’s behavioural state. This is likely to affect x, making it smaller or larger. Eventually, such changes will happen to produce a state with large x, which then automatically reduces subsequent behavioural variability, thus keeping x large for a while. In effect, this mechanism drives the organism towards states with large x, because the organism is driven away from states with small x (because of the large behavioural variability there). Thus, the mechanism displays an inherent goal-directedness, with large x as the goal. Importantly, the goal-directedness is not explicitly built in, as a deterministic regularity, but emerges from modulating the statistics of randomly varying behavioural dispositions. The emergence is ontological: it is ‘out there’ in the world and not just a consequence of how theories represent the world. Moreover, it is a strong kind of emergence, because the emergent causal efficacy of the mechanism is not fully produced by the determinate causation of its parts and their interactions. Instead, its causal efficacy comes partly from slowly accumulating effects produced by randomness. Therefore, the causal mechanism is non-deterministic (i.e., neither determinate nor indeterminate). Nevertheless, it is still fully naturalistic, because randomness is naturalistic, even though the specific outcome of individual random events is not determined by laws. The mechanism produces ontological irreducibility, because randomness lacks a determinate ontology: before a random event occurs (say a die throw), it does not yet exist, whereas when the event is realized (say showing 3), it ceases to be random. Large x is, thus, a strongly emergent goal of an organism that has this mechanism. It complies with the strongly emergent norm that the larger the x, the more adequate it is. But what property could x measure? In principle, it could be anything influenced by behaviour. However, organisms have to struggle for resources to live and procreate, thus x better be a goal that is beneficial to the organism’s propensity to survive and reproduce (which is a basic definition of evolutionary fitness). In fact, the optimal x is a measurement of an organism’s own evolutionary fitness. Then high x as a goal is guaranteed to serve actual fitness better than any other goal. Any other type of x would be unstable in the long run. Moreover, organisms with this mechanism can be shown to outcompete organisms without it. Evolutionary fitness is not a simple variable that can be readily measured by an organism. It has to be assessed by combining many indicators of fitness, such as about an organism’s internal state (e.g., health), its environment (e.g., availability of food), and its social conditions (e.g., presence of potential mates). Fitness can only be assessed approximately. Then ‘measurement’ is not the right term, because that is usually used for assessing simple variables in a straightforward way. When considerable (implicit) inference is involved, the proper technical term is ‘estimation.’ Thus, x must have evolved to be an internal estimate of an organism’s own fitness. Although large x is the overall goal of an organism, in practice this will consist of many sub-goals that each contribute to large x. Acts are then behaviours produced by behavioural dispositions corresponding to sub-goals. Agency is an aspect of the internal process that produces x and that modulates the variability of behavioural dispositions. Agency has goal-directedness (because large x is an inherent goal) complying with a norm (the larger x, the better). Although the norm has emerged internally, it is not arbitrarily related to reality, because evolutionary pressure keeps x close to being an estimate of the actual fitness of the organism (as it did for its ancestors and as it will do for successful descendants, if any). Evolutionary pressure is always present, because it involves not only reproduction, but also staying alive. 97

Hans van Hateren

Other properties of minimal agency are present as well. Behavioural freedom follows from the fact that the mechanism is non-deterministic. Behavioural dispositions are produced by a determinate x that drives the statistics of randomness. As a result, determinacy and randomness cannot be separated here. Therefore, the much-professed dichotomy that behaviour is either determinate or random (as an argument against free acts) is false: an act is neither. Individuality follows from the fact that x provides unity to an organism. This is so because x estimates fitness, which quantifies how well the organism can survive and reproduce. Both surviving and reproducing concern the organism as a unit, not as a collection of parts. Directedness follows from the fact that x is an estimate, pointing to something other than itself. This applies to sub-goals of x as well. Finally, the causal efficacy of agency is distinct and independent, because it is strongly emergent. Despite this explanatory success, fitness estimation theory has a weakness too. The mechanism is conjectural, as there is no direct empirical evidence yet that it is present in organisms. There is circumstantial evidence and all parts of the mechanism have plausible physiological counterparts. But direct empirical investigations are required in order to test whether it exists in specific species, and if so, how exactly. Such investigations are lacking at the time of writing.

2.6 Summary of theories Table 7.1 summarizes the theories that are discussed above.

3 Beyond minimal agency How does minimal agency relate to full-blown agency? Several basic properties of fullblown agency are already present in minimal agency, in particular, minimal forms of teleology, normativity, freedom, individuality, representation, and causal efficacy. But others are missing, such as consciousness, symbolic representation, social and cultural influences on norms, and social extensions of agency. I believe that there are well-defined paths to get to Table 7.1 Features of theories of minimal agency Explanatory role Ontology of Ascribed agency Source of goals Ascribed Source of norms Source of freedom Source of individuality Source of directedness Dominant weakness

Ascribed None

Intrinsic power

Evolutionary etiology

Organizational theory

Fitness estimation

Mysterious

None

Self-generated

Strongly emergent

Immanent teleology To flourish

Biological functions Past evolution

Ascribed

Immanent causation Intrinsic

Ascribed

Intrinsic

No real role Nonnaturalistic

Self-maintenance Internal fitness estimate To exist High fitness estimate None None Self-modulated randomness Reproduction Operational Estimated fitness closure Biological None Estimation functions Lacks actuality Groundless norm Empirical validation

98

Minimal agency

those properties as well. Fitness estimation theory must then extend fitness beyond its minimal form (i.e., individual survival and reproduction) to include wider factors (such as kin, peers, and culture). This is complex but feasible. Nevertheless, the distance between basic mechanisms and actual agency would then become so large that only ‘in principle’ kinds of explanation could be given. This is still important, because it demystifies agency and its properties. But for more specific understanding, analysis is better done at the appropriate, higher levels of abstraction.

Related topics Agency and causation; Agency, function, and teleology; Animal agency; Artificial and machine agency; Agency, will, and freedom; Agency and autonomy; Agency and normativity.

Further reading Barandiaran, X. E., Di Paolo, E., and Rohde, M. (2009) Defining agency: individuality, normativity, asymmetry, and spatio-temporality in action. Adaptive Behavior 17:367–386. Discusses the properties of minimal agency from the perspective of the leading organizational theory. Burge, T. (2009) Primitive agency and natural norms. Philosophy and Phenomenological Research 79:251–278. Proposes that natural flourishing of living organisms leads to natural functions and natural norms. van Hateren, J. H. (2015) The origin of agency, consciousness, and free will. Phenomenology and the Cognitive Sciences 14:979–1000. Explains the fitness estimation theory of minimal agency and sketches ways to extend it.

References Barandiaran, X. E., Di Paolo, E., and Rohde, M. (2009) Defining agency: individuality, normativity, asymmetry, and spatio-temporality in action. Adaptive Behavior 17: 367–386. Breidenbach, J. C. (2018) Action, animacy, and substance causation. In: W. M. R. Simpson, R. C. Koons, and N. J. Teh (eds.) Neo-Aristotelian Perspectives on Contemporary Science. New York: Routledge, Chapter 10. Burge, T. (2009) Primitive agency and natural norms. Philosophy and Phenomenological Research 79: 251–278. Dennett, D. C. (1996) Kinds of Minds: Toward an Understanding of Consciousness. New York: Basic Books. Di Paolo, E. A. (2005) Autopoiesis, adaptivity, teleology, agency. Phenomenology and the Cognitive Sciences 4: 429–452. Dretske, F. I. (1999) Machines, plants and animals: the origins of agency. Erkenntnis 51: 523–535. Faisal, A. A., Selen, L. P. J., and Wolpert, D. M. (2008) Noise in the nervous system. Nature Reviews Neuroscience 9: 292–303. Frankfurt, H. G. (1978) The problem of action. American Philosophical Quarterly 15: 157–162. Fulda, F. C. (2017) Natural agency: the case of bacterial cognition. Journal of the American Philosophical Association 3: 69–90. Kauffman, S., and Clayton, P. (2006) On emergence, agency, and organization. Biology and Philosophy 21: 501–521. Laskar, J., and Gastineau, M. (2009) Existence of collisional trajectories of Mercury, Mars and Venus with the Earth. Nature 459: 817–819. Millikan, R. G. (1989) In defense of proper functions. Philosophy of Science 56: 288–302. Moreno, A., and Mossio, M. (2015) Biological Autonomy: A Philosophical and Theoretical Enquiry. Dordrecht: Springer. Neander, K. (1991) The teleological notion of ‘function.’ Australasian Journal of Philosophy 69: 454–468. Oderberg, D. S. (2013) Synthetic life and the bruteness of immanent causation. In: E. Feser (ed.) Aristotle on Method and Metaphysics. Basingstoke: Palgrave Macmillan: 206–235. Sterelny, K. (2001) The Evolution of Agency and Other Essays. Cambridge: Cambridge University Press.

99

Hans van Hateren Steward, H. (2012) A Metaphysics for Freedom. Oxford: Oxford University Press. van Hateren, J. H. (2015) The origin of agency, consciousness, and free will. Phenomenology and the Cognitive Sciences 14: 979–1000. Varela, F. G., Maturana, H. R., and Uribe, R. (1974) Autopoiesis: the organization of living systems, its characterization and a model. BioSystems 5: 187–196.

100

8 ANIMAL AGENCY Helen Steward

Do any non-human animals have agency? The answer, one might think, unless one is using the term ‘agency’ in an unusually inflated sense, is obviously ‘yes’. Agency is the power to act – and surely animals act in various ways? Some animals hunt; some build dams; some collect and store food for the winter. Many animals play, and fight, and tend to their young. Some groom one another; some engage in complex courtship rituals. Some can be trained to help with human tasks such as rounding up sheep (dogs) or retrieving fish (cormorants); ‘chimpanzees, arguably’, the most intelligent of all, have even been trained to communicate quite complex messages by use of a rudimentary sign language. In view of these facts, it is surprising, then, that animals have figured so little in the philosophical discussions of the nature of agency which have taken place in the last century or so, and indeed that their status as agents continues to be denied (e.g. Stoecker, 2009). Of course, human agency is distinctive in numerous respects – many of which are associated with other important philosophical concerns, for example, with practical reason, with morality, with self-control – which helps to explain the focus of philosophers on specifically human action. But as I shall make clear below, the neglect of animal agency by philosophers has not only prevented us from having a clear view of the animal phenomenon but also obscured our view of the metaphysics of the human one as well. For we humans – let it not be forgotten – are animals. Human agency is, first and foremost, a species of animal agency (though admittedly a very special one) and attempts to say what agency is which entirely ignore that fact are likely to be doomed to failure.

1 The causal theory of action According to what has arguably been the dominant theory of action over the past 50 years or so, the causal theory of action, or CTA, actions are merely events, generally speaking, bodily movements, such as my right arm’s rising or my left eye’s closing. What marks actions out from events which are not actions, on this view, is a distinctive sort of causal provenance. My arm’s rising counts as an action if and only if it is brought about by the right sorts of things and in the right sort of way1 – for example, perhaps if it is brought about by my desire that my arm should rise (which perhaps is, in turn, brought about by my desire to ask a question in a seminar, together with my belief that I can signal my desire to ask a question by raising my arm), these beliefs and desires together constituting the reasons for which I act (Davidson, 1963). DOI: 10.4324/9780429202131-11101

Helen Steward

Other philosophers have adopted slightly different, but still basically causal/functionalist strategies, in which the distinctive precursors of actions are not beliefs and desires, but rather other sorts of mental state and event, for example, intentions (Bratman, 1987; Mele, 1992) or volitions (Mill, 1970[1872]; Locke, 1975[1689]). Fundamentally, though, the idea behind all these views of action is the ultimately dualistic one that an action is essentially a physical thing which is produced by a mental thing, or set of mental things (or perhaps the ultimately not very different view that we should think of an action as a complex entity composed of both the mental cause(s) and the physical effect together). The question whether animals act, then, given a theoretical background of this kind, becomes the question whether the purposive bodily movements which we undoubtedly observe taking place in their limbs, wings, legs, fins, and so on have the appropriate sorts of mental cause. Do these bodily movements have the appropriate sorts of mental cause? This is a difficult question to answer. For there have always been many people who have doubted the wisdom of ascribing mental states to animals. Some, for example, have argued that animals cannot have beliefs (Davidson, 1975, 1982; Stich, 1979). Beliefs are propositional attitudes, ascribed by way of conceptual content – but how do we know whether, for example, to ascribe a dog the belief that the cat is up the tree when it isn’t at all obvious that the dog can distinguish cats from a range of other small animals that it also likes to chase (e.g. squirrels), or whether it makes any distinction between trees and bushes (to use an example taken from Stich)? And if we have no idea how to provide content for beliefs of a sort that would fit the doggy mind, how sensible is it to think of dogs as having beliefs at all? It might seem less easy to see what is wrong with supposing that animals have intentions, but doubt here has also been expressed. One might wonder, for example, whether we can really be sure that animals know what they are doing in the way that is arguably required to justify the attribution of intentions to them. Birds certainly behave purposively when they build nests, for example, choosing moss and grasses carefully and weaving the whole together so as to produce astonishingly well-crafted receptacles for their eggs and for the subsequent rearing of chicks. But do they have any idea why they are building the nests? Do they have any conception of what is to take place, shortly, within them, particularly during the first year of breeding, when they have had no previous experience? Or is the whole process merely instinctive, done without any sense of the end to which the nest is a means? And if the answer is that the process is instinctive and comes with no knowledge or understanding of its point or purpose, can it be correct to suppose that the bird really intends to build a nest? One person’s modus ponens, however, is another person’s modus tollens. If a certain philosophical view of action leads to the conclusion that animals can’t act, perhaps it is that view of action that should be questioned. Harry Frankfurt is one of the most influential philosophers to have pointed out how wrong the conclusion that animals aren’t agents seems. Frankfurt (1978) discusses the difference between what happens when a spider moves its legs as it travels across the ground, and what goes on when its legs move in exactly the same way because they are being manipulated by a child who has attached strings to them and is moving the spider like a puppeteer. Frankfurt notes that in the former case, unlike the latter, the movements are attributable to the spider as the agent of the movements, and that this would seem to be just the very same distinction as the distinction between a person’s actively raising their arm and a person’s arm going up because someone else has raised it. Spiders, then, would seem to be agents too. Moreover, it seems that we can know that spiders are agents in this sense, without knowing anything about whether they have mental states worth calling beliefs or intentions. Even if agency does demand the existence of certain forms of mentality, might it not be overkill to assume that it requires these comparatively sophisticated states? 102

Animal agency

Perhaps if animals can, for example, see certain objects, want certain things, and try to get them (as might seem much more plausible than that they have beliefs or intentions), these capacities would be sufficient to constitute appropriate precursors for exercises of something we could recognise as agency? What kind of theory of action might make possible such an extension of agency to animals? One possibility, as suggested above, might be simply to downgrade the sophistication of the mental states required to serve as causes of action to things which might be simpler than beliefs and intentions. Another, more radical suggestion might be that we should try a different way of privileging as definitional the feature of agency to which Frankfurt’s argument draws attention – the fact that we attribute certain of the movements which go on in the world to agents. Our dualistic heritage encourages to think about the production of effects in the world by agents as though that were equivalent to the production of effects in the world by minds, by mental states, and by events – but this may not be the only way in which one might be able to think about what it means to attribute some happening to an agent as its source. The idea of an agent as a thing which has a body, rather than just being a body, is an interesting one – and even if a thing’s having a body entails that it has some mental capacities (as seems likely), it might not follow that each occasion on which an agent controls its body in the way we think distinctive of action is an occasion on which the agent’s mental states are directly involved in the exercise of that control. To see this, it helps to think about types of human activity which are a bit less deliberative and intentional than the sorts of examples that philosophers tend mainly to choose – in particular, it can be useful to consider what have sometimes been called ‘sub-intentional actions’, such as absent-mindedly scratching an itch or distractedly fiddling with one’s jewellery. Are these actions? I think so. Like the movements of the legs of Frankfurt’s spider, they are due to the agent, even if they are not intended, or even wanted by that agent. But if they are actions, we are going to need a different and less mentalistic way of thinking about what is essential to agency than the one which the CTA and its descendants have suggested to us, one which might, in turn, make it seem much more straightforward to accord agency to at least some animals.2

2 Which animals? As that last phrase ‘at least some’ suggests, though, it is not yet clear how widely even a more generous and flexible conception of agency can be extended across the animal kingdom. Are all animals agents? One might worry that extending the attribution too widely risks losing sight of the nature of the distinctive phenomenon we are trying to capture. The movements of some very simple creatures, for example, seem explicable as fairly simple mechanisms, and one might think that the attribution of agency is excessive in such cases. Take a paramecium, for example. A paramecium is a single-celled creature which moves through water by means of cilia on the outside of its body. But is it really right to say that the paramecium is beating its cilia? One reason for thinking the answer might be ‘no’ might be the thought that no consciousness, no thought, no real mentality could really be supported by a single cell. Another, though, might relate to the idea that in a case like this, an entirely mechanistic explanation for the relevant motions threatens to make the invocation of agency redundant. There is detailed research, for example, showing that the direction of movement of a paramecium depends on a fairly small number of factors, including water temperature and pH value (Glaser, 1924; Chase and Glaser, 1930). I think it would not be surprising if it were to turn out that the motions of paramecia are pretty much entirely predictable given knowledge only of a fairly manageable number of environmental variables. But if this did turn out to be the case, 103

Helen Steward

one might think that the paramecium seems to fall into the same category as a robot or a self-driving car, something whose movements are settled by a complex interaction between its (in this case, biological) programme and the environment, but perhaps not one we need to think of as agential. This raises the important question, though, of what creature-environment interactions have to be like in order to warrant the attribution of agency to a creature. Some philosophers would doubtless argue that the movements of all animals (including even those of humans) are also settled by complex interactions of a basically mechanistic sort, differing from those we can observe in the paramecium only in degree of complexity. One interesting and extremely influential manifestation of this view is encapsulated in Daniel Dennett’s idea that we adopt an ‘intentional stance’ in order to explain the behaviour of certain kinds of systems – humans, at least some animals, and potentially also some artificial systems, such as chess computers and thermostats – and that there is nothing more, or less, to the question whether a system has intentional states like beliefs and desires than there is to the question how well the intentional stance works to explain and predict the behaviour of those systems (Dennett, 1987). Dennett himself is mainly interested in the question what is required for the ascription of intentional mental states, rather than in our question, which is what agency requires. But one can see that a similar strategy might work in the case of agency too. Perhaps agents are those things which it proves fruitful to treat as agents – that is to say, with respect to which it pays off to consider what informational state the creature may be in and what its goals are, in order to decide what it may be about to do. Of course, as Dennett himself points out, one can in a sense treat almost anything as an agent – for example, he notes that it would be possible to predict the behaviour of a lectern by ascribing it the belief that it is currently located at the centre of the civilised world and desiring above all else to remain at that centre. We’d then predict that it wouldn’t do anything – and hey presto! – we’d be right. But in this case, as Dennett says, ‘we get no predictive power from [the intentional strategy] that we did not antecedently have. We already knew what the lectern was going to do – namely nothing’ (p. 23). Perhaps, then, the mark of a true agent is that what one might call the agential strategy, a strategy which assigns the system or creature in question such things as perceptual knowledge and a range of goals, gives us predictive power that we did not have available before. It is not that we could not in principle predict the motions of an agent from what Dennett calls ‘the physical stance’, a stance which involves determining the physical constitution of a system and the physical nature of the impingements upon it and using the laws of nature to predict the result. But that would be an immensely impracticable task, because the calculations would be so astronomically large and complex. Where agents are concerned, the agential strategy simplifies matters for the purposes of explanation and prediction. This Dennettian view is designed to be compatible with the view that, ultimately, everything is mechanism – and if there is any kind of differentiation to be made amongst the various varieties of mechanism, the differences will be matters of degree, not of kind. Because the idea is that whether or not a system is an agent is a matter of how fruitful it is for prediction and explanation to apply the agential stance, the line between the agents and the non-agents will not be sharp. An alternative view, though, might be that agency is distinctive, in part, because of its modal features – that an exercise of agency is always such that it does not have to happen, at least not in quite the way or at quite the time that it does happen, so that agency is an essentially indeterministic phenomenon. If the motions of the paramecium are all strictly determined by such things as the temperature and pH of the water, then intuitively, nothing really remains to be settled by the paramecium – and this might be thought to justify the view that the paramecium does not count as an agent in respect of those motions – any more than 104

Animal agency

I count as an agent of my motions when my unconscious body is flung from the top of a building and I fall to the ground. But when a cow wanders across a field, say, perhaps we might be more inclined to believe that it makes motions happen in its body that need not have happened in precisely those ways and at precisely those times, even holding fixed all prior circumstances. Of course, animals act in accordance with their instincts – and so there are limits to what any given animal might possibly do on any given occasion. Cows are not free to decide not to graze, for example. But where we attribute agency, this view would have it that we attribute at the same time a certain freedom – a movement only really being worth attributing to a creature (rather than, say, to antecedent events and interactions going on in its parts) if the creature has the power to settle something of what results. Of course, much more needs to be said about what kind of metaphysical picture of the world could allow for such settling – how certain results can obtain neither because they are determined to do so nor by chance. But if such a metaphysical picture is available, this view might hold out the promise of respecting Frankfurt’s point that we need an account of agency which explains properly what it means to attribute some movement to a creature, and which does not simply identify that with attributing the movement to certain mental states and events.3 Even if some animals do have agency, it has to be admitted, of course, that human agency is a distinctive phenomenon. In the next two sections, I want to turn to look at two important potential differences between animal and human agency.

3 Rationality Are animals rational? That depends, of course, what is meant by ‘rational’ – a matter on which philosophers do not agree. It seems fairly clear that animals can do certain sorts of reasoning task. Some crows, for example, can solve puzzles of quite considerable complexity to release tools by means of which to extract pieces of food which they can see, but which are inaccessible except by way of those tools.4 Some have suggested that since animals can reason, since they seem to be able to make inferences, it settles the question in favour of their rationality. But this is not really clear. The concept of rationality has a long history in philosophy – and one very powerful conception of rationality connects rationality with normativity. As Robert Brandom puts it, ‘To be a rational being in this sense is to be subject to a distinctive kind of normative appraisal: assessment of the reasons for what one does … Rational beings are ones that ought to have reasons for what they do and ought to act as they have reason to’ (Brandom, 2009, pp. 2–3). Ought animals to have reasons for what they do? It seems quite natural to suppose that sometimes at least some animals do have reasons for what they do – that a cow might run at a dog because she wants to protect her young. But in many ways, it seems strange to bring animals within the fold of our ‘ought’ judgements. We can perhaps make sense of a quasi-normative ‘ought’ attached to an imputed teleology: ‘if that nest is to stay in the tree, the blackbird ought to choose a stronger branch’; ‘the salmon ought to swim up the fish ladder if they’re to reach their spawning grounds’, and so forth. There are better and worse ways for animals to achieve what we take to be their ends, and so we can make some kind of sense of the idea that an animal ought to do something in one way rather than another. But other parts of the structure that generally pervades the ‘space of reasons’ (Sellars, 1956) are manifestly lacking in the case of non-human animals. Animals are not answerable to anyone for their mistakes; we do not (because we cannot) demand justifications, excuses, or explanations from them. One can train a dog but arguably one cannot get it to see that there are better ways for it to do things. 105

Helen Steward

One line of thinking which I find helpful when considering the question whether any non-human animal is rational draws on what can reasonably be supposed to be an Aristotelian thought. Aristotle is well known for his claim that human beings are rational animals – and moreover that it is in some sense part of the essence of a human being to be rational. But of course, if this idea is not to fall at the first hurdle, it needs to be explained why it is not disproved by the existence of manifestly non-rational human beings – humans, for example, who are unable to speak or think and who do not show any evidence of the powers associated with reason or reflection. One response is that the Aristotelian account does not intend to assert that rationality is part of the essence of any individual human beings; it is rather part of the essence of the kind, human being, in that in some sense or other it characterises what human beings are like when development follows a certain teleologically structured path in the way it is supposed to. Moreover, Aristotle clearly thought that ‘rational’ was a special differentia of the genus ‘animal’ – unlike ‘mammal’ or ‘two-footed’, which arguably are also parts of the essence of the kind human being (Boyle, 2012). But in what way special? Boyle argues, to my mind very convincingly, that rationality was thought of by Aristotle not merely as a power or collection of powers which is possessed uniquely by human beings, but rather as a distinctive form of psyche, or soul, as important and as fundamental a kind of soul as the Aristotelian vegetative and animal souls. Just as animality might be thought of as a way of being which fundamentally builds on, but also utterly transforms, the kind of life found in plants, so he suggests, rationality can be thought of as a way of being which fundamentally builds on, but also utterly transforms, the nature of the animal life to which it belongs. Different participants in the debate about animal rationality, then, may simply be talking past one another. Some participants think of rationality as an individual power, or perhaps a connected set of powers, which can be possessed to a greater or lesser degree – and according to this conception, there can be no very hard and fast divides between the rational and the non-rational, and everything argues for respecting the continuities in nature. For others, though, to call a being ‘rational’ is not merely to indicate that it is possessed of particular powers, but rather to indicate that the whole form of its life is of a specially transformed kind – a kind which involves the ability to respond to reasons as such (McDowell, 1994), to justify one’s actions, and to operate on principles.

4 Moral responsibility Another interesting (and related) question is whether animals can ever be morally responsible for what they do. Although I suspect most of us think the answer to this question is ‘no’, it is harder than one might think to justify this answer. One traditional route to the view that animals are not morally responsible agents, of course, goes via the idea that animals are not agents at all – that they have none of the requisite freedom to do otherwise – but as I have tried to suggest above, and have argued in more detail elsewhere (Steward, 2012), that may not be correct. Moreover, it does not seem to be true that no non-human animals have systems that might count as moral codes. Some animal societies appear to be based upon certain pro-social norms, the violation of which attracts considerable opprobrium, and which members of the animal society make the appearance of having internalised. Moreover, some animals seem to be motivated by what might be argued to be moral reasons – they feel sad, for example, at the suffering of others and make attempts to ameliorate it (Rowlands, 2013). It might be pointed out that non-human animals have no capacity to scrutinise or discuss their motivations, but perhaps it might reasonably be replied that there are many perfectly 106

Animal agency

good – even virtuous – human beings who do little or no such scrutinising either. Might it not be overkill, then, to demand a practice of scrutiny in a morally responsible agent? Some animals, then, may be involved in a kind of norm-following which might count as a primitive morality, and some may be responsive to moral reasons. But I think it continues to seem implausible that animals are morally responsible for their actions. It has been suggested that the key factor which distinguishes what we might call (following Rowlands) morally sensitive subjects from morally responsible agents concerns the capacity to intervene with respect to one’s potentially morally sensitive habits and behaviours, in the light of such things as reflection or the acquisition of new information (Rowlands, 2013; Musschenga, 2015). On Rowlands’ view, ‘What demarcates moral subjects from moral agents … is a kind of level of understanding’ (Rowlands, 2013, p. 239). It is a lack of understanding, though, rather than a lack of the capacity for agency, which appears to be the obstacle to the assignment of responsibility. It has long been recognised, of course, that animals should be thought of as moral patients – that because of their capacity to undergo suffering, we have duties to them. I should like to end on the thought that if we are prepared, in addition, to recognise the status of at least some of them as agents (albeit not morally responsible ones), as I believe we should, we might need to give a different and more radical account than is sometimes given of where those duties end. Agents make choices, have ends, and live lives – and prima facie, we have a duty to agents to permit them, so far as is consistent with other reasonable moral demands, to exercise that agency in the course of those lives, in the ways that they wish, or may be presumed to wish. We therefore have prima facie duties not merely to avoid causing animal agents pain but also to refrain from curtailing or interfering unduly with their natural lives. These are duties we have not thus far sufficiently recognised.

Related topics Agency and causation; Minimal agency; Rational agency; Agency and responsibility; Agency and morality

Notes 1 ‘In the right kind of way’ because it is accepted by the proponents as well as the detractors of the CTA that certain sorts of counterexample show conclusively that more is needed for an action to have taken place than merely the causation of the appropriate sort of bodily movement by the appropriate sort of mental cause; see Davidson (1973) for a discussion of the so-called ‘deviant causal chains’ problem as it applies in the case of action. 2 See my (2012) for a detailed discussion of how this might be done by appeal to the idea of topdown causation. 3 See my (2012) for an attempt to develop the wanted metaphysics. 4 See e.g. BBC (2014), accessed 8th January 2019 https://www.wimp.com/a-crow-solves-an-eight-steppuzzle/ to see a crow in action.

Further reading Boyle, Matthew (2012) ‘Essentially rational animals’, in G. Abel and J. Conant (eds.) Rethinking Epistemology. Berlin: Walter de Gruyter Verlag. In this paper, the author attempts to argue the case for thinking that being an animal is fundamentally transformed when the powers distinctive of rationality are present, in a way that justifies the idea that humans are animal agents of an utterly distinctive kind, in many ways non-continuous in virtue of that transformation, with other animal agents.

107

Helen Steward Frankfurt, Harry. (1978) ‘The problem of action’. American Philosophical Quarterly 15: repr. in his The Importance of What We Care About (Cambridge: CUP): 69–79. In this paper, Frankfurt argues against the causal approach to making the distinction between what an agent does and what merely happens to him; and makes out the case for the view that the conditions for attributing the guidance of bodily movements to a whole creature, rather than to a mechanism within one, obtain outside the human arena. Hurley, Susan and Nudds, Matthew (2006). Rational Animals. Oxford: Oxford University Press. A useful collection of articles, which asks interesting questions about what it might mean to attribute rationality to a creature and brings to bear on the issues many empirical studies of the kinds of tasks animals are able to perform and psychological theorising concerning the processes which might underlie them. Rowlands, Mark. (2013) Can Animals Be Moral? Oxford: Oxford University Press. This book argues, against centuries of tradition, that animals can act for moral reasons and thereby count as what Rowlands calls ‘moral subjects’. Steward, Helen. (2012) A Metaphysics for Freedom Oxford: Oxford University Press, Ch 4. This chapter defends the view that many animals are agents, and also considers the difficult question of how can be decided which animals are agents, and why.

References Boyle, M. (2012) “Essentially rational animals,” In G. Abel and J. Conant (eds.) Rethinking Epistemology. Berlin: Walter de Grutyer Verlag. Brandom, R. (2009) Reason in Philosophy. Cambridge: Harvard University Press. Bratman, M. (1987) Intention, Plans and Practical Reason. Cambridge: Harvard University Press. Chase, A. and Glaser, O. (1930) “Forward movement of paramecium as a function of the hydrogen ion concentration,” Journal of General Physiology 13: 627–636. Davidson, D. (1963) “Actions, reasons and causes,” Journal of Philosophy 60: 685–700, repr. in Davidson Essays on Actions and Events. Oxford: Oxford University Press, 1980: 3–19. Davidson, D. (1973) “Freedom to act,” In T. Honderich (ed.) Essays on Freedom of Action. London: Routledge and Kegan Paul, 137–156. Repr. in Davidson (1980): 149–162. Davidson, D. (1975) “Thought and talk,” In S.D. Guttenplan (ed.) Mind and Language. Oxford: Clarendon Press: reprinted in Davidson Inquiries into Truth and Interpretation. Oxford: Oxford University Press. Davidson, D. (1982) “Rational animals,” Dialectica 36(4): 317–327. Dennett, D. (1987) The Intentional Stance. Cambridge: MIT Press. Frankfurt, H. (1978) “The problem of action,” American Philosophical Quarterly 15, repr. in his The Importance of What We Care About (Cambridge: Cambridge University Press): 69–79. Glaser, O. (1924) “Temperature and forward movement of paramecium,” Journal of General Physiology 7: 177–188. Locke, J. (1975)[1689]. An Essay Concerning Human Understanding, ed. P. Nidditch. Oxford: Oxford University Press. McDowell, J. (1994) Mind and World. Cambridge: Harvard University Press. Mele, A. (1992) Springs of Action. Oxford: Oxford University Press. Mill, J.S. (1970)[1872] A System of Logic, 8th ed. London: Longman. Musschenga, Albert (2015) “Moral animals and moral responsibility,” Les ateliers de l’éthique 10(2): 38–59. Rowlands, Mark. (2013) Can Animals Be Moral? Oxford: Oxford University Press. Sellars, Wilfrid S. (1956) “Empiricism and the philosophy of mind” Minnesota Studies in the Philosophy of Science 1: 253–329 Steward, H. (2012) A Metaphysics for Freedom. Oxford: Oxford University Press. Stich, S. (1979) “Do animals have beliefs?” Australasian Journal of Philosophy 57: 15–28. Stoecker, R. (2009) “Why animals can’t act,” Inquiry 52: 255–271.

108

9 INTENTIONAL AGENCY Lilian O’Brien

We have many pre-theoretical ideas about freedom or consciousness or knowledge. But we probably have few, if any, about intentional agency as such. The intentional agents that we consider are usually also persons or moral agents. And it is the features that qualify them for these categories that tend to catch our eye. And within the philosophy of agency there is more energy devoted to understanding intentional actions than to understanding intentional agency. But what intentional agents are deserves our attention. Within the philosophy of agency, controversies about intentional action sometimes turn on claims about what role the agent must play in intentional action and whether intentional agency can be captured in, say, event-causal terms (e.g. Mele 2003, Ch. 10). Outside of the philosophy of agency, questions about intentional agents arise in many contexts. Advances in artificial intelligence have given rise to machines that produce complex behaviour. But are they intentional agents? And for that matter, if we, the adult humans, our children, and our pets, are intentional agents, what qualifies us for this? Without an understanding of the concept of intentional agent and the category of things it delineates, we are at a loss in addressing important questions concerning our similarities to, and differences from, co-inhabitants of our social world.

1 Preliminary profile: purposefulness, intentional states, explanation When we think of a human agent building a house or a cat stalking prey or a machine assembling another machine, the behavioural outputs of these things are, we might ordinarily say, purposeful. I will assume that an intentional agent is something that has capacities for producing purposeful behaviour. What is purposeful behaviour? It is not promising to equate purposeful behaviour with an agent having a purpose in mind in directing that behaviour. Someone who scratches out the eyes in a photograph of a rival or who whistles a tune while they work seems to act purposefully without having a purpose in mind in acting thus (Hursthouse 1991; Mele 1988). Paradigm cases of purposeful behaviour have structure – purposeful behaviour progresses, not randomly or haphazardly, but along a trajectory, such that some eventualities would count as disruptions of the structure and others would count as proper parts of it. In addition, some parts can be explained in terms of their relationship to other parts. I will assume the following necessary condition on intentional agency: DOI: 10.4324/9780429202131-12109

Lilian O’Brien

Purposefulness: If S is an intentional agent, then S has capacities that allow her to produce purposeful behaviour. I will rely on two further assumptions. First, intentional agents are, as the name suggests, beings with intentional states. These are states that represent, or are about, or directed upon, things and states of affairs beyond themselves. Common examples are belief and desire. Second, intentional states help to define this kind of agent by playing a role in generating and guiding their purposeful behaviours. Intentional: If S is an intentional agent, intentional states of S play a role in generating and guiding S’s purposeful behaviour. Given this role in generation and guidance, a distinctive kind of explanation of the behaviour of an intentional agent becomes possible. This is an explanation of the behaviour that draws on the guiding intentional states: Explicability of the Purposeful Behaviour of an Intentional Agent (EPIA): If B is a purposeful behaviour of an intentional agent S, then S’s intentional states play a role in explaining B. Purposefulness, Intentional, and EPIA offer a preliminary sketch of intentional agents. They are a kind of causal unit, one that produces a certain kind of behaviour, and this behaviour is explicable in a distinctive way.

2 Lower limits on the category Bees work together to build hives, make honey, and feed larvae. Given such complex behaviours, they seem to satisfy Purposefulness. Bees’ behaviour seems guided by informational states that, let’s assume, vary reliably with the facts – bees typically register the fact that nectar is nearby, say, only when there really is nectar nearby (Goldman 1979). Some philosophers think that these informational states qualify as intentional states (Dretske 1981, 1986). Although questions about the nature of intentional states are relevant and interesting, they go beyond this discussion, so let’s grant this. We can also suppose that the bees’ intentional states concerning nectar play an important role in explaining their behaviours, such as their flight to gather nectar – bees also seem to satisfy EPIA, at least as it is loosely stated above. We might nevertheless be sceptical that bees are intentional agents. Is this scepticism an objectionable kind of intellectualism about intentional agency? Enactivists argue that creatures even simpler than bees are agents (Hutto and Myin 2013). Are intentional agents simply that subset of agents who have intentional states guiding their behaviour? It is not clear that this wide notion of intentional agent would be theoretically useful. Such a category would include insects and higher animals, and it would also include creatures as sophisticated as statistically normal adult humans, who are capable of planning for the future, valuing and desiring things, acquiring skills, and evaluating a wide variety of practical options – courses of action that are open to them to pursue – as morally permissible or not, as prudent or not, and so on. Given this diversity, we would not be able to make too many philosophically interesting generalizations about the members of such a group, such as general claims concerning their distinctive characteristics, their capacities, their moral 110

Intentional agency

entitlements, and what kinds of action they are capable of. We would have to look to a potentially high number of sub-categories of intentional agency to frame and answer our questions. One may worry that there is really no useful category of intentional agent. After all, it does not seem to be a category that significantly shapes our everyday thought and talk, nor does it loom large in scientific or legal contexts. But I think that it is relied upon in philosophical contexts. Many philosophers are interested in what makes a rational practical agent autonomous, free, or morally responsible. They presuppose that there are intentional agents who nevertheless fall short of being fully autonomous, metaphysically free, or morally responsible. Human children, sophisticated non-human animals, and adult humans acting under duress are some examples. Having a better understanding of this basic general category of agent that we seem to rely on in philosophical contexts would be useful. In what follows, I consider whether there are grounds for excluding bees and other creatures from a category that has real boundaries. Let’s suppose that as long as a bee senses the proximity of nectar, abundant evidence of an insuperable barrier between her and it will not stop her from trying to fly towards the nectar. This behavioural flexibility is notoriously, and somewhat poignantly, reported to be missing from the Sphex wasp (Dennett 1984; Wooldridge 1963). This lack of flexibility may be one reason to doubt that the bee is an intentional agent. But human intentional agents may engage in very repetitive – so-called compulsive – behaviour that is not flexible, and yet we may think that they qualify as intentional agents even when they engage in such behaviour. Given this, it is more promising to think that intentional agents must have capacities for flexible behaviour even if they don’t always exercise them: Flexibility: If S is an intentional agent, S has capacities that allow her to exhibit a degree of behavioural flexibility when faced with obstacles to her purposeful behaviour, and this behavioural flexibility helps her to successfully complete her purposeful behaviour. What capacities must be involved in Flexibility? For starters, the bee would have to have an increased capacity for information-bearing intentional states, and she would also have to have capacities for updating existing information-bearing intentional states in light of new information. This requires ordering the contents of such states so that rational revision of existing informational states becomes possible. If this is correct, we should also accept this further condition on intentional agency: Rational revision: If S is an intentional agent, then S is able to update informational states in light of new information where updating involves rational revision of informational states. In being prompted to fly towards the nectar, I am imagining that the bee does not have a desire to fly there. Rather, the intentional state registering the presence of nectar triggers this flying behaviour directly. In imagining her in this way, she lacks a kind of intentional state that other agents have – a desire or desire-like state. Roughly, a desire is a state whose primary function is not to accurately represent how things are, but to represent a way that things could be, and to motivate behaviour that will bring this state of affairs into existence. If the bee had a desire-like state whose motivational force were suppressed when there is information to the effect that the desired state of affairs can’t be attained in the current 111

Lilian O’Brien

conditions, she could be released from fruitless repetitive behaviour, thereby having an increased capacity for flexible behaviour. Let’s add this to the capacities that are required for Flexibility: Desire: If S is an intentional agent, then S has desires or desire-like states that motivate behaviour in conjunction with informational states. In addition to increased behavioural flexibility in responding to changing states of affairs, a desire or desire-like state gives the creature greater causal independence from her immediate environment. If we assume that the state does not originate in the immediate environment, it allows the creature to have a source of behaviour that stems from within her. This resonates with the idea that an agent is some kind of causal unit that enjoys a degree of causal independence from its surroundings. Nevertheless, if the desire-like state is merely given by a creature’s biological nature or by her designer, and if the motivational force of the state is dampened or extinguished by informational states that are caused by the environment, then the causal independence that she has is quite slight. In fact, she seems to be a ‘site’ of causal processes rather than a source of them. This suggests that an intentional agent should also have some causal independence from her desire-like states. She should have, at a minimum, a capacity for suppressing desire-like states that is not determined by inputs from her immediate environment. This suggests the following further condition on intentional agency: Independence: If S is an intentional agent, S has a certain degree of causal independence from designer-given or biologically given desire-like intentional states in the production of her purposeful behaviours. We have considered what conditions might allow us to make sense of a general category of agent, who falls short of autonomous, free, morally responsible agency, but is nevertheless more complex than a bee. Creatures that lack the capacities that undergird flexible behaviour, such as information-updating that is governed by rational rules, who lack desires or desire-like states, who lack causal independence from their environment and from their own externally determined motivational states, may seem to be more like automatons – beings lacking independence from the past or their make-up or their environments – rather than intentional agents.

3 The agent it/her/himself Suppose that a complex computer has the two kinds of intentional state discussed above, and the operation of these states is not wholly determined either by the immediate environment or by a designer-given programme. Suppose that the computer is such that in informational state, IN, there is an 80% chance that a designer-given desire-like state is suppressed and a 20% chance that it will not be suppressed but will motivate and guide a bit of purposeful behaviour. Because of the probabilistic relationship among the computer’s states, whether or not the motivational state is suppressed is not directly determined by the designer. Suppose also that the computer produces purposeful behaviours that satisfy its desire-like states and that these behaviours are explained in terms of these desire-like states and its other intentional states. Would this computer qualify as an intentional agent? Does the computer itself produce these behaviours? Although the computer may be composed of complex causal chains that enjoy considerable independence from its immediate 112

Intentional agency

environment and its designer, the computer is not clearly something that itself exercises control. To come at this idea from a different direction, let’s suppose that a paradigmatic product of intentional agency is an intentional action involving bodily movement. But intentional actions are not things that are produced by parts of agents, but are performed by, and attributable to, the agent herself. Intentional actions are ‘personal’ rather than ‘sub-personal’ phenomena; they are performed by the agent rather than caused by an event in the agent. To come back to the computer, does it fulfil the following condition? Source: If S is an intentional agent, then S has capacities that allow her to exercise control over her purposeful behaviour. Event-causalists about intentional action characterize the agent’s exercising control in intentional action in terms of certain kinds of causal pathway obtaining between mental events, such as desire and belief, or intention (see Bratman 1999; Mele 1992; Sinhababu 2017 for discussions of the psychological states involved), and behaviours that are also events (e.g. Bishop 1989; Enç 2003; Mele 2003; Shepherd 2014; Smith 2010). This is an appealing view, because it does not involve an irreducible or supernatural entity in the role of agent. But there has been a debate about whether event-causal views suffer from a so-called ‘disappearing agent’ problem (Hornsby 2004, 2010; Mele 2003, chapter 10; Nagel 1986; Schlosser 2011). The event-causal view, it is argued, yields a picture in which one kind of event (mental) causes another (bodily behaviour), but the latter is not an agent’s performing an intentional action. In attempting to reduce the operations of agency to causal relations between mental and bodily events, the agent has disappeared altogether. A related complaint is that typical eventcausal views that characterize actions as caused by intentions or desire-belief pairs can only account for actions from which the agent is alienated, rather than actions ‘par excellence’, roughly, actions with which the agent identifies (Mele 2003, chapter 10; Velleman 1992). Two issues are intertwined here. One is conceptual: what is required for the agent herself to perform an intentional action – must she be a controller of her action, or the source of her action, or must she identify with the action or something else? And what is it to be any one of these things? The second issue concerns what it would take, ontologically speaking, for a creature to be the controller or source, or to identify with an intentional action. Would the agent have to be a substance who gives rise to, and intervenes in, causal processes, but whose operations are irreducible to causal relations among events? Or can the characteristic activities of the agent be characterized in event-causal terms? Event-causalists make a number of responses. They stress that the mental events involved in the production of intentional action are agent-involving psychological states: they are thoughts of the agent. Our talk of the computer’s ‘belief-like’ or ‘desire-like’ states, they might say, mislead us – a richer account of the psychology of intentional agency than we have considered so far will assuage at least some of the worry about disappearing agents. And we also need to characterize the ways in which such psychological states play a role, not just in causing movements but also in guiding these movements as they unfold over time (e.g. Bishop 1989). It is also argued that the disappearing agent’s worries rest on muddy intuitions that risk exaggerating the role of the agent, of ignoring continuities in agency between human and non-human animals, and misunderstanding the reductive naturalistic aims of the event-causalists (Mele 2003, chapter 10; Schlosser 2011). For worries that event-causalism can only capture alienated actions, it is argued that as long as there is a certain kind of psychological complexity in the production of actions, an event-causalist can capture the role of the agent in the production of intentional actions ‘par excellence’ (Mele 2003, chapter 10; Velleman 1992). 113

Lilian O’Brien

Those who reject event-causal responses may opt for agent-causal views (e.g. Brent 2017; Mayr 2011), characterizations of agency in terms of two-way powers (e.g. Alvarez 2013), or volitionist views (e.g. O’Shaughnessy 1986/2008; Pietroski 2000), and recent work on the ontology of processes also challenges event-causal approaches to intentional agency (e.g. essays in Stout 2018). Agent-causal views capitalize on, and contribute to, a rich ongoing research programme into substance causation, presenting an alternative to the Humean event-causal picture. Volitionist or trying theories offer accounts of willing that are irreducible to the operation of psychological states, such as desire. All of these views carefully articulate a number of worries about the event-causal programme in mind and action. It goes well beyond us to consider the details of the competing views, but as they centrally concern how to capture the difference between being a mere site of causal processes and being a performer of intentional actions, they are an excellent place to start in trying to capture what it takes for a condition like Source to be satisfied. One final thing should be noted: although I have here talked about intentional actions involving bodily movements, we must also consider what it takes to capture the role of the intentional agent in the production of other behaviours such as mental actions, omissions, refrainings, and lettings happen.

4 Unity in the category So far, I have motivated some conditions on intentional agency. Much more must be said to develop such conditions, but let’s turn now to the issue of whether the category, even as briefly sketched, promises to have the internal unity that was hoped for – a unity that would allow us to make fruitful generalizations about its members. A hungry cat sees a pigeon and begins to stalk it. But then she notices that there is a large dog approaching the pigeon. The cat promptly gives up the pursuit and retreats to lie in the sun. She watches and waits, attending to whether the dog moves to a safe distance. Let’s suppose that when the cat notices the pigeon, her mouth begins to water, she feels excitement, she feels motivated to move towards the pigeon, and these thoughts and feelings play a role in causing her to stalk it. When she notices the dog, she finds it threatening, she feels fear, and these thoughts and feelings cause her to stop in her tracks, to crouch down, and retreat. In breaking off her pursuit of the pigeon, let’s suppose that the cat doesn’t judge, nor does she have the capacity to judge, that the threatening dog would make her pursuit of the pigeon a mistake. She doesn’t treat the presence of the dog as a reason to retreat (Kauppinen 2021; Schlosser 2012). And let’s suppose that she doesn’t think that it would be bad or a mistake for her to cross the dog’s path. Furthermore, let’s suppose that the cat doesn’t have either the capacity to recognize that there is more than one practical option open to her, or the capacity to deliberate about which of two or more practical options to take, finally making a decision about which one is best. Rather, her finding the pigeon desirable or the dog threatening leads directly to attentional and behavioural changes that constitute her goal-directed behaviour of stalking the pigeon or avoiding the dog. Contrast the case of the cat with the case of a hungry child who sees a juicy apple. It looks delicious to her, her mouth begins to water, she feels excitement, she feels motivated to get it, and these thoughts and feelings play a causal role in her reaching for it. As she reaches, she notices that her friend frowns at her and begins to utter words of protest. The child hesitates. Let’s suppose that her hesitation is not just a conditioned response to frowns and protests – she has matured beyond that. Rather, she thinks that she shouldn’t take the apple because it is her friend’s, that it would be wrong of her to do so. These thoughts and feelings play a 114

Intentional agency

causal role in her bringing her action to a halt. After a little more thought about what to do, she decides to leave the apple to her friend. When looked at ‘from the inside,’ the cat and the child are very different: the cat doesn’t take states of affairs as reasons to act or not to act, but the child does; the cat does not engage in reflexive thought, but the child does; the cat doesn’t recognize that there is more than one practical option to choose from, but the child does. And the cat doesn’t regard herself as potentially criticizable, but the child does. Although they both plausibly fit the conditions on intentional agency that we have considered so far, their differing capacities for deliberative, normative, and reflexive thought make them very different from one another. Should we tighten the conditions on intentional agency to exclude the cat? For example, should it be required that intentional agents have the following capacities? Reasons: If S is an intentional agent, then S has the capacity to recognize and respond to reasons as such. Decision: If S is an intentional agent, then S has the capacity to decide to take a course of action from among two or more courses of action that seem to her to be open to her. Alternatively, should we use such conditions only to specify sub-categories of intentional agent? I will next offer some reasons to think that we can have a meaningful general category that includes creatures such as our imagined cat, the child, and also sophisticated adult human agents. Reasons and Decision could be used to delineate subspecies of the category of intentional agent, but the diversity would remain tolerable because of significant similarities across otherwise diverse agents. In spite of their differences there are striking similarities between the cat and the child: both have first-person perspectives on the world such that there is something that it is like to be them as they navigate that world (Nagel 1986). The phenomenal aspects of their intentional states seem to play roles in prompting and guiding their actions. Note that when adult human agents engage in sophisticated practical deliberation and action, this process often originates in an occurrence that is common to cat, child, and adult: they are moved by how things in the world strike them – as desirable or undesirable, as disgusting, delicious, ugly, beautiful, and so on. And even though a cat may not have a capacity for valuing things, for planning to do things in the future, or for self-criticism (Hieronymi 2009; O’Brien 2019), our capacities for these may be grounded in our capacity to find things beautiful, delightful, painful, and so on (Carruthers 2018; Jaworska 1999). It is also worth noting that a first-person perspective may play a role in making behaviours intelligible in a distinctive way. It may be that it is because we, statistically normal adult human planning agents know what it is like to feel very hungry and to find something delicious, that we may grasp in a special way why an agent would reach for it quite spontaneously. And it is because we know what it is like to find something threatening or forbidding, that we grasp why an agent would cower and stop in their tracks. Although the cat and the child may be very different from us in their capacities for normative or reflexive thought, their actions may nevertheless be at least partially intelligible to us in a distinctive way precisely because of similarities in the way things seem and feel to all of us, and because of similarities in how we all act in response. Given this, we should consider whether a condition such as the following is also necessary for intentional agency: FPP (First-Person Perspective): If S is an intentional agent, then there is something it is like for S to navigate the world and these phenomenal aspects play a role in generating and guiding her action. 115

Lilian O’Brien

I am tentatively suggesting that the right way to proceed in understanding the category of intentional agent is that, in spite of diversity in capacities for complex thought, all members of the category satisfy FPP. In addition, we should consider whether earlier conditions, such as Intentional and EPIA, should be developed in ways that reflect FPP. In the earlier, somewhat curt, introductions of belief-like and desire-like intentional states, they were not characterized in terms of how they feel to the agent. For all that was said, an account of them in terms of direction of fit (Anscombe 1957/2000; Frost 2014), or in purely causal-functional terms (Lewis 1972; Putnam 1975) – terms that makes no essential reference to phenomenal aspects – might have been acceptable. Perhaps zombie minds – minds in which there is nothing it is like to have a desire or belief – are possible (Chalmers 1998). Perhaps there are artificial agents who lack phenomenal aspects of intentional states, but seem to be intentional agents. Clearly, FPP is no shoo-in. Nevertheless, it is important to consider whether FPP, together with the other conditions, is one of the keys to the internal unity of a philosophically fruitful category of intentional agent. And we should think carefully about whether any such philosophically fruitful grouping can countenance the satisfaction of FPP by some of its members but not by all.

Related topics Animal agency; Rational agency; Planning agency; Agency, reasons, and rationality.

Recommended reading Alvarez 2013 develops a view of human agency in terms of two-way powers. Baker 2013, especially Chapters 2 and 6, offers a fresh and accessible discussion of the first-person perspective in different types of agent. Bratman 1999, especially Chapters 2 and 3, develops the planning theory of intention and defends the irreducibility of intentions to desires and beliefs. Hieronymi 2009 develops an account of the will that emphasizes the criticizability of the agent herself for her intentional actions. Mele 2003, especially Chapters 2, 3, and 10, discusses a range of issues concerning the defensibility of a naturalistic event-causal view of intentional action and agency.

Bibliography Alvarez, M. 2013. “Agency and Two-Way Powers”, Proceedings of the Aristotelian Society, 113:1, pp. 101–121. Anscombe, G.E.M. 1957/2000. Intention. Cambridge, MA: Harvard University Press. Baker, L. 2013. Naturalism and the First-Person Perspective. New York: Oxford University Press. Bishop, J. 1989. Natural Agency: An Essay on the Causal Theory of Action. Cambridge: Cambridge University Press. Bratman, M. 1999. Intention, Plans, and Practical Reason. Stanford: CSLI. Brent, M. 2017. “Agent Causation as a Solution to the Problem of Action”, Canadian Journal of Philosophy, 47:5, pp. 656–673. Carruthers, P. 2018 “Valence and Value”, Philosophy and Phenomenological Research, 97:3, pp. 658–680. Chalmers, D. 1996. The Conscious Mind. Oxford: Oxford University Press. Dennett, D. 1984. Elbow Room: The Varieties of Free Will Worth Wanting. Cambridge, MA: MIT Press. Dretske, F. 1981. Knowledge and the Flow of Information. Cambridge, MA: MIT Press. ———. 1986. “Misrepresentation”, in R. Bogdan (ed.), Belief: Form, Content, and Function. Oxford: Oxford University Press, pp. 17–36. Enç, B. 2003. How We Act: Causes, Reasons, and Intentions. Oxford: Oxford University Press. Frost, K. 2014. “On the Very Idea of Direction of Fit”, The Philosophical Review, 123:4, pp. 429–484.

116

Intentional agency Goldman, A. 1979. “What Is Justified Belief ?” in G.S. Pappas (ed.), Justification and Knowledge. Dordrecht: Reidel, pp. 1–25. Hieronymi, P. 2009. “The Will as Reason”, Philosophical Perspectives, 23, pp. 201–220. Hornsby, J. 2004. “Agency and Actions”, in Hyman and Steward (eds.), Agency and Action. Cambridge, UK: Cambridge University Press, pp.1–24. ——— 2010. “The Standard Story of Action: An Exchange (2)”, in Aguilar and Buckareff (eds.), Causing Human Actions. Cambridge, MA: MIT Press, pp. 57–68. Hursthouse, R. 1991. ‘Arational Actions’, Journal of Philosophy, 88, pp. 57–68. Hutto, D. and Myin, E. 2013. Radicalizing Enactivism: Basic Minds without Content. Cambridge, MA: MIT Press. Jaworska, A. 1999. “Respecting the Margins of Agency: Alzheimer’s Patients and the Capacity to Value”, Philosophy and Public Affairs, 28:2, pp. 105–138. Kauppinen, A. 2021. “Rationality as the rule of reason”, Noûs, 55:3, 538–559. Lewis, D. 1972 “Psychophysical and Theoretical Identifications”, Australasian Journal of Philosophy, 50, pp. 249–258. Mayr, E. 2011. Understanding Human Agency. Oxford: Oxford University Press. Mele, A. 1988. ‘Effective Reasons and Intrinsically Motivated Actions’, Philosophy and Phenomenological Research, 48:4, pp. 723–731. ———. 1992. Springs of Action. New York and Oxford: Oxford University Press. ———. 2003. Motivation and Agency. New York and Oxford: Oxford University Press. Nagel, T. 1986. The View from Nowhere. New York and Oxford: Oxford University Press. O’Brien, L. 2019. ‘The Subjective Authority of Intention’, The Philosophical Quarterly, 69:275, pp. 354–373. O’Shaughnessy, B. 1986/2008. The Will. Volumes 1 and 2, 2nd edition. Cambridge: Cambridge University Press. Pietroski, P. 2000. Causing Actions. Oxford: Oxford University Press. Putnam, H. 1975. “The Nature of Mental States”, Mind, Language, and Reality: Philosophical Papers, Volume 2. Cambridge: Cambridge University Press. Schlosser, M. 2011. “Agency, Ownership, and the Standard Theory”, in Aguilar, Buckareff, and Frankish (eds.) New Waves in the Philosophy of Action. London: Palgrave-MacMillan. ——— 2012. “Taking Something as a Reason for Action”, Philosophical Papers, 41:2, pp. 267–304. Shepherd, J. 2014. “The Contours of Control”, Philosophical Studies, 170:3, pp. 395-41. Sinhababu, N. 2017. Humean Nature: How Desire Explains Action, Thought, and Feeling. Oxford: Oxford University Press. Smith, M. 2010. “The Standard Story of Action: An Exchange (1)”, in Aguilar and Buckareff (eds.), Causing Human Actions. Cambridge, MA: MIT Press, pp. 45–56. Stout, R. (ed.) 2018. Process, Action, and Experience. Oxford: Oxford University Press. Velleman, D. 1992. “What Happens When Someone Acts?” Mind, 101:403, pp. 461–481. Wooldridge, D. 1963. The Machinery of the Brain. New York: McGraw Hill.

117

10 RATIONAL AGENCY Eric Marcus

According to the Aristotelian definition, the human being is the rational animal. Many would for this reason reject it, contending that the rational patterns exhibited by human behavior and cognition are present, even if to a lesser degree, in nonhumans as well. One way of defending the definition is to argue that the practical and theoretical capacities of humans constitute a distinctive form of agency: rational agency. Unlike many of the topics in this handbook, this is a relatively unexplored area of research. There is no extensive literature to survey and no warring factions whose battle lines might be usefully charted. Since many philosophers simply have no idea what a rational form of agency would be, this entry will focus on the very idea. In a suitably broad sense of activity, one might say that activity marks the distinction between the animate and the inanimate. Living things are active insofar as they grow and reproduce. Animals are self-movers, and as such are active in a more demanding sense. Self-movement can be understood in terms of an animal’s acting on the basis of perception in pursuit of its aims. The network of capacities underlying this activity is constitutive of animal agency. But animals are passive in a way in which humans are not. Whereas we are capable of stepping back and reflecting on the goodness of our aims and making a judgment about what to do on the basis of this reflection, nonrational animals are driven entirely by their aims (see Korsgaard 2009). Similarly, a nonrational animal is not capable of believing on the basis of consideration of what to believe, but rather simply believes what it, so to speak, finds itself believing. This is to gesture in the direction of the idea of a distinctively rational agency. We can begin to flesh the idea out by considering the nature of the capacities that manifest themselves in episodes of rational reflection. Such episodes involve what one might call self-conscious engagement with normative questions—questions paradigmatically of whether to do x or believe p. If a poker player is deciding whether or not to call a large, surprising bet, she is asking herself questions about, for example, what her opponent’s behavior on prior rounds of betting indicates, the significance (and ingenuousness) of his body language, what she knows about his general tendencies as a player, what impact her possible actions will have on her “table image,” whether what might ordinarily be an overly cautious fold would be prudent given how close she is to “making the money,” and so forth. These reflections require the possession of the concepts that figure in the specification both of the actions being considered and the facts on which the relative wisdom of the various choices depend. 118

DOI: 10.4324/9780429202131-13

Rational agency

Someone who engages in such reflection does not merely manifest the possession of concepts in the sense in which a dog, who reacts excitedly upon seeing his master reach for the leash, manifests the concepts of “leash” and “walk.” A rational subject’s manner of possession of these concepts is linked to an understanding of the contributions they make to truth-conditions, to some awareness of the evidential significance of their exemplification, and to a capacity to grasp and evaluate the corresponding propositions. It is not unreasonable to think that these are the abilities that are limited to language users. We are, after all, unwilling to credit someone with thoughts of the sort that figure in the above reflections if the thinker displays no mastery of the linguistic forms that are used to express them. Acts of reflection also require the possession of the concepts that articulate the form of the questions themselves. Our gambler must understand the bearing of the considerations she brings to consciousness on questions of whether to believe something, for example, that her opponent is bluffing, and so whether to do something, for example, call the bet. She must thus possess the concepts of belief and of action. Furthermore, she must understand the very idea of considerations showing a proposition to be one that should be believed, that is, to be true, and the very idea of considerations showing an action should be performed, that is, to be good. For this understanding is an element in the relevant forms of rational responsiveness. The subject must thus have some (very general understanding) of the framework of theoretical and practical justification. We might summarize these points by saying that rational reflection depends upon the sort of grasp of concepts that makes it possible to explicitly consider normative questions. It is, in this sense, a self-conscious activity: a rational deliberator understands what she’s doing as aiming to make up her mind about what to believe or do. Suppose this much is granted. Suppose that it is also granted that nonhuman animals lack the cognitive wherewithal to engage in rational reflection, so understood. Finally, suppose it is allowed that such reflection substantially affects the lives of the deliberators. It would still not follow that humans possess anything that deserves to be called a distinctive form of agency. For it might nonetheless be contended that both human and nonhuman animals act on the basis of prior thought in the same sense that the difference between them is only in how conceptually sophisticated those thoughts are. This, some will argue, is just a difference of degree. And so what we really have is just ordinary animal agency, which varies in the contents of the associated mental states according to the relevant creature’s cognitive abilities. The nature of human thought and action and the way the former affects the latter are not, it will be concluded, tied to any of the specifically rational capacities that we are now supposing belong exclusively to human beings. Thus, we still have no case for or even a clear conception of rational agency. Our imagined skeptic of rational agency is right about this much: for the possession of a network of capacities to amount to a form of agency, their exercise would have to be the central element in a distinct kind of change, distinct in the way that, for example, growth and self-movement are distinct kinds of change. One might express this point by saying that for rationality to amount to a kind of agency there has to be a distinctive form of causation: rational causation. (If “causation” seems too metaphysically loaded, then read: “distinctive form of causal-explanation: rational causal-explanation.”) But what does this mean? Deliberation culminates, in ideal cases, in the making up of one’s mind: to do x, or to believe that p. These conclusions are based on the reasons for action and belief that proved decisive. We often ask after these bases: “Why did you call the bet?” “Because I thought he was bluffing.” “Why did you think he was bluffing?” “Because he always smiles like that 119

Eric Marcus

when he’s bluffing.” Here the gambler gives rational explanations of her action and belief. It is characteristic of such explanations that what’s offered as explanation is at the same time justification. It is an explanation that consists of a justification. But it is not just a justification, it is her justification: what, at the time, the deliberator took to justify the relevant belief or action. It was, according to the explanation given, in virtue of her taking the stated reason to justify that the action was performed or the belief adopted. It has vexed scores of philosophers working in various subfields to say more precisely what it is for the justification to be hers. Donald Davidson famously points out that (in the practical case) it is not just a matter of the action being caused by the agent’s desiring an end and believing that the relevant action will bring the end about. His infamous mountain climber does not let go of the rope, killing his companion, in order to save himself; rather, the recognition that he could save himself by doing so stuns him, thereby causing him to let go (Davidson 1963). This (“deviant”) causation circumvents rather than serves his agency, even as it leads to the realization of his aim. We have, on the one hand, paradigmatic exercises of rational powers: the making up of one’s mind to do x or to believe p on the basis of considerations. We have, on the other hand, certain causal phenomena: someone performs an action or adopts a belief for reasons. There is, apart from any interest in our topic here, a general sense that we are still in the dark as to how to understand the connection between the former and the latter. There would be rational agency if the solution to this mystery went as follows: facts of the former sort constitute facts of the latter sort. That is, if the causal connection between a reason and the action or belief that rests on it were nothing beyond the exercise of rational powers— specifically the power to make up one’s mind about what to do or believe on the basis of considerations—then there would be a manner of bringing things about that belonged exclusively to rational creatures. Rational agency, then, is (or would be) the power to decide normative questions in a manner so as to constitute facts about the causes (or, if you prefer, causal explanations) of certain events and states. For a subject to accuse the butler of doing it because the butler lacked an alibi or to believe that the butler did it because the butler lacked an alibi is, on this view, nothing over and above the subject’s viewing the action of accusing the butler as to be done or the proposition that the butler did it as to be believed in light of his lacking an alibi. According to this proposal, in saying that someone is x-ing or believes that p for the reason R, we attribute to her a normative judgment that x-ing is to be done or p is to be believed in light of R. This is closely connected to the “Guise of the Good” thesis, according to which acting involves a normative judgment that the action is good. Here the idea is that in acting for a reason one performs the action under the guise of R’s contributing to its goodness. A defense of this approach will thus need to reply to those who contend that someone might act for a reason they knew did not establish the goodness of the relevant action—as, it might be contended, a certain sort of akratic does—or for a reason they took to establish the badness of an action—as, it might be contended, an aspiring super-villain does (see the essays in Tenenbaum 2010 for a recent discussion). The thesis that in inferring p from q, one draws the conclusion in the light of the support q provides p is referred to by Paul Boghossian as the Taking Condition (Boghossian 2014). Here, too, defense will be required: against the critics of the Taking Condition, who argue variously that there is no way of spelling it out without regress, that it requires too much conceptual sophistication, that it makes false claims about the phenomenology of inference, or that it ignores the possibility of the doxastic equivalent of akrasia (See, e.g., McHugh and Way 2016). 120

Rational agency

The most controversial aspect of the thesis of rational agency is the central causal claim: that certain practical and theoretical normative judgments constitute the causal-explanatory nexus between the reason and the relevant act or state. This idea runs deeply against the grain of conventional thinking about causation, according to which the obtaining of a causal connection is the wrong sort of thing to be constitutively dependent on a subject’s representing the world a certain way. It is, of course, uncontroversial that a subject’s representation can influence and be influenced by the world. But rational agency, as I suggest it must be understood, involves the idea that a subject’s representing R as conferring to-be-done-ness on an action or to-be-believed-ness on a proposition constitutes the obtaining of a causal connection between R and the relevant action or belief. The resultant conception of causation must be defended (see Marcus 2012). To think of people as possessing distinctively rational agency is to see them as authors of a distinctive kind of change, one that is tied essentially to their rational sensitivity to reasons. The agent, in concluding that x is to be done on the basis of R, thereby is x-ing on the basis or R. The subject, in concluding that p is to be believed on the basis of R thereby believes p on the basis R. It follows that there is no gap between R and the action or belief—no “Reasons Gap,” as we can call it. It’s not that, say, one judges that one should put out the trash and this causes, via nonrational mechanisms, certain bodily movements that accomplish one’s aim. One’s judgment that it is to be done constitutes the fact that one is taking out the trash. And so it is precisely this judgment that then makes it possible to perform further actions because one is taking out the trash (See Thompson 2008). The conclusion of practical reasoning, as Aristotle held, is an action. It is the action of x-ing itself that is constituted by the subject’s judgment that x is to be done. Similarly, it is not that one judges that one should believe that it’s Tuesday, and that this then contributes to the formation (in an optimal case) of the disposition in which believing that it’s Tuesday consists. One’s judgment that it is to be believed constitutes the fact that one believes it’s Tuesday. And so it is precisely this judgment that makes it possible to then infer that it’s not a weekend because it’s Tuesday. The conclusion of theoretical reasoning is belief, and thus it is belief that p itself that is constituted by the subject’s judgment that p is to be believed. Human action and belief are, according to the thesis of rational agency, themselves manifestations of the very capacities exercised in explicit acts of critical reflection. The thesis of rational agency also eliminates what we can call “The Knowledge Gap.” To accept the gap is to think that what one knows simply in x-ing for the reason R or knows simply in believing that p for the reason R falls short of the fact that one is x-ing because R or believes that p because R. To reject this gap is thus to hold that one who is x-ing on the basis of R or who believes that p on the basis of R knows, simply in doing so, that she is x-ing because R or believes that p because R. Anscombe’s Intention begins from the assumption that there is no Knowledge Gap. She argues that someone who performs an action can “give application” to the rational “why?,” and not on the basis of observation or evidence (Anscombe 2000). The agent, on her view, knows non-empirically what she’s doing and why (in the relevant sense) she’s doing it. This claim can be made with equal plausibility about belief and our reasons for belief. Insofar as action and belief are constituted by judgment, we can begin to make sense of this knowledge. Since they are constituted by normative judgments, rational creatures have the kind of epistemic access to action and belief that they have to their own judgments. And if there is nothing more to the causation underlying someone’s x-ing for the reason that R or believing that p for the reason that R than her viewing R as conferring to-be-done-ness on x-ing or to-be-believed-ness on p, then this non-empirical knowledge of our reasons would be intelligible. The subject can speak with special authority 121

Eric Marcus

about the question of why the action is performed or the proposition believed because the relevant facts are constituted by how she herself takes matters to be, specifically: her taking the action as to be done or the proposition as to be believed in light of the justification provided by the relevant reasons. It should now be evident that there are many challenges facing the defender of rational agency. As mentioned above, it will be argued that people act and believe against their normative judgments, and that such judgments are neither necessary nor sufficient for belief and action. The Aristotelian doctrine that action is the conclusion of practical reasoning can seem especially problematic, given that people may never get around to doing what they sincerely and non-akratically judge is to be done, and that people may make such judgments where the relevant action simply cannot ever be performed. Others will hear in the doctrine the implausible suggestion that what we do takes place inside the mind. But, as I have emphasized, the idea is rather that an action must be understood as itself a manifestation of (as opposed merely to an effect of ) human practical rationality, as the agent’s answer to the question: “What should I do?” Actions are, on this view, elements of the space of reasons, and not simply via the proxy of causally efficacious psychological states. Though Anscombe and those who follow her take the rejection of the Knowledge Gap as a datum to be explained, others would reject it outright. People who are self-deceived might be described as not knowing why they believe and act as they do. Implicit bias might be interpreted as a matter of possessing beliefs of which one had no awareness. And an agent who acts, for example, selfishly, might prefer to believe a more favorable account of her own action and this in turn might, through selective attention and motivated reasoning, lead her to adopt one. To claim as an advantage a superior position from which to explain our non-empirical knowledge of our actions, beliefs, and the reasons on which they are based, an advocate of rational agency would have to deal with the apparent threat to the datum posed by these sorts of cases. If acting and believing for reasons require—or, more precisely, if the correct application of the corresponding linguistic forms to people require—that the agents and believers possess sophisticated conceptual understanding, then it follows that nonrational beings do not (in the same sense) believe or act for reasons. Some account will then be required of what we are talking about when we describe a dog, for example, as running down the stairs because its master called for him. Such descriptions impute to animals actions performed on the basis of thought. It is unlikely that any view of the sort I describe in this entry will find a widespread following until a plausible account of nonrational thought and action can be formulated. This remains the most significant challenge for advocates of rational agency. Although a detailed account has yet to be given, the form that a satisfying response will take is clear. John McDowell argues that although rational and nonrational creatures are perceptually sensitive to their environment, the perception of rational creatures necessarily draws into operation conceptual capacities that the nonrational lack (McDowell 1994). Continuing along this path we might say that whereas thought quite generally puts the thinker in cognitive contact with the world, mediating between perception and action, when a species or individual acquires the suite of cognitive abilities that constitute rationality, their perception, thought and action are not merely supplemented with additional contents, but transformed into qualitatively different capacities. When a rational creature sees, thinks, and acts, she exercises distinctively rational conceptual powers (in the manner sketched above). A successful elaboration of this idea must exhibit the commonalities that make rational and nonrational agency instances of a common genus and the differences that make them distinct species of that genus (cf., Boyle 2012). 122

Rational agency

Above, I said that episodes of rational deliberation involve self-conscious engagement with normative questions. Such episodes are marked by the requirement of conceptual sophistication and culminate, in ideal cases, in causation-constituting normative judgments that utilize this sophistication and exhibit the just-discussed rational self-consciousness. But there is no requirement that such judgments be preceded by deliberation. In fact, it is surely the exception. Work needs to be done to understand the nature of these judgments and what underlies the mutual interdependence of their salient characteristics—that they require conceptual sophistication, that they are made as if in answer to questions about what to believe or do, that they put the rational being in a position to speak authoritatively about the relevant causal matters. But this further understanding will not contradict the self-evident truth that the sorts of judgment in which deliberation culminates are sometimes made without deliberation. It’s not as if someone whose deliberation leads him to say that p must be true in light of R is expressing something different from another who makes the same claim without deliberation. Note that it does not follow from the fact that much of what we believe and do for reasons is not the result of episodes of deliberation that someone who is incapable of engaging in conscious episodes of deliberation could make the relevant judgments. It is only to say that those judgments, with the interlocking characteristics that we introduced by way of considering rational deliberation, can occur even without such deliberation. In fact, it is highly implausible that a creature that is incapable of explicitly taking up normative questions could nonetheless possess the ability to make judgments that are the taking of stands on them. Finally, it is worth emphasizing that this sketch of the very idea of rational agency wards off two common misunderstandings of the thesis that humans are a distinctive kind of agent in virtue of their rationality. First, it is not equivalent to the absurd thesis that humans are unerring optimizers and Spock-like cogitators. And not because these are ideals of which we fall short. Rather, these archetypes simply do not personify the rationality that figures in the thesis under discussion. Rational, in the relevant sense, does not contrast with irrational, but with nonrational. Second, the thesis is not that, over and above the exercise of nonrational cognitive powers, we are also capable of exercising rational cognitive powers. This flawed conception would be that while we go about our ordinary business, we think and act in the manner of the nonrational. Then, in occasional episodes of critical reflection, we exercise a capacity that is exclusively our own. These exercises can then impact what we believe and do, where these thoughts and actions are still understood as of the sorts of states and events that figure in the lives of animals more generally. The thesis is rather that the thoughts and actions of human beings quite generally are themselves manifestations of the very capacities that occasionally also manifest themselves in episodes of explicit critical reflection.

Related topics Agency and causation; Intentional agency; Mental agency; Agency and practical knowledge; Agency and autonomy

Recommended reading Boyle, M. (2011). ‘“Making up Your Mind” and the Activity of Reason’. Philosophers’ Imprint 11. Rational creatures can make up their minds about what to believe on the basis of reasons in favor of believing. According to this essay, a belief just is the actualization of this power. Boyle, M. and Lavin, D. (2010). ‘Goodness and Desire’ in Tenenbaum 2010, p 161–202.

123

Eric Marcus The essay argues that the specific form that goal-directedness takes in rational creatures is the capacity to act under the guise of the good, i.e., to act in light of her answer to the question ‘What should I do?’. Marcus, E. (2012). Rational Causation. Cambridge, MA: Harvard University Press. An articulation of a conception of rational agency along the lines presented in this entry and of the challenges it poses to the naturalistic orthodoxy in the philosophy of mind. Moran, R. (2001). Authority and Estrangement. Princeton, NJ: Princeton University Press. An exploration of the connection between the first-person perspective and the phenomenon of normative self-consciousness: our (fallible) ability to know what we believe or want (and why) in virtue of it reflecting our judgments about what is to be believed or to be wanted (and why). Rödl, S. (2007). Self-Consciousness. Cambridge, MA: Harvard University Press. It is argued that the determination of one’s belief as a matter of what to believe and one’s action as a matter of what to do is at the same time an exercise of the ability to think of oneself as a subject—as “I.”

References Anscombe, G.E.M. (2000). Intention. Cambridge, MA: Harvard University Press. Boghossian, P. (2014). ‘What Is Inference?’ Philosophical Studies 169 (1):1–18. Boyle, M. (2012). Essentially Rational Animals, in Rethinking Epistemology, Abel, G. and Conant, J., eds., Berlin, Germany: Walter de Gruyter. Davidson, D. (1963). ‘Actions, Reasons, and Causes’, Journal of Philosophy 60 (23):685–700. Korsgaard, C. (2009). ‘The Activity of Reason’, Proceedings and Addresses of the American Philosophical Association 83 (2):23–43. McDowell, J. (1994). Mind and World. Cambridge, MA: Harvard University Press. McHugh, C. and Way, J. (2016). Against the Taking Condition. Philosophical Issues 26 (1):314–331. Tenenbaum, S. (ed.) (2010). Desire, Practical Reason, and the Good. Oxford: Oxford University Press, 161–202. Thompson, M. (2008). Life and Action: Elementary Structures of Practice and Practical Thought. Cambridge, MA: Harvard University Press.

124

PART 3

Agency and ability Introduction to Part 3 Luca Ferrero

Although the use of the term ‘agency’ might make us think that we are talking about a single and monolithic capacity, agency might be better thought in terms of a distinctive constellation of powers, capacities, and dispositions, which operate, at least normally, in a sufficiently integrated fashion. These powers, capacities, and dispositions, however, do not operate simply as the natural powers of non-agential entities. When agents come to the scene, so do the concepts of ability and skill, and the correlated notions of success, mastery, and expertise, on the one hand, and mistakes, disabilities, and pathologies, on the other. These are the topics of the chapters in this section. The section starts with a survey of the distinctive features of genuinely agential powers, their nature, and their role in the explanation of action. It then moves to a discussion of the questions raised by skillful performance and expert agency. What is the relation between habitual, skilled, and expert agency? Is expert performance more or less effortful than novice performance? How does expert performance relate to standard cases of intentional agency? The discussion then moves to what we can learn from what might be called ‘privative’ manifestations of agency: first, are there any specific mistakes of agency? And if so, what do they reveal about the nature of agency? A similar set of questions arise in connection not just with occasional failures, but with a systematic lack of certain capacities and powers, as it happens in disabilities and pathologies of agency.

Agency, powers, and skills In ‘Agency, powers, and skills,’ Will Small addresses the question of the nature of agential powers. Unlike natural powers, agential powers are not just manifested but exercised. They are typically exercised in intentional actions, they can be exercised on different occasions, and they persist between exercises. Although standard explanations of intentional actions are cast in terms of reasons, agential powers play an important explanatory role as well. At the very least, they seem to play an explanatory role at the point when deliberation comes to an end and the agent is supposed to do something ‘just like that’—that is, to perform a teleologically basic action, one for which the question of which means to take no longer arises. But the explanatory role of agential powers does not seem to be restricted to basic actions. Many skills and abilities are primarily exercised in how an agent perceives of a certain situation and deliberates about how to do things in that situation, rather than in some bodily DOI: 10.4324/9780429202131-14125

Luca Ferrero

movement (which might be relatively easy and available to an unskilled agent—consider, for instance, the ability to play chess, where the mere moving of the pieces on the chessboard requires no special skill). Another issue concerns the specification of agential powers. Schematically, an ability is individuated as the ability to x, where the x stands for an act-type. But how fine or coarse-grained is the act-type? Which ability is exercised, Small asks, when one butters a toast with a knife at midnight in the kitchen? Typically, the proper specification of an agential power is more coarse-grained than the individuation of the action that is performed in exercising that power. We often use the word ‘skill’ to refer to practical abilities whose exercise is more complex and requires higher levels of competence. The question then arises of how intelligence, cognition, and knowledge relate to agential powers in the case of skill. Finally, Small discussed different philosophical views of the relation between agential powers and natural powers or dispositions. A disposition is usually individuated by its stimulus and its manifestation: X is fragile iff X is disposed to break if struck in a certain way. Those philosophers who claim that agential powers are ultimately dispositions need to characterize what is distinctive of the stimuli and manifestations of agential powers. But some philosophers argue that agential powers are of a different kind altogether. Whereas dispositions are one-way powers, agential powers are ‘two-ways,’ since they can be exercised in two correlated ways: it is up to the agent whether or not to exercise her power when one has the opportunity to do so; or, according to a different view, a two-way power maybe be exercised either in x-ing or in ‘contra-x-ing,’ as in the case of medicine, where a skilled doctor is the one who has the power to either heal a patient or to harm her.

Expert agency The question of expertise is the topic of Barbara Montero’s ‘Expert agency.’ There is a common view about skilled agency according to which, as expertise develops, both effort and agency diminish: expert agency ultimately becomes so effortless that, at its pinnacle, the agent disappears. Montero argues for the opposite view: experts generally act effortfully when engaging in skillful actions, their effort does not usually interfere with their skills, and expert agency is not agent-less. Montero distinguishes everyday expertise from professional-level expertise. Everyday expertise (e.g., the one exercised when climbing stairs or swimming recreationally) is largely built out of relatively mindless repetition. This agency might indeed be experienced as effortless. But the mindset of professional-level expertise (professional athletes, artists, doctors, scholars, etc.) is quite different. These experts not only worked hard to achieve their expert status but they continue to practice their skills in order to improve them. Unlike everyday expertise, in professional expertise, one does not settle for ‘good enough.’ So it is a mistake, Montero argues, to take the investigation into everyday expertise to reveal the nature of professional expertise. Agency and effort are not the same, since one might experience agency in performing very easy, effortless tasks. But Montero argues that consciously experienced effort comes with the feeling of agency, the feeling that one is doing something. Hence, effortful expertise implies agentful expertise. In optimal skillful action, the agent is there. Evidence to the contrary seems to come from three sources: expertise-induced amnesia, Csikszentmihalyi’s concept of flow, and the Taoist concept of the wu-wei. But Montero does not find any of these convincing. First, although it is true that post-performance amnesia occurs in a wide range of skills, failures to register effort and agency in past skillful performance do not 126

Agency and ability 

necessarily prove that no effort or agency was experienced at that time. After all, one might consciously attend to one’s skillful actions yet forget them immediately afterward. Conscious attention might have been so keenly focused on the moment that it impeded memory formation. Alternatively, if the experts are exclusively focused on nonverbalizable aspects of their skills, they might be unable to form any declarative memory of their past effortful performance. Second, the phenomenon of ‘flow’ is often taken to illustrate the dissolution of effort in expert action and the resultant disappearance of agency. Against this interpretation, Montero argues that Csikszentmihalyi’s research on ‘flow’ does not support the claim of effortless expertise, since what he investigated was not peak performance, but peak experience. The two should not be confused: doing one’s best is not necessarily the same as feeling one’s best. Finally, Montero is unconvinced by the interpretation of Taoist texts as supporting the idea that expert action is effortless and agentless. In closing, Montero warns us against the temptation to fall for the appearances that the expert actions of others are effortless. This is often an illusion: Montero’s own experience in engaging in professional activities such as ballet and philosophical writing proves the opposite: her expertise in either of them has been proportional to her effort. As she concludes, “expertise is effortful and agentful, which can at times turn what was enjoyable into a nightmarish trial.”

Agency and mistakes It is uncontroversial that we are imperfect and fallible agents; we make many mistakes. But, as Santiago Amaya argues in ‘Agency and mistakes,’ our imperfection and fallibility are too often overlooked in the philosophical reflection on agency, which tends to emphasize the successful exercises of our agential abilities. Mistakes, however, can reveal important features about what makes us the kind of agents that we are. Our actions are subject to two different kinds of standards: constitutive standards (which our behavior must meet to be considered an instance of agency in the first place) and standards of success (which our behavior must meet for an action to succeed as an action—such as achieving what one intends to do). When an agent makes a mistake, she does so in acting, that is, her behavior meets the standards constitutive of agency (given that she is acting) but fails to meet standards of success. There are two kinds of mistakes. First, some mistakes are due to the attitudes that explain our conduct, such as our beliefs, intentions, and desires (e.g., having problematic or false beliefs, desiring something that is not really desirable, having inconsistent or misaligned attitudes) In these cases, however, the trouble is not primarily about failing at meeting standards of success of ‘agency,’ but rather standards of the correctness of the attitudes in question. What is more revealing about the nature of our agency are instead what Amaya calls ‘performance mistakes’: the mistakes we make not in our capacity as subjects of mental attitudes but as creatures who need to translate the combination of these attitudes into actual behavior. Even when our attitudes are unproblematic, we can still (and often do) fail at translating them into actions—as it happens, for instance, in action slips. More generally, performance mistakes occur whenever there is a disconnect between our actions and the relevant attitudes that should be reflected in them. This fallibilism is a pervasive feature of our agency and cannot be fully explained in terms of the defects and shortcomings that we might find in the acquisition, retention, and revision of our mental attitudes. Performance mistakes are the distinctive and non-derivative mistakes in the exercise of our agency. What they reveal about us is that, no matter how perfect we might be in our 127

Luca Ferrero

appreciation of the circumstances, in our motives, and in our deliberation, we still need to translate our decisions, intentions, and plans into actions and to sustain them over time, which requires the proper and successful functioning of our agential powers and abilities. Against the idealization and perfectionism that is pervasive in the philosophical discussion of agency, Amaya concludes by urging us to embrace the fallibilism of our agency, to acknowledge that “the possibility of making mistakes is built into whichever powers make us agents. Ultimately… the gap between the constitutive standards of our agency and our standards of success is not entirely accounted for by the limitations we have in other domains of our rational lives.”

Agency and disability A further push away from an idealized picture of agency comes from the study of disabilities. As argued by Kevin Timpe in ‘Agency and disability,’ the primary lesson that we learn by considering how disabilities affect human agency is that we should take seriously the degreed nature of the capacities and behaviors involved in human agency. Agency itself might turn out to be a degreed concept. Timpe takes a ‘ground-up’ approach, by exploring first a variety of specific disabilities rather than starting to treat disability as a monolithic and unified concept with an unequivocal impact on agency. There is much heterogeneity in the kinds of disabilities, and even the same disability might manifest itself in a wide range of ways, across different individuals, circumstances, and social contexts. Timpe discusses examples of physical disability (esp. akinesia, the lack of movement of slowness of initiating and maintaining movement in Parkinson’s disease), emotional disability (e.g., emotional blunting or flattening of affect, commonly associated with schizophrenic syndrome and frontotemporal dementia), and intellectual disabilities (which can impact the agent’s capacity to plan, learn, adapt their conduct, and interact socially). The reflection on the relationship between disabilities and agency teaches us several important lessons about human agency in general. First, we should take more seriously the degreed character of our agential abilities and agency. Second, against an atomistic or individualized approach to human agency, disability shows us how much of our agency depends on the agent’s social and environmental circumstances. Third, we learn that, even when one succeeds in one’s agency, one does so against difficulties that come in different degrees and relative to the agent’s social and environmental contexts. Similar to what Amaya argued in the earlier chapter, Kimpe insists on the general fallibility of our agential abilities. Finally, there are important normative implications: not only does agency come in degrees, but so does our moral responsibility, and the extent of our blameworthiness and praiseworthiness.

Pathologies of agency In ‘Pathologies of agency,’ Lubomira Radoilska shows us how the investigation of these pathologies makes it possible to differentiate between various agential abilities, which are difficult to tell apart in standard instances of successful agency. This is because in successful agency, these abilities are exercised as part of a bundle that obscures their functional and conceptual separation. This has, as she writes, a twofold payoff: we gain both “a clearer insight into manifestations of agency beyond straightforward intentional actions, including negative, omissive, and second-order exercises, such as permitting, preventing, facilitating,

128

Agency and ability 

or contravening,” and a better understanding and assessment of “competing attributions of responsibility by mapping them onto different patterns of agency.” Two preliminary distinctions are in order: first, pathologies of agency (which affect the contours of agency) are different from mere circumstantial sources of impediments or obstacles to non-pathological agency. Second, there are two complementary dimensions of agential achievement: production (bringing about the intended effect) and assertion (the articulation of the agent’s reasons and commitments). In a fully successful action, these two dimensions are well-aligned. But they can come apart. For instance, in both compulsion and weakness of will, the action is successful as production but unsuccessful as assertion, since the action does not match the agent’s better judgment or policy. If the misalignment is sustained over time, the agency might become, as Radoilska calls it, ‘necessarily less-than-successful.’ Partial, unavoidable, and repeated failures of agency along one of these dimensions are constitutive of pathological agency. It is important to notice that these failures do not put the agent outside of the space of reason, so there is still room for some criticism of irrationality. A telling example of pathology of agency is apathy or auto-activation deficit (AAD) syndrome, a neurological disorder where self-generated voluntary and purposive actions are virtually absent, while externally driven behaviors are normally executed. In apathy, the bond between practical knowledge and knowledge of one’s motivating reasons is broken: people with apathy can lucidly report what they are doing but not their motivating reasons. In this case, agents do not appear to be responsible in the mode of ‘answerability,’ since their behavior is only sensitive to incentives rather than responsive to any reason that they might be able to articulate. Even if some mental disorders and anomalous psychological conditions might stand in the way of responsiveness to reasons within a particular context, this does not imply that the affected agents are not members of the moral community. Radoilska concludes by drawing some implications for the attribution of decisional capacity and autonomy in the psychiatric practice. On the one hand, we should recognize the presence of reason-responsiveness (and thus decisional capacity and autonomy) in mental disorders and abnormal psychological conditions. On the other hand, we should avoid treating those aspects of behavior that are genuinely not reason-responsive as if they were. Here is where the notion of necessarily less-than-successful agency, which distinguishes between the different dimensions of achievement, can be of help.

129

11 AGENCY, POWERS, AND SKILLS Will Small

Powers of various sorts—capacities, abilities, and skills—play a large role in ordinary thought and conduct concerning human agency.1 Their acquisition, maintenance, and development are at issue in many developmental, educational, and therapeutic contexts. Examples of the kind of power I will discuss include the abilities to walk, to speak a first (or second) language, to juggle, to play the piano, to sail, to ride a bicycle, to operate a lathe. (Some of these powers are more naturally described as abilities, some as skills; that distinction will be discussed below.) Call them agential powers. Characteristically, agential powers are not merely manifested, but exercised (contrast the power of the sun to melt a frozen puddle). And characteristically the exercises of agential powers are intentional actions. In ascribing such powers to agents, we typically talk about what those agents can do. The same power may be exercised on different occasions, and it persists between exercises. Agential powers can, however, be lost in various ways; most—arguably all—are acquired. So even when an agential power is not currently being exercised, it would appear to have some kind of reality, unlike mere possibility. An agential power would seem to be a ‘positive explanatory factor in accounting for the performance of an agent’ (Kenny 1975: 133), again unlike mere possibility: possibly p does not contribute to an explanation of its being the case that p, whereas an agent’s ability to juggle contributes to an explanation of its being the case that she is juggling. This entry considers some of the most important questions concerning the nature and specification of agential powers, and their place in an account of human agency.

1 Agential powers and the explanation of action Despite the seemingly great theoretical, practical, and metaphysical significance of agential powers, relatively little attention has been paid to them by recent work in the philosophy of agency.2 This may be because much work on the nature of exercises of agency has focused on the explanation of action, specifically on explanations of action in terms of the agent’s reasons. Asking why an agent acted as she did (e.g. why she crossed the street) is likely to uncover her intention, desire, or motive in so acting (e.g. to get to the café) and what she knew or believed about the world that connected what she did to that motivating state (e.g. that crossing the street was a way of getting to the café). But her ability to walk (if it was by walking that she crossed the street) is part of the background that this explanation assumes; the 130

DOI: 10.4324/9780429202131-15

Agency, powers, and skills

explanation doesn’t make it salient. And when a reasons explanation of what someone did is used as the template for a metaphysical account of what it is for an event to be an action—as it is by the so-called standard story of action, according to which, in its simplest form, actions are bodily movements caused (in a way to be specified such as to exclude so-called ‘deviant causation’) by belief-desire pairs—the fact that in crossing the street the agent exercised her ability to walk (and perhaps, if it was a busy street, a more specific traffic-negotiating ability or skill) is left by the wayside, seemingly irrelevant to the nature of actions and the form of explanation characteristic of them (Hornsby 2004; Small 2017a).3 However, agential powers come into view when we consider that which a reasons explanation aims to reconstruct: an agent’s practical reasoning (or possible course thereof ). The agent wants to get to the café, on the other side of the street. Her deliberation aims to deliver, minimally, a practicable answer to the question ‘How am I to get there?’ Among other things, her conception of her agential powers plays a constraining role: she knows she cannot fly, but she knows she can walk, and that she can cross by walking. Many theories of intentional action (including the standard story, but also many views that oppose it) maintain that deliberation can come to an end successfully only when it yields something an agent can do ‘just like that’—something for which the question ‘How am I to do that?’ doesn’t arise. When an agent does something ‘just like that’—without doing anything else as a means to doing it—she is said to have performed a (teleologically) basic action.4 So one natural place at which agential powers might come into an account of intentional action is in an account of basic action. For instance, a causal theorist might hold that the belief in the belief-desire pair identifies an ultimate means, something the agent can do ‘just like that,’ and the belief-desire pair causes a bodily movement by ‘activating’ an agential power, such as the ability to walk (in something like the way in which striking a glass triggers its disposition to break); another agent could desire and believe the same things, but if she lacked the relevant agential power, her belief-desire pair would not cause a bodily movement of the relevant type. Alternatively, a causal theorist who gives intentions an essential role might hold that the ultimate intention activates and guides the exercise of the relevant agential power (Clarke 2010).5 According to these conceptions of basic action, agential powers take over where practical thought and deliberation give out. But there is something paradoxical about this (Lavin 2013). In exercising an agential power, the agent is supposed to be acting intentionally; yet she can have no idea how she’s doing what she’s doing. One conclusion that might be drawn from this is that the agential powers whose exercises are basic actions cannot be ‘alien’ to practical reason; the powers must be ‘intelligent powers’ (see below) which are such that their exercises are known by their agents (Small 2019). Theorists of basic action disagree about what can be done ‘just like that.’ According to some (e.g. Hornsby 1980), tying one’s shoelaces and riding a bicycle might be basic actions, suggesting that agential powers to do such things figure as fundamental ‘positive explanatory factors’ in agents’ performances. However, others (e.g. Danto 1963) think these would not be basic actions because they are (supposedly) done by means of making certain bodily movements; this suggests that such powers as the ability to tie one’s shoelaces and the skill of riding a bicycle do not ultimately contribute to the explanation of agents’ performances, and that the only agential powers that are genuine positive explanatory factors are an agent’s powers of bodily movement. (Note that this dispute is not typically framed in terms of the nature and explanatory significance of agential powers, but rather in terms of the nature of basic action and the specificity of practical thought.) However, agential powers seem to have explanatory significance beyond that which theories of basic action accord them. First, there are many skills and abilities whose exercises 131

Will Small

involve deliberating about how to do things, notably when skills comprise subskills, and the agent might achieve her end in exercising her skill in different ways, using different combinations of exercises of its subskills (consider carpentry.) Secondly, agential powers—and skills, in particular—frequently include perceptual dispositions: becoming a skilled tennis player involves not only developing the ability to hit a top-spin cross-court forehand, but also becoming disposed to recognize opportunities for doing so. Indeed, in the case of many skills, an agent’s expertise is manifested primarily in what she sees and in how she determines what to do, the doing itself being something that an unskilled agent could often do with little difficulty (consider chess.) These aspects of agential powers raise important questions for philosophers of agency. What, if anything, is distinctive about the practical reasoning of the bearer of a skill? Can the deliberative, perceptual, and practical aspects of agential powers and their explanatory significance be understood independently, or must a satisfactory account treat them together?

2 The specification of agential powers In virtue of possessing agential powers, it may be said of an agent that she can do certain things, or that she is able to do them. But the converse is not the case: that someone cannot (or is not able to) φ does not entail that she lacks the ability to φ (or skill of φ-ing, etc.). She may lack the authority (in the right context, a registrar can effect a marriage by saying certain words and signing certain documents; someone lacking the authority cannot, even if she has the abilities to say and do the relevant things). An agent may lack the opportunity to φ (I cannot cook an omelet without any eggs). And sometimes an agent cannot φ because, even though she has the ability to φ, her exercise of it is impeded (and not by removing the opportunity to φ): a sprained wrist or stage fright might prevent a skilled pianist with ready access to a piano from playing.6 Some philosophers who discuss abilities distinguish between ‘general’ and ‘specific’ abilities (e.g. Mele 2003). I am said to retain my ‘general ability’ to cook an omelet in the absence of eggs, but to lack the ‘specific ability’ to cook an omelet because I cannot cook one ‘right now.’ As Maier (2015: 123) puts it, when an agent has a specific ability to do something, “there is, as it were, nothing between her and the deed.” But as this distinction is usually understood, the ascription of a specific ability to φ is simply the ascription of both the (‘general’) ability to φ and the opportunity to φ; if this is so, the term ‘specific ability’ is something of a misnomer, for it does not refer to a kind of ability at all.7 There is, however, a real question about the specification of agential powers. It is common for philosophers to talk about ‘the power (capacity, ability) to φ,’ or ‘the skill of φ-ing.’ But what does the variable ‘φ’ range over in such expressions—what are its legitimate substitution instances? Clearly not particular actions (for they are the paradigm exercises of agential powers, and the same power may be exercised on different occasions in different actions), but rather something general: act-types, perhaps.8 But what are act-types? According to one influential view (Davidson 1967; Hornsby 1980), what is done on an occasion is an act-type, and the doing of it on an occasion is a token action—or, as I’ll say, an action. A sentence reporting an action (someone’s doing of something on an occasion—a dated, unrepeatable particular) thereby specifies what that person did (something they might have done before or do again, something someone else might do or have done). Thus, ‘Jones buttered the toast in the bathroom with a knife at midnight’ reports a particular action, but what Jones is thereby said to have done—namely, butter the toast in the bathroom with a knife at midnight—is something he might do every night, and that Smith might do too. An action sentence typically entails 132

Agency, powers, and skills

further action sentences (e.g. ‘Jones buttered something in the bathroom,’ ‘Jones buttered some toast in a bathroom,’ ‘Jones buttered the toast with a knife’) each of which specifies a thing done, something that might be done again by the agent or someone else. There are good reasons to deny that any act-type is, in principle, a legitimate substitution instance for ‘φ’ in ‘the power to φ,’ ‘the skill of φ-ing,’ and so forth. In acting as he is reported to have done, did Jones exercise the power to butter the (that? or some?) toast in the bathroom with a knife at midnight? Did he also exercise the power to butter toast with a knife and the power to butter toast at midnight? It is implausible to maintain that Jones possessed, let alone exercised, a distinct power corresponding to each of the (perhaps indefinitely) many act-types he instantiated. The ability to use a knife for spreading is something Jones acquired, probably as a child, over some more-or-less determinate period of time. But when did he acquire the ability to butter toast in the bathroom with a knife at midnight? If he ‘counts’ as possessing that ability in virtue of his performance, how could that ability have been a ‘positive explanatory factor’ in the performance? Wouldn’t he have needed the ability before acting? Moreover, nothing further by way of ability (as opposed to opportunity and motivation, etc.) is needed to explain Jones’s buttering of the toast in the bathroom with a knife at midnight than his ability to use a knife for spreading; this ability is sufficient also to explain9 Jones’s spreading jam on a croissant at breakfast and clotted cream on a scone at tea. Typically, then, the proper specification of an agential power will be more coarse-grained than the specification of what was done on an occasion of its exercise. If this were not so, it would be impossible for finite agents to be in a position to perform actions of infinitely many types (Small 2017a).

3 Skills versus abilities? Skills are connected with knowledge, whereas there seem to be agential powers (call them ‘mere abilities’) that are not. Certainly not every agential power would be called a skill (e.g. the ability to wiggle one’s ears, the ability to do 50 push-ups). The word ‘skill’ (and its cognates) has an honorific use, which may be withheld due to a contextually salient lack of complexity in the relevant activity or (relative) lack of competence on the part of the agent. Further, not every case of one agent’s being a more able φ-er than another would be aptly characterized in terms of her being more skilled (the difference might be one merely of e.g. speed or strength).10 How is intelligence (cognition, knowledge) related to agential power in skill? Two main strategies present themselves, though variations of each are possible. According to the first, skill is a hybrid, comprising mere ability plus cognitive states suited to guide its exercise (see e.g. the ‘intellectualist legend’ criticized by Ryle 1949: ch. 2; and for a more sophisticated contemporary expression, Stanley and Krakauer 2013).11 According to the second, a skill is an agential power of a specifically intelligent form (see e.g. Ryle 1949: ch.2; Small 2019).12 On the first view, a skill comprises one or more agential power (mere ability) and something that is not an agential power (cognitive states, which are typically, though not necessarily, construed as propositional attitudes); the cognitive states are intelligent, whereas the mere abilities are not: a skill counts as intelligent in virtue of the cognitive states it comprises. On the second, skills are intrinsically intelligent agential powers: a skill is both a distinctively practical form of knowledge and a distinctively intelligent form of causal power. The second view may seem to imply that mere abilities and skills are kinds of agential powers that are disjoint—skills the intelligent agential powers, mere abilities the unintelligent ones (Annas 2011). But it is consistent with it to construe mere abilities as privative cases of skill, where there is perhaps very little to learn (Small 2020). 133

Will Small

4 Are agential powers… 4.1 …dispositions? Many philosophers think that agential powers are dispositions. Like abilities, dispositions are modal properties and may be ascribed using ‘can.’13 Perhaps agential powers are dispositions, possessed by an agent, the manifestations of which are intentional actions. Dispositions are usually individuated by both their stimuli and their manifestations: for example, X is fragile iff X is disposed to break if struck; X is soluble iff X is disposed to dissolve when placed in water; and so forth.14 If agential powers are dispositions, what are their stimuli and manifestations? According to one approach (e.g. Vihvelin 2013), to have the ability to φ is to be disposed to φ intentionally when one tries (or intends) to φ; one exercises one’s ability to φ when one φ-s intentionally as the upshot of trying (or intending) to φ. This proposal faces problems. Though I have suggested that the exercises of agential powers are characteristically intentional actions, the present proposal would seem to exclude altogether the possibility of non-intentional (including unintentional) exercises of agential powers, and this may seem too strong (Vetter 2019). Moreover, according to some, not every case of φ-ing intentionally requires the agent to try to φ; according to others, when an agent φ-s intentionally her trying to φ is not distinct from her φ-ing (see e.g. Hornsby 2010). Finally, the so-called ‘Simple View’ that intentionally φ-ing requires an intention to φ is rejected by many (e.g. Bratman 1987); and even when an agent who is φ-ing intentionally does intend to φ, if her intention is construed as ‘stimulating’ the agential powers involved in her φ-ing, it seems the picture is not of her exercising her powers, but merely of her powers being manifested—the agent ‘disappears’ from the execution of her intention (Hornsby 2004). According to another approach (e.g. Sosa 2015), to have the ability to φ is to be disposed to φ well or successfully when one φ-s. It is true that explicit ascriptions of abilities and (in particular) skills—that is, ascriptions in which ‘ability’ or ‘skill’ figure, as opposed merely to, for example, ‘can’—often connote that the agent is good, and perhaps better than many or even most at φ-ing. But what is it for A to be better at φ-ing than B? There does not seem to be a single answer: for instance, in some cases, it is for A’s attempts to φ to more often result in φ-ings than B’s; in others, it is for A to be able to φ in a more difficult range of circumstances than B; in others, it is for A’s φ-ings to be better instances of φ-ing than B’s (Small 2017a). Further, there are some kinds of action (those predicated using ‘success verbs’) such that, if they are engaged in at all, they are engaged in successfully; yet (as Vetter 2019 argues) the present view seems to imply that, where ‘to ψ’ is a success verb, every agent has the ability to ψ, because every agent is such that, were she to ψ, she would ψ successfully. But this is absurd, because surely some agents do in fact lack agential powers that are specified using success verbs.

4.2 …two-way powers? The above proposals both assume that agential powers differ from other dispositions only by having a special kind of bearer, stimulus, and/or manifestation (an agent, intention, attempt, intentional action, successful intentional action): the formal character of the relation between disposition and manifestation is the same in an agential power as it is in fragility.15 However, some have proposed that agential powers are powers of a distinctive form: they are ‘two-way’ powers, whereas non-agential powers are ‘one-way’ powers. Whereas a one-way power has one kind of exercise (fragility is manifested in breaking), a two-way power can be exercised in two ways. 134

Agency, powers, and skills

What are the two-ways in which an agential power can be exercised? According to some (e.g. Kenny 1975; Alvarez 2013), a two-way power to φ may be exercised in φ-ing and in refraining from φ-ing.16 But though an agent could properly be said to have refrained from φ-ing only if she possessed the power to φ, it does not follow that her so refraining was an exercise of that power; it is natural to think she precisely refrained from exercising it. In fact, it seems that talk of a ‘two-way’ power is a distraction here: what is crucial to this conception of agential powers is that, unlike in the case of non-agential powers, ‘when the conditions for the exercise of the power obtain, the power need not be manifested’ (Alvarez 2013: 102). Rather, it is up to the agent whether or not to exercise her agential power when she has the opportunity to do so—this is something she determines through her choice. This proposal may seem similar to the view that to have an agential power to φ is to be disposed to φ when one so intends. But they are very different: that view treats the agent’s intention as among the conditions necessary for the exercise of the power (the totality of conditions necessary for the disposition’s manifestation are sufficient for its manifestation), whereas this proposal insists that the conditions necessary for the power’s exercise are not sufficient for its exercise, and thus that the agent’s choice or intention is no such condition or circumstance (see Kern 2017: 163ff ). On an alternative conception, deriving from Aristotle (Metaphysics Θ.2), the two ways in which a two-way power may be exercised are in φ-ing and ‘contra-φ-ing’ (Makin 2006: 44): the doctor can exercise her medical skill either in healing the patient or in harming (‘contra-healing’) him, and it is up to her not only whether, but also how, to exercise her power.17 What explains the possibility of contra-φ-ing (and thus the ‘two-wayness’ of twoway powers), on this view, is that two-way powers are partly constituted by knowledge of how to φ, which implies knowledge of what not to do in φ-ing—the knowledge exercised in contra-φ-ing. Skills are therefore plausible candidates to be two-way powers in this sense; whether mere abilities are two-way powers will depend on one’s conception of the skill/ ability distinction (see above).

5 Conclusion There is greater sympathy for powers in contemporary philosophy today than there was in the second half of the 20th century, when many of the analytic tradition’s most influential contributions to the philosophy of agency were written. Clarke is surely correct: “Abilities are fundamental to agency; we don’t have a decent comprehension of agency without an understanding of them” (2015: 893). But there remains much work to do in order to understand agential powers, abilities, and skills. One key issue, not discussed here, is the significance of agential powers to the theories of free will and moral responsibility, which are often concerned with the question of whether someone who φ-d could have done, or had the ability to do, otherwise. I hope to have brought out in this entry that the significance of agential powers for the philosophy of agency extends beyond these concerns, by focusing on the possibility that agential powers play a fundamental role in our understanding of the nature of intentional action. I have sought to draw attention to a number of important questions that remain unresolved. Is the role of agential powers in a satisfying philosophical account of the nature of intentional action restricted to cases of so-called ‘basic actions’? What constraints would recognizing a fundamental role for agential powers in such an account impose on our conceptions of intentional action and of basic action in particular? How coarse or narrow should our canonical specifications of agential powers be, and why? Are skills modes of practical intelligence insofar as they combine abilities 135

Will Small

with knowledge, or is their intelligence a matter of the specific kind of abilities that they are? What, precisely, is the difference between a practical ability and a practical skill? How do agential powers differ from natural powers, or dispositions: do they differ only in their stimuli and manifestations or (also?) in their form?

Related topics Agency and causation; Agency, events, and processes; Expert agency; Agency and disability; Agency and mistakes; Pathologies of agency.

Notes 1 Not every manifestation or exercise of a power possessed by a human agent is a case of agency (e.g. the capacities to understand French, to tolerate spicy food, to fall asleep on buses). But such powers are not at issue here. Nor are the agential powers of nonhuman agents. 2 An important exception is the central role given to ‘embodied coping skills’ by the ‘analytic phenomenological’ tradition of Hubert Dreyfus and his followers, which lack of space precludes me from discussing. Note that Dreyfus (2001) claims that exercises of such skills are, as such, not intentional actions. 3 Certain dissatisfactions with the standard story encourage some to introduce not only a wider range of mental states but also various kinds of mental powers into their accounts of what must happen causally upstream of a bodily movement for it to qualify as an action—e.g. willpower, or capacities for instrumental rationality and self-governance. Though such capacities play an integral role in our lives as agents, they are not agential powers in the sense at issue here, for their exercises are not themselves (characteristically) intentional actions. 4 More precisely, given certain views about the individuation of action, the description of her action under which it is true to say that she ___-ed ‘just like that’ is the basic description of her action. 5 Can an intention to perform a basic action can guide the performance, if (by definition) it contains no information about how the action is to be performed (Small 2019)? 6 Though one might say that the agent cannot (or isn’t able to) exercise her ability to φ, this should not suggest that she has lost the ability to exercise her ability to φ, an ability she will regain when she is no longer impeded in exercising her ability to φ. 7 The same concern applies to the view that a third ‘kind’ of ability should be distinguished, to be ascribed to an agent who, though possessing the ability and opportunity to φ, cannot φ because of some physical or psychological impediment to the exercise of her ability (e.g. the ‘narrow ability’ of Vihvelin 2013). However, see Maier (2015) for an attempt to explain general ability in terms of specific ability; for criticism, Clarke (2015: 894). 8 Though note that there is not always a direct route from an ascription of skill (e.g. ‘is a skilled carpenter,’ ‘is a skilled doctor’) to any act-types. 9 More precisely, to contribute what agential powers contribute to the explanation of action. 10 Suppose Amy can do 50 push-ups, whereas Ben can do only 30. If the remarks about the specification of agential powers in the previous section are correct, we should say, not that Amy has an ability Ben lacks (the ability to do 50 push-ups), but rather that she possesses the ability they each have (to do push-ups) to a greater degree. 11 Stanley and Williamson (2017) contend that skills are dispositions to form (or activate) knowledge suitable to guide action—and thus, presumably, to guide the exercise of mere abilities (see Riley 2017 for criticism). As some such knowledge is, arguably, ability-entailing, the possession of a skill entails the possession of relevant mere abilities, and their proposal can be viewed as an idiosyncratic version of the hybrid view. 12 Much recent discussion of the issue focuses on whether what skilled agents know (as such) can be exhaustively represented as propositional knowledge; for discussions that focus on the role of skill in the explanation of action, see Fridland (2013, 2014) and Small (2017b, 2019). Phenomenologists such as Dreyfus can be read as adopting the second strategy distinguished in the text, though the views of Dreyfus (e.g. 2001) are importantly different from those of e.g. Ryle (1949); see Small (2017b) for discussion.

136

Agency, powers, and skills 13 However, some dispositions (such as character traits) are not capacities, but rather tendencies; they are more aptly ascribed using a habitual expression (‘does φ,’ ‘φ-s’). 14 Though Vetter (2015) maintains that dispositions are individuated solely by their manifestation-types. 15 The same goes for the dispositional conception of skill proposed by Stanley and Williamson (2017; see n. 11 above), though the distinctive manifestation there is not any kind of action but rather the forming/activation of knowledge apt to guide action. 16 Indeed, Alvarez contends that ‘there is human agency whenever there is the exercise of … a twoway causal power’ (2013: 102); but because certain ‘not-doings’ are supposedly among the exercises of two-way powers, this means that not every case of human agency is a case of intentional action. 17 For contrasting conceptions of Aristotle’s conception of two-way powers, see e.g. Beere (2009); Kern (2017: ch. VI); Frost (2020); Small (2021).

Further reading Alvarez, M., 2013. Agency and Two-Way Powers. Proceedings of the Aristotelian Society (113), 101–121. An important statement of the ideas that agential powers are powers of a distinctive form, and that manifestations of human agency should be understood in terms of the exercises of such powers (rather than in terms of the occurrence of intentional actions). Clarke, R., 2015. Abilities to Act. Philosophy Compass (10), 893–904. An overview of recent literature on abilities, primarily focusing on texts that concern the connection between abilities and freedom of the will. Kenny, A., 1975. Will, Freedom and Power. Oxford: Basil Blackwell, ch. VII. An important discussion of abilities, and of how they differ from and relate to opportunities. Ryle, G., 1949. The Concept of Mind. London: Hutchinson, chs.2, 5. The locus classicus in the analytic tradition of an approach to agency that gives pride of place to powers and skills. Small, W., 2017a. Agency and Practical Abilities. Royal Institute of Philosophy Supplements (80), 235–264. A discussion of recent approaches to abilities, paying special attention to the relationship between their specification and their role in the explanation of action.

References Alvarez, M., 2013. Agency and Two-Way Powers. Proceedings of the Aristotelian Society (113), 101–121. Annas, J., 2011. Practical Expertise. In J. Bengson and M. Moffett eds., Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press, 101–112. Beere, J., 2009. Doing and Being: An Interpretation of Aristotle’s Metaphysics Theta. Oxford: Oxford University Press. Bratman, M., 1987. Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Clarke, R., 2010. Skilled Activity and the Causal Theory of Action. Philosophy and Phenomenological Research (80), 523–550. ———, 2015. Abilities to Act. Philosophy Compass (10), 893–904. Danto, A., 1963. What We Can Do. The Journal of Philosophy (60), 435–445. Davidson, D., 1967. The Logical Form of Action Sentences. Reprinted in his Essays on Actions and Events, 105–148. Oxford: Clarendon Press (1980). Dreyfus, H., 2001. Phenomenological Description versus Rational Reconstruction. Revue International de Philosophie (55), 181–196. Fridland, E., 2013. Problems with Intellectualism. Philosophical Studies (165), 879–891. ———, 2014. They’ve Lost Control: Reflections on Skill. Synthese (191), 2729–2750. Frost, K., 2020. What Could a Two-Way Power Be? Topoi (39), 1141–1153. Hornsby, J., 1980. Actions. London: Routledge and Kegan Paul. ———, 2004. Agency and Actions. Royal Institute of Philosophy Supplements (55), 1–23. ———, 2010. Trying to Act. In T. O’Connor and C. Sandis, eds., A Companion to the Philosophy of Action, Chichester: Wiley-Blackwell, 18–25. Kenny, A., 1975. Will, Freedom and Power. Oxford: Basil Blackwell.

137

Will Small Kern, A., 2017. Sources of Knowledge: On the Concept of a Rational Capacity for Knowledge. Trans. D. Smyth. Cambridge, MA: Harvard University Press. Lavin, D., 2013. Must There Be Basic Action? Noûs (47), 273–301. Maier, J., 2015. The Agentive Modalities. Philosophy and Phenomenological Research (90), 113–134. Makin, S., 2006. Aristotle’s Metaphysics Book Θ. Oxford: Clarendon Press. Mele, A., 2003. Agents’ Abilities. Noûs (37), 447–470. Riley, E., 2017. What Skill is Not. Analysis (77), 344–354. Ryle, G., 1949. The Concept of Mind. London: Hutchinson. Small, W., 2017a. Agency and Practical Abilities. Royal Institute of Philosophy Supplements (80), 235–264. ———, 2017b. Ryle on the Explanatory Role of Knowledge How. Journal for the History of Analytical Philosophy (5), 57–76. ———, 2019. Basic Action and Practical Knowledge. Philosophers’ Imprint (19), 1–22. ———, 2020. Practical Knowledge and Habits of Mind. Journal of Philosophy of Education (54), 377–397. ———, 2021. The Intelligence of Virtue and Skill. Journal of Value Inquiry (55), 229–249. Sosa, E., 2015. Judgment and Agency. Oxford: Oxford University Press. Stanley, J., and Krakauer, J., 2013. Motor Skill Depends on Knowledge of Facts. Frontiers in Human Neuroscience (7), 1–11. Stanley, J., and Williamson, T., 2017. Skill. Noûs (51), 713–726. Vetter, B., 2015. Potentiality: From Dispositions to Modality. Oxford: Oxford University Press. ———, 2019. Are Abilities Dispositions? Synthese (196), 201–220. Vihvelin, K., 2013. Causes, Laws, and Free Will: Why Determinism Doesn’t Matter. Oxford: Oxford University Press.

138

12 EXPERT AGENCY Barbara Gail Montero

My fingers emit sparks of fire with expectations of my future labours. —William Blake

Think back to a time when you learned a new skill. Perhaps it was speaking a foreign language, playing the saxophone, or swimming. Let’s take swimming. At first, the prospect of making it across the pool was daunting. Legs would kick; arms would thrash; yet rather than being propelled forward, you plummeted. Try as hard as you could, you’d end up clutching the side of the pool, white-knuckled, while you caught your breath. But with practice, it became easier, and now you can shimmer across without thinking about it, without exertion, without even being aware of what you are doing. What had once been onerous is now easy and effortless. Your experience of learning to swim exemplifies a common view about skilled agency: that the progression from novice to expert follows a trajectory from the effortful to the effortless, ultimately becoming so effortless that the agent disappears. Your expert actions are not done by, but rather happen to you. To put it somewhat paradoxically: when you are highly proficient at a task, you can perform it without you even being there. The view that as expertise develops, effort diminishes is widespread. It plays a role in Guthrie’s (1952) classic definition of expert skill as “the ability to bring about some end results with maximum certainty and minimum outlay of energy, or of time and energy” (p. 136);1 it finds its way into Wulf and Lewthwaite’s (2010) thought that “relative effortlessness is a defining characteristic of [expert] motor skill” (p. 75); and it is supported by at least some people’s experience of learning a skill: as Millican and Wooldridge (2014) observe, “a great many of the actions that we perform intentionally are done without explicit consciousness of them, and the more expert we become at a skill (such as driving, riding a bike, typing, or playing the piano), the more likely we are to perform the actions that it involves with minimal consciousness of what we are doing (and indeed trying to concentrate on what we are doing is quite likely to disrupt our performance).” This inverse relation between skill and effort is also sometimes thought to find support in controlled studies (Beilock et al. 2003; Dugdale and Eklund 2003), phenomenological investigations (Dreyfus and Dreyfus 1986),

DOI: 10.4324/9780429202131-16139

Barbara Gail Montero

Csikszentmihalyi’s work on flow (1990), and some of Taoism’s central texts (Slingerland 2007; Velleman 2008). A common theme in these investigations is that, at its pinnacle, expertise is effortless and agentless. I plan to argue, however, that support for this theme is weaker than is often thought and to offer an opposing theme: effortful expertise.

1 The hypothesis of effortful expertise According to the hypothesis of effortful expertise, the expertise characteristic of professional athletes, artists, doctors, and scholars—or, more broadly, of anyone who achieves their professional status through hard work and continues to practice their skills with the intent of improving them—is generally effortful; experts generally act effortfully when engaging in actions in their domain of skill, and their effort does not generally interfere with their skills. Furthermore, although agency and effort are not the same thing—you might experience agency in performing very easy tasks—because my concern is with felt agency (the feeling that you are doing something) and consciously experienced effort (the feeling that you are making an effort to do something), effort drags agenthood into expertise: if an agent is consciously aware of making an effort to achieve a goal, agency is at work.2 Thus, the thesis of effortful expertise implies the thesis of agentful expertise; it implies that, in optimal action, there (generally) is a you there. Pacherie (2007, p. 195) distinguishes a reflective conception of agency from a “minimal sense of agency.” The minimal sense is nonreflective and “fully immersed,” while the reflective conception comes in two forms: either taking a third-person stance toward one’s own actions or taking a first-person stance and being aware of the thoughts and experiences one has in action. Though I shall not concern myself with the third-person stance, the effort of expertise that concerns me is frequently the “reflective sense of agency.”

2 Everyday expertise is different from professional-level expertise Your experience with swimming may suggest that the better you get at performing a skill, the easier it is to perform it. But I’m assuming you’re not an Olympic swimmer since, according to hypothesis of effortful expertise, if you were this would not be the case. As I see it, there is a significant distinction between, on the one hand, “everyday expertise”—the type of expertise you have in tying your shoes, climbing the stairs, and, for some of the recreational swimmers among you, swimming—and, on the other hand, “professional level expertise”—the expertise characteristic of professional athletes, artists, doctors, scholars, and others who achieved their professional status through hard work and continue to practice their skills with the intent of improving them. Everyday expertise is largely built up out of relatively mindless repetition. When first learning how to tie your shoes, you were taught what to do. And you likely struggled to wrap one loop around the other. Yet you have tied your shoes so many times now that the actions have become so proceduralized that they can be performed effortlessly and, arguably, without conscious thinking about them (see Bargh and Chartrand 1999). Furthermore, it may be that thinking about tying your laces while you are tying them hinders your shoelace-tying prowess. Everyday action, at least at times, seems to crumble under real-time reflection. Ponder how to type in your password while attempting to do so and you’ll end up needing to call the help desk. And it is easy to see how an evolutionary advantage could accrue to those who could think about more important things during grooming and other such activities. Thus, everyday activities, for all I shall argue, may fit into the effortless expertise mold.3 140

Expert agency

Although fitting the mold means that these skills are typically performed effortlessly and that effort does not generally hinder them, it doesn’t mean that reflecting on them can’t benefit their performance. You can improve the way you tie your shoes, or walk, or even reach for the shampoo, and to do so, you will need to focus on what you are doing and make an effort (Shusterman 2008). As Richard Shusterman argues, attention is required to improve everyday skills, but, ultimately, their execution should become invisible. The agent might not disappear: you may still have a sense of making an effort when walking to work, for example, as you effortfully consider how to respond to the plaintive email from your colleague who was just denied the salary bump you received. Nonetheless, as long as you are not pushing yourself to walk faster than usual, the walking itself often seems to be performed effortlessly, without any attention to what is being done. Professional-level expertise, however, is different since such experts have not only reached professional status but are keen on improving. Might there be people whom we would call “experts” and who have dedicated their lives to an activity without attaining professional status? And might some professionals slip through the cracks, attaining their position through, say, nepotism rather than hard work and thus, even if they are set on improving, are rather lacking at anything we would want to call expertise until they do improve? “Yes” and “yes,” but rather than expatiating on the concept of expert let me say that, in general, experts, for present purposes, are those who have attained professional status and manifest a continued desire to improve (see, however, Montero 2016, 2018). While we are happy to let our everyday skills remain at the same good-enough level, the expert does not settle for good enough; for the professional-level soccer player, for example, there is not something more important to focus on during a tournament than the game at foot, and the drive to win may be seen as almost a life or death matter. And for current purposes, this makes for a significant difference between the mindset someone engaging in everyday skills and the mindset of someone engaging in professional-level skills. Nonetheless, some take investigations into everyday expertise to reveal the nature of professional expertise. For example, a central proponent of the effortless expertise hypothesis, Hubert Dreyfus, in the research he conducted with his brother, Stuart Dreyfus (Dreyfus and Dreyfus 1986), generalizes from driving a car to piloting a military airplane. As he puts it, “once Stuart had worked out the five stages [of expertise] using his driving skills as his example, we just changed car to plane and driver to pilot and wrote a report for the Air Force” (p. 32). Yet Stuart was a lay driver. And, for better or worse, everyday driving is not a skill we care enormously about. It is a life or death matter, but unlike air force pilots, we do not treat it as such, and so would rather chat with the person in the passenger seat than give it our all. And because driving it is not something that we train to improve (apart from an initial learning period and the modicum of studying required to renew a license), it is not well suited as a generalization base for theories about professional-level expertise. Professional-level experts not only approach their tasks differently but also train for them differently. Research by Anders Ericsson (2008) indicates that when one is already at a high level of achievement, improvement requires “deliberate practice,” which involves “[d]eliberate efforts to improve one’s performance beyond its current level [and] demands full concentration and often requires problem-solving and better methods of performing the tasks” (p. 31); deliberate practice is “very high on relevance for performance, high on effort, and comparatively low on inherent enjoyment” (Ericsson et al. 1993, p. 373). For example, a dancer might engage in deliberate practice when she works time and time again on a certain step, analyzing what went wrong and trying to incorporate corrections (from herself and others). A chess player’s deliberate practice might encompass, among other things, studying 141

Barbara Gail Montero

past games, exploring lines, and working on openings. As such, deliberate practice involves pushing yourself beyond what you can already do.4 While the proceduralization of everyday actions leads, according to Ericsson (2008), to stagnation, or “arrested development,” deliberate practice prevents the attainment of automaticity and thus allows for ongoing improvement. “The key challenge for aspiring expert performers,” Ericsson tells us, “is to avoid the arrested development associated with automaticity,” and the way to accomplish this is by “actively setting new goals and higher performance standards, which require them to increase speed, accuracy, and control over their actions” (p. 991). Setting new goals and higher performance standards requires making an effort. Yet, although deliberate practice is often effortful and analytical, deliberate practice is not contrary to the idea of effortless and agentless expertise, which concerns the performance of a skill rather than the practice of it. Rather, it is not contrary to it unless practice, itself, is a form of expertise, which, arguably, it is as one may get better at the skill of practicing and try to get better at it. If this is correct, there is at least one type of expertise—expertise in practicing— that even the proponents of effortless expertise hypothesis would agree is neither effortless nor agentless. More importantly, effortful, consciously focused training seems to open the door to effortful, consciously focused performance. If you practice performing a skill while thinking about what you are doing, then thinking during performance should feel natural. If you think about how to type in your password, you may be paralyzed; however, if you were to practice thinking while you type, after a while your thinking will, presumably, become coupled with action in such a way that you may be thrown off by not thinking. In fact, performance—at least for skills, such as performing in music or dance, that do not have as their ultimate goal winning—may incorporate an element of practice. The ballerina Natalia Makarova claimed that for her, “it is physically impossible … to do the same performance twice” (Sherman 1980/1977). What she means by “physically impossible” is not simply that, because the human body has so many degrees of freedom, one could never perfectly reproduce a movement even if one tried. Rather, she cannot bring herself to try to reproduce it again. This may be in part because such an approach makes performing more enjoyable, but it is also in part because she is always trying to improve.

3 What post-performance amnesia does not imply We are all familiar with the long-distance truck driver who, after a stretch of road, realizes that she had been driving without any awareness of what she had been doing (Armstrong 1981). She realizes this, presumably, because she has no memory of her actions. But what, exactly, does a memory gap show? It is tempting to think that highly skilled performances are sometimes not remembered because experts perform on autopilot: the truck driver was not unconscious, but she experienced a limited form of consciousness, one that did not include consciousness of agency. Post-performance amnesia apparently turns up in a wide range of skills. See Simon Høffding’s (2018) interviews with members of the Danish String Quartet (DSQ) and the anecdotes in (Bermúdez 2017; Hoffding and Montero 2020). It is well illustrated by the Danish ballet dancer Eric Bruhn’s account of his utter lack of memory of one performance of Giselle. After leaving the stage, he reports, he became suddenly afraid that everything had gone terribly wrong since he had no recollection of his dancing or of anything that occurred while he was performing (Meinertz 2008, p. 117). Moreover, some experiment-based studies of post-performance amnesia purport to indicate an inverse relationship between athletes’ level of expertise and the extent of their memories for performance in their domain of skill, that 142

Expert agency

is, they seem to indicate that the better you are at performing a skill, the less you will recall about any particular performance of that skill (Beilock and Carr 2001 and Beilock et al. 2002; though see Hoffding and Montero 2019 for a criticism). Although agency may sometimes disappear in skilled action, post-performance memory failures, themselves, do not entail that agents have been on holiday since it is possible to consciously attend to one’s actions yet forget them immediately afterward. Individuals who suffer from anterograde amnesia provide a dramatic example: they are unable to form new long-term memories yet appear to have abundant conscious declarative thoughts in action (Baddeley 2014). Post-performance amnesia is clearly not a case of clinical amnesia, but I would like to suggest that something analogous may be occurring. Armstrong is struck by the fact that long-distance truck drivers can manage the complex sub-routines of driving yet may not be conscious of driving and so he implicates “unconscious attention” to account for tasks; however, I would like to suggest that when performance is complex and environments are uncertain, memory failures might be caused not by online attentional failures, but, somewhat ironically, by intense attention during action. It is known that a distraction task immediately following a memory task interferes with long-term memory formation. In Muller and Pilzecker’s (1900) pioneering study, the participants, after they attempted to memorize a list of syllables, were presented with a new syllable list either 17 seconds or 6 minutes later. In the 17-second condition, participants recalled 28% of the syllables, whereas in the 6-minute condition this increased to 49%. This and numerous recent studies of both humans and nonhuman animals, using both behavioral and neuroimaging approaches, seem to support the idea that distraction immediately following a memory task can serve as a “retroactive interference” impeding long-term memory formation (see, for example Dewar et al 2009 and, for a review, Wixted 2004). Perhaps something similar occurs in post-performance amnesia: experts fail to recall what they have just done, not because they were missing in action but because their conscious attention was so keenly focused on the moment that each subsequent period of attention impedes memory formation of the previous period. It also may be that post-performance amnesia occurs when performers are highly focused on aspects of their skills that are not readily expressible in words. If conscious awareness of our movements and positions outstrips what we can presently be put into words, it may be that an exclusive focus on nondeclarative elements or qualities of our movements and positions will result in our inability to say anything, or at best very little, about what had just transpired. If performers are exclusively focused on nonverbalizable aspects of their skills, they might be left without any declarative memory of what they had done even if the action had been effortful. On the standard account, post-performance amnesia results when experts’ conscious minds are not engaged during performance. By contrast, according to the two alternative explanations I’ve tendered, post-performance amnesia occurs when the nature of the performers’ conscious experience leaves them unable to recall what they had done (cf. Bermúdez 2017). If either of these explanations are correct, experts may be present in action without having any recollection of what they had just done.

4 Agency in flow David Velleman (2008) suggests that we achieve excellence only when we are “transcending reflective agency” (p. 182). Although reflective agency—that is, thinking about and deliberating over our actions as they unfold—is a stepping-stone to developing expertise, we 143

Barbara Gail Montero

perform at our best when we attain “self-forgetful spontaneity” or “flow” (p. 187). Expressing a version of the effortless expertise hypothesis, he tells us that in highly skilled actions, “the capacity to monitor … performance, to consider how it falls short of an ideal, and to correct it accordingly … is no longer exercised” (p. 188). Rather, after the requisite training, according to Velleman, “evaluative judgment is suspended,” and experts act “without deliberate intention or effort” (p. 185). Velleman leans on the research by psychologist Mihaly Csikszentmihalyi into when people feel happiest, which occurs when they are in what Csikszentmihalyi calls “flow.” In flow “there is no need to reflect, because the action carries us forward as if by magic” (Velleman 2008, p. 186, quoting Csikszentmihalyi 1990, p. 54). As being propelled forward as if by magic seems rather effortless, Velleman understands the phenomenon of flow as illustrating how effort dissolves expert action. And with the dissolution of effort comes the disappearance of agency. “One of the most universal and distinctive features of optimal experience,” Csikszentmihalyi tells us, is that “people become so involved in what they are doing that the activity becomes spontaneous, almost automatic; they stop being aware of themselves as separate from the actions they are performing;” and what this means is that in flow there is a loss of consciousness of the self (Velleman 2008, p. 186, quoting Csikszentmihalyi 1990, p. 53). Csikszentmihalyi, as Velleman interprets him, opens the door to a way in which expertise may be manifested effortlessly (as if by magic) and agentlessly (without a consciousness of the self ). But does Csikszentmihalyi’s research based on the “experience sampling method,” which prompts subjects at random times to write down what they are doing and thinking about and then rate their state of contentment, show that expert action is generally effortless and agentless? As Csikszentmihalyi (1990) makes clear, he is not investigating peak performance, but rather peak experience. And doing your best and feeling your best are different. It feels great to take a leisurely swim across the lake, but when you’re swimming in a race and want to win, you’ll put up with some suffering. Furthermore, the “magic” that characterizes flow does not necessarily result in effortless and agentless action since Csikszentmihalyi (2014) also seems to think, “intense concentration [is], perhaps the defining quality of flow” (p. 92). And according to Jackson and Csikszentmihalyi (1999), “without self-awareness an athlete misses important cues that can lead to a positive change in performance” (p. 105). Thus, Csikszentmihalyi’s research does not support the hypothesis of effortless expertise.

5 Act without acting? Taoism, in its two primary texts, the Laozi and the Zhuangzi, is also understood by some as advocating a way of acting that is unfettered by effort and devoid of agency. According to Edward Slingerland (2007, 2014), these texts describe how, at the pinnacle of skill, an expert is guided by the wu-wei, which Slingerland understand as the injunction to act yet avoid action.5 How to interpret these texts is controversial, and, because of its paradoxical nature, understanding the wu-wei’s injunction is particularly fraught with difficulty. Nonetheless, Velleman (2008), building on Slingerland (2007), finds in them an argument for “acting without deliberate intention or effort,” and so marshals them to further support his view that in performing optimally, one need not “keep one’s eye on an ultimate goal, or … follow the precepts of a method, or even … focus on one’s actions themselves” (p. 184). Rather, according to Velleman, we learn in the Zhuangzi that for the accomplished artisan, actions are not guided by the self but they just happen. Although he sees value in thoughtful training that involves self-reflection, he sees the Zhuangzi as offering further support for the idea that, at its best, expert action leaves the agent behind. Experts, as he sees it, “have acquired their 144

Expert agency

skills through training that involved self-scrutiny, self-criticism, and self-correction … [but this capacity to reflect] is no longer exercised after they have perfected their skill” (p. 188).6 For support, Velleman quotes from the Zhuangzi’s account of the skill of butcher, who, after years of practice, encounters oxen no longer with his eyes, but with his spirit: “sensible knowledge stops and spiritual desires proceed” (Ivanhoe and van Norden 2005, p. 225, as quoted by Velleman 2008, p. 183). The butcher’s actions are not guided by the butcher, but by “what is inherently so” (Ivanhoe and van Norden 2005, p. 225). Yet what does this mean and how accurate is it as a description of expertise? It is an open question whether the stories one finds in the Zhuangzi about individuals who perform their skill entirely effortlessly provide a recipe for individual action rather than a metaphor for political rule (specifically, that leaders ought not to govern by force). But more relevant to my hypothesis is whether butchers who want to butcher well do act effortlessly. In answering this question, research into the practice of butchery is needed.

6 My final effort It is unknown whether the Zhuangzi quotes an actual butcher, but it is perhaps telling that its authors did not rely on their own skills to make their point. I am often tempted to look at the actions of others and think: it’s so easy for them. Yet despite the persuasiveness of these appearances, I also have come to realize that sometimes, at least, this ease is an illusion. Perhaps nowhere is the illusion more powerful than in ballet. Slingerland contends that effortlessness is “unfakeable” (2014, p. 190). Yet in my prior career as a ballet dancer, I was paid to fake effortlessness. There are situations when experts feel that, suddenly, things have gotten easier, such as when marathon runners “break though the wall.” This break happens, it is thought, when one’s fuel source switches from glucose to ketones, which are a more efficient source of energy (Prince et al. 2013). However, when you are highly motivated to reach that finish line as fast as you can, once switched, you are not going to relax and let the ketones take over. Rather, after perhaps a moment to feel no longer on the brink of death, you’ll push yourself to the point at which any further expenditure would be detrimental to performance. Rather than discussing the phenomenology of butchers, I would have loved to hear Velleman focus on the phenomenology of writing philosophy. I cannot know what it is like for him to write—from my own perspective it always seems easy for others. However, rarely when I am writing philosophy does thinking stop and spiritual desire proceed; rather, self-scrutiny, self-criticism, and self-correction—what Velleman claimed was missing in expertise—are frequently making my life unpleasant. In fact, far from expertise being inversely proportionate to effort, in writing skill and effort increase proportionately. In their 1987 study of children’s writing skills, Carl Bereiter and Marlene Scardamalia found that “the paragons of effortless performance were fi fth-graders who, given a simple topic, would start writing in seconds and would produce copy as fast as their little fingers could move the pencil.” Increasing difficulty also characterized my experience of ballet. Because my standards for what counted as excellence and my ability to evaluate my technique develops faster than my dancing ability, it became increasingly difficult to meet my self-imposed demands. Degas’s insight that “painting is easy when you don’t know how; but very difficult when you do,” applied perfectly well to my case (in Bammell 1961). Expertise is effortful and agentful, which can at times turn what was enjoyable into a nightmarish trial. But this trial does have a saving grace, as sometimes the struggle makes the prize all the more rewarding. 145

Barbara Gail Montero

Further topics Agency, powers, and skills; Agency and mistakes; Agency and disability.

Notes 1 He might also mean that an expert uses energy efficiently, not that she uses less of it. 2 According to Bermúdez (2010), there is no sensory component to agency. I, however, understand sensory component of agency educes to that of effort, trying, conscious control, will power, and so forth. 3 That said, the highly automatic act of breathing, though it changes with conscious attention, seems to benefit from conscious attention. Conscious attention is not the same as making an effort. One can consciously attend to the movements of a limb that one doesn’t control or to a goal while reclining on the coach and not trying to achieve it. However, not only does it seem to be a necessary condition for making an effort, but it also seems that the mere attending to an action one does bring with it a minimal level of trying. I leave it to readers to test this for themselves. 4 For example, although all professional-level ice skaters practice an enormous number of hours every day, Deakin and Cobley (2003) have found that the best of them challenge themselves more by spending more time on jumps that they have not mastered, while the others spend more time on jumps that they have already mastered. 5 Onuma et al. (1993), see the goal in Japanese archery as “not the elimination of thought … [but rather] the elimination of the remnants of thought: that which remains when thought is divorced from action” (p. 22). 6 For a criticism of Slingerland see Fraser (2007), where he questions whether the passages about skill in the Zhuangzi are intended to express wu-wei.

Recommended reading * Christensen, W., Bicknell, K., McIlwain, D., and Sutton, J. (2015). “The sense of agency and its role in strategic control for expert mountain bikers.” Psychology of Consciousness 2(3): 340–353. An exploration of agency in expert mountain biking that resonates strongly with my view that expert agency is often guided by the conscious mind. * Høffding, S. and Montero, B. G. (2019) “Not being there: Reconciling expertise induced amnesia and the possibility of total recall,” Mind and Language 35(4): 621– 640 An article that lays out the idea of post-performance amnesia and presents a fuller explanation of how post-performance amnesia can be compatible with conscious attention in action. * Montero (2018) Thought in Action: Expertise and the Conscious Mind, Oxford: Oxford University Press, particularly chapter 9, “The Pleasure of Movement and the Awareness of the Self.” Here, I discuss why I think it is a mistake to interpret Csikszentmihalyi’s concept of flow as implicating a loss of self. * Slingerland, Edward (2007). Effortless Action: Wu-wei as Conceptual Metaphor and Spiritual Ideal in Early China (New York: Oxford University Press). Slingerland’s presentation of his conception of the wu-wei as effortless and agentless, a view which contrasts with the view I advocate. * Velleman, J. David (2008). “The way of the wanton,” in C. Mackenzie and K. Atkins (eds), Practical Identity and Narrative Agency (New York: Routledge). Velleman’s application of Slingerland’s and, as he understands them, Csikszentmihalyi’s ideas to action theory.

References Armstrong, D. M. (1981) “The nature of mind” in his The Nature of Mind and Other Essays. Ithaca, NY: Cornell University Press. Baddeley, A. (2014). Essentials of Human Memory (Classic Edition) (London: Psychology Press). Bammell, Ives (1961) The Shop-talk of Edgar Degas Paperback (Boston: University Press).

146

Expert agency Bargh, J. and Chartrand, T. (1999). ‘The unbearable automaticity of being,’ American Psychologist 54: 462–479. Beilock, S. and Carr, T. H. (2001). “On the fragility of skilled performance: What governs choking under pressure?” Journal of Experimental Psychology: General 130(4): 701–25. Beilock, S. L., Carr, T. H., MacMahon, C., and Starkes, J. L. (2002). “When paying attention becomes counterproductive: Impact of divided versus skill-focused attention on novice and experiences performance of sensorimotor skills,” Journal of Experimental Psychology: Applied 8: 6–16. Beilock, S. L., Wierenga, S. A., and Carr, T. H. (2003). “Memory and expertise: What do experienced athletes remember?” in J. Starkes and K. A. Ericsson (eds.), Expert Performance in Sports: Advances in Research on Sport Expertise (Champaign, IL: Human Kinetics): 295–320. Bermúdez, J. (2010). Action and awareness of agency: Comments on Chris Frith. Pragmatics & Cognition, 18: 576–588. Bermúdez, J. P. (2017). “Do we reflect while performing skillful actions? Automaticity, control, and the perils of distraction,” Philosophical Psychology 30(7): 896–924, DOI: 10.1080/09515089.2017.1325457 Csikszentmihalyi, Mihaly (1990). Flow: The Psychology of Optimal Experience (New York: Harper and Row). Csikszentmihalyi, Mihaly (2014). “The concept of flow,” in M. Csikszentmihalyi (ed.), Flow and the Foundations of Positive Psychology (Dordrecht: Springer): 239–263. Deakin, Janice M. and Cobley, Stephen (2003). “A search for deliberate practice: An examination of the practice environments in figure skating and volleyball,” in J. Starkes and K. A. Ericsson (eds), Expert Performance in Sports: Advances in Research on Sport Expertise (Champaign, IL: Human Kinetics): 90–113. Dewar, M., Garcia, Y. F., Cowan, N. & Della Sala, S. (2009). “Delaying interference enhances memory consolidation in amnesic patients.” Neuropsychology 23(5): 627–634. Dreyfus H. L. and Dreyfus, S. E. (1986). Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer (New York: Free Press). Dugdale, J. R. and Eklund, R. C. (2003). “Ironic processing and static balance performance in high-expertise performers,” Research Quarterly for Exercise and Sport 74 (3): 348. Ericsson, K. A., Krampe, R. T., and Tesch-Romer, C. (1993). “The role of deliberate practice in the acquisition of expert performance,” Psychological Review 100(3): 363–406. Ericsson, K. A. (2008). “Deliberate practice and acquisition of expert performance: A general overview,” Academic Emergency Medicine 15: 988–94. Fraser, Chris (2007). “Review: On Wu-Wei as a unifying metaphor,” Philosophy East and West 57 (1): 97–106. Guthrie, E. R. (1952). The Psychology of Learning (New York: Harper and Row). Høffding, S. (2018). A Phenomenology of Musical Absorption (Hampshire: Palgrave Macmillan). Høffding, S. and Montero, B.G. (2020) “Not being there: An analysis of expertise-induced amnesia,” Mind & Language 35 (5): 621–640. Ivanhoe, P. J. and van Norden, Bryan W. (2005). Readings in Classical Chinese Philosophy (Indianapolis, IN: Hackett). Jackson, S. A. and Csikszentmihalyi, M. (1999). Flow in Sports: The Keys to Optimal Experiences and Performances (Champaign, IL: Human Kinetics). Meinertz, A. (2008). Erik Bruhn: Billedet indeni. København: Schønberg. Millican, P. and Wooldridge, M. (2014). “Them and us: Autonomous agents in vivo and in silico,” in A. Baltag and S. Smets (eds), Johan van Benthem on Logic and Information Dynamics (Outstanding Contributions to Logic Volume 5) (Berlin: Springer-Verlag). Montero, Barbara Gail (2016). “Aesthetic Effortlessness,” in Sherri Irvin (ed.), Body Aesthetics (Oxford: Oxford University Press): 180–192. Montero, B. G. (2018). “The elusive concept of Expertise: Who counts as an expert?” Archives of Psychology 2(5): 1–16. Pilzecker, A., Müller, G. E. (1900). Experimentelle beiträge zur lehre vom gedächtniss. Germany: J.A. Barth. Onuma, Hideharu, DeProspero, Dan, and DeProspero, Jackie (1993). Kyudo: The Essence and Practice of Japanese Archery (New York: Kodansha America). Pacherie, Elisabeth (2007). “The sense of control and the sense of agency,” Psyche 13 (1): 1–30. Prince, Allison, Zhang, Yifan, Croniger, Colleen, and Puchowicz, Michelle (2013). “Oxidative metabolism: Glucose versus ketones,” Advances in Experimental Medicine and Biology 789: 323–328.

147

Barbara Gail Montero Sherman, Robert (1980). “Natalia Markarova,” Great Artists Series (excerpted from a 1977 interview) (WNYC archives id: 70039). Shusterman, Richard (2008). Body Consciousness: A Philosophy of Mindfulness and Somaesthetics (New York: Cambridge University Press). Slingerland, Edward (2007). Effortless Action: Wu-wei as Conceptual Metaphor and Spiritual Ideal in Early China (New York: Oxford University Press). Slingerland, Edward (2014). Trying Not to Try: The Art and Science of Spontaneity (New York: Crown Publishers). Velleman, J. David (2008). “The way of the wanton,” in C. Mackenzie and K. Atkins (eds), Practical Identity and Narrative Agency (New York: Routledge). Wixted JT. (2004). “The psychology and neuroscience of forgetting.” Annu Rev Psychol 55: 235–269. Wulf, G. and Lewthwaite, R. (2010). “Effortless motor learning? An external focus of attention enhances movement effectiveness and efficiency,” in B. Bruya (ed.), Effortless Attention: A New Perspective in Attention and Action (Cambridge, MA: MIT Press): 75–101.

148

13 AGENCY AND MISTAKES Santiago Amaya

Mistakes are part of our life as agents. Although we tend to be shy about them, preferring sometimes not to talk about them, everyone recognizes having made a mistake at some point or another. Philosophers are no different in this regard. In theorizing about agency, we tend to emphasize the successful cases. Even when we recognize our imperfections, we mostly focus on those that only derivatively show up in what we do. Mistakes, however, are more than a pervasive feature of our lives. Even if one worries about making them, there is still something positive about them. Not only, as the saying goes, can we learn a lot from our mistakes; we can also learn a lot about what kind of agents we are by thinking about the mistakes we make. There is, in fact, as we shall see, reason to think that mistakes are built into the capacities that make us agents. In this chapter, I discuss the relationship between agency and mistakes. I begin with some clarifications that show why this relationship matters (Sections 1 and 2). Then, I offer a quick survey of the mistakes often discussed by philosophers of action, noting the tendency to focus on those explained by errors in some of our attitudes (Section 3). With this, I discuss the category of performance mistakes and some strategies that have been deployed to downplay their importance (Sections 4 and 5). I conclude with some brief remarks about the sources of our fallibility as agents (Section 6).

1 Standards of agency Human creatures have a distinct set of abilities in virtue of which we qualify as agents. We can, for instance, form desires and intentions. We can normally figure out which courses of action are better to satisfy those desires and are consistent with our intentions. And we can settle on those courses of action or renounce from embarking on others, as the case might be. Philosophical accounts of agency emphasize different abilities in identifying what is constitutive of human agency. Humans, of course, are not the only agents out there. The animal kingdom, as some have noted (see, for instance, Andrews, 2013; Burge, 2009), abounds with instances of nonhuman agency: creatures coordinating their behaviors to achieve goals in ways that are attributable to them. For present purposes, however, we shall restrict ourselves to what might be

DOI: 10.4324/9780429202131-17149

Santiago Amaya

described as distinctively human agency. That is, we shall focus on instances of agency that require the exercise of abilities that seem exclusively within the purview of humans. Even then, there is room for disagreement. For many theorists, interest in human agency reflects foremost an interest on intentional actions: roughly, the things that you do in the execution of your intentions (Brand, 1984; Mele, 1992). Others have thought that our capacity to form intentions and plan ahead of time has a certain priority here, so their focus has mostly been on our capacities for planning agency (Bratman, 1987). Finally, some take their interest in agency as derivative on questions related to autonomy and freedom. Hence, they have argued that in theorizing about human agency, the spot should not shine equally on all intentional actions but, instead, on full-blooded instances of them (Velleman, 2000). Being an agent, however, requires more than the possession of these abilities, regardless of which of them one’s preferred theory identifies as central. It is a matter as well of shaping what we do by virtue of their exercise. This last bit is important. Desiring good things or having true beliefs about how to achieve them, forming coherent plans, and making decisions that accord to them are certainly remarkable things. But, unless you can translate those attitudes into behavior and structure what you do accordingly, you won’t be an agent. You might, at best, be a good thinker. Thus, a widely agreed feature of actions is that, by virtue of being brought about by agents, they are open to certain forms of evaluation. This is true whether we speak about minimal or distinctive human agency, of intentional planning, or of full-blooded agency (for an interesting discussion, see Shepherd, 2021, ch. 6). In brief, being capable of behaving in certain ways creates reasonable expectations that one would behave in those ways, at least in some circumstances. In doing so, it gives rise to standards by which one’s doings can be evaluated. Naturally, different standards matter for different kinds of agents. In the case of minimal forms of agency, for instance, actions are typically evaluated in terms of survival, reproductive success, and the like. With respect to our actions, standards of evaluation vary widely with circumstances and interests. Although some are specific to a domain, a few seem to be domain-general. In principle, any of one’s actions can be judged in terms of how prudent, moral, or rational it is.

2 Which standards? In speaking of standards of evaluation, we should be careful to distinguish two different notions. By not doing it we might end up with a picture that has a hard time accommodating some of our mistakes (for one crisp way of bringing up these difficulties, see Lavin, 2004, sect 8) There are, first, what have been called constitutive standards. These are standards that any behavior would have to meet to be considered an exercise of our agency. In addition, there are what we might call standards for success: those that exercises of agency need to meet to be considered successful, that is, to be successful qua actions. Most theorists agree that the constitutive standards of human agency are tied to certain expectations of rationality, which is why human agency is often glossed as rational agency. In short, any behavior that qualifies as a human action ought to have some sort of rational structure. For different kind of theorists, constitution will mean different things: being caused in certain ways, being responsive to certain norms, having some distinctive kind of explanation, and so forth. And rationality will be tied to different set of norms. But, ultimately, the underlying thought is that human actions are somewhat rational responses (not necessarily the most rational responses) to the circumstances we occupy. 150

Agency and mistakes

Undoubtedly, more variation can be found with respect to the expectations by which actions get evaluated for success. Among other things, what counts as successful varies from one domain to another. But, arguably, a minimum success standard for intentional action, and for more sophisticated forms of agency that depend on it, is that it satisfies the intentions that give rise to it (see Amaya, 2017; Radoilska, 2013). A successful action, on this view, is one whether the agent achieves what she originally intended with it, given the attitudes that the agent had at the time. Now, constitutive and success standards, even those that are domain general, can evidently come apart. Cases of causal deviance, such as those illustrated by Donald Davidson’s (1978) famous climber, are a good example of how success standards could be met, except that the conduct in question fails to meet constitutive standards for agency. In such cases the behavior matches the intentions of the agent, so in that respect what the agent does would seem a success. But to the extent that it is not the product of the rationality of the agent, it falls short of exhibiting the structure that, according to many theorists, is required of genuine action. Mistakes show how these standards can come apart, but here the dissociation goes in the opposite direction. At least, this is true of the kind of mistakes that matter in thinking about our agency. When we make a mistake, what we do can be described as meeting the constitutive standards of agency, say, by exhibiting the rationality characteristic of actions. But it fails to meet some reasonable standard of success. Thus, although at the time we behave in ways that are somewhat reasonable, at least reasonable enough to be considered actions, what we do falls below some reasonable standard. At the very minimum, we fail to act as intended. Consider, in this regard, two common ways of rejecting imputations of a mistake. First, one can disavow the imputation by questioning whether one’s actions actually violated a reasonable standard of success. Expressions such as “I didn’t know,” “It was too hard for me to do it,” or even “I did mean to do that” often mark this sort of response. Alternatively, instead of questioning whether a mistake was made, one can reject the imputation by questioning the status of what one did as an exercise of one’s agency. “It was an accident,” “I could not help it,” “I was just bad luck,” are expressions typically used to indicate this second way of distancing.1 We can now say more precisely why our mistakes are a window into the kind of agents that we are. By dissociating constitutive and success standards, reflection on our mistakes (which are they and why we make them) require us to think what are the conditions that exercises of our agency need to meet to be such, as opposed to those that we simply find desirable that our actions meet. Doing so requires not just that we specify clearly what those standards are, but also that we ask ourselves how much imperfection our conception of agency can tolerate. Ultimately, the size of the gap between constitutive and success standards, especially those that are domain general, will reveal how accepting we truly are of our mistakes.

3 Which mistakes? Some philosophers might not worry whether their accounts of agency are actually instantiated by imperfect human beings (see, for instance, Velleman, 2004; Ford 2011). Idealization is somewhat widespread in this corner of philosophy. But many of them aim at providing theories that seek to recognize our human fallibility. So, even if explicit theorization of mistakes is rare, mention of them tends to show up in discussions about human agency. 151

Santiago Amaya

Yet, it is important to observe which kinds of mistakes normally get recognized and what role they are typically called to play in the construction of these accounts. Often the mistakes discussed can be traced to problematic features of the attitudes behind our actions, the goals we desire, how we think they can be achieved, and so forth, as opposed to being non-derivative mistakes of our agency. As a consequence, most theorizing in this area tends to overlook the extent of our imperfections as agents.2 It is easy to recognize this trend once one’s attention is drawn to it. In his seminal essay on Agency, for example, Davidson (1971) sought to identify what sets our mistakes apart from unintended things that merely happen to us. Among the examples discussed, he mentions the man who unintentionally spills his coffee, the agent who boards the wrong plane, the prince who stabs an innocent man, and similar cases. In all of them, the source of the mistake is the same: the person thought the cup had tea (not coffee), he believed the plane was heading to London, England (not London, Ontario), and Hamlet didn’t know Polonius was behind the tapestry. As Davidson knew well, mistakes originating in our conative, as opposed to our doxastic, attitudes are also possible and part of our life. These have also received a fair share of philosophical attention. In their groundbreaking discussions of autonomous agency, for instance, Harry Frankfurt (1971) and Gary Watson (1975) brought to light cases of agents alienated from their desires, asking to what extent their actions could be considered free and attributable to them. Less dramatic, but equally problematic perhaps, are David Gauthier’s (1986) examples of agents that form preferences lacking any real information about the options presented to them. In addition, philosophers have also discussed cases of attitudes that do not fit well with each other. For this lack of alignment might also be conducive to making mistakes. Discussion about weakness of will is good example here. As Amelie Rorty (1980) noticed, episodes of weakness of will can involve a multiplicity of different conflicts. They could happen, as Alfred Mele (1988) has argued, when we form intentions that run contrary to the agent’s considered judgments. Or, as Richard Holton (1999) described them, the conflicts can be among our current intentions and resolutions to overcome temptation that we made in the past. 3 It is worth noting the pattern behind these well-known discussions. Although the examples brought up there involve recognizable mistakes, they are ones that only derivatively we make as agents. In them, agents fail to succeed in some significant respects, say, failing to obtain what they “really” want or what they would have intended had they controlled themselves. But they fail at it because they antecedently have problematic beliefs, they desire things that they do not really want or they think they shouldn’t want, or because the relevant attitudes behind their actions are not properly aligned with each other. The standards by which these actions are mistakes are the standards of correctness for those attitudes. Unfortunately, this leaves too many of our imperfections out. As we mentioned at the outset, human agency is not exercised simply by having attitudes of a certain sort (say, those that track true and good things) or that bear certain kind of relations (of consistency, for instance) with each other. In order to be considered agents, we also need to effectively translate those attitudes into action. And here, in the process of making these transitions, the possibility of making a variety of mistakes of a different sort opens up. Despite having attitudes that are not, in principle, questionable, we can end up doing things that fall short of some reasonable expectations of success. Let me make the overall pattern more perspicuous by distinguishing two kinds of mistakes. There are, first, mistakes that can be traced to the attitudes that lie behind the actions 152

Agency and mistakes

that we undertake. In these cases, the attitudes themselves fall below certain standards applicable to them or they instantiate patterns that otherwise violate some standard of correctness. These are the mistakes that mostly figure in the philosophical literature. But there are, in addition to them, performance mistakes. These are ones that we make, not as believers, desirers, and so forth, but as creatures who can translate their beliefs, desires, and so forth into actions. In so far as these do not necessarily impugn our attitudes, they are mistakes as judged by standards of success independent of whichever measure of correctness we use to evaluate our attitudes. Because of this, they illustrate how our fallibility extends beyond our problematic attitudes.

4 Performance mistakes To get a sense of how this happens, consider the following quotation from Michael Bratman: Suppose that this morning I formed the intention to have a milkshake at lunch, lunchtime arrives, my intention remains, and nothing unexpected happens. In such a case, I do not normally need to tote up the pros and cons concerning milkshake drinking. Rather, in the normal course of events I will simply proceed to execute (or, anyway, try to execute) my intention and order a milkshake. My intention will not merely influence my conduct, it will control it. (1987, p. 16) The passage is meant to introduce Bratman’s influential idea that intentions involve a characteristic commitment, something that sets them apart from other attitudes. But what he says in it is literally false. Intentions, we can agree, control conduct in ways in which, say, desires don’t. They involve a commitment to act when time comes. Yet, living up this commitment requires more than forming an intention and waiting for that time to arrive. Imagine some mundane ways in which Bratman’s story might unfold instead. You decide to have a milkshake, the intention remains, and so forth. But, as soon as you get to lunch, you see in the menu a combo that includes a soda and that looks like a good deal. So, you go ahead and order it. It is not as though you changed your mind and desist from having the milkshake. Simply, you ordered something else without even remembering your prior intention. Alternatively, at noon you leave the office to get lunch, while still thinking about the paper you were writing. But out of habit you end up walking to the place with the salad bar where you normally go, instead of walking to burger place where you had earlier planned to go. Only after you finish your salad and the mineral water you order with it, you realized you never had the milkshake. These stories should sound familiar, as they illustrate a common kind of performance mistake: action slips. Each, however, instantiates a slightly different pattern (Norman, 1981; Reason, 1984). Sometimes, as in the first story, you form an intention to do something later and plan accordingly, but the thought of your early plans simply does not come to mind when the time is right. So, you end up doing something that would otherwise make sense to do then (Amaya, 2013). Or, as in the second story, being engrossed in some mental task, a habitual routine kicks in and you end up doing something that predictably would not allow you to do what you intended first (Amaya, 2020). Slips, to be sure, tend to result in minor inconveniences (although some slips are genuinely catastrophic, see Amaya and Doris 2014). And normally they do not occur because of some problematic attitude that you need to change. Certainly, it was not wrong for you to 153

Santiago Amaya

find the lunch combo appealing; the habit of going to the salad place is definitely a good one to have. In any other occasion, acting on that desire or following that habit would have been an acceptable thing to do. However, on this occasion, by acting on them, you ended up acting contrary to the intentions and plans that you had. By doing so, you ended up making a mistake. The point can be put in more general terms. Being an agent, let’s agree, requires the ability to form intentions and plans for executing them. But having that ability, even exercising it, only makes you a good planner. This is what Bratman fails to note in the passage above. To be an agent, you also have to exercise the power of bringing to mind those intentions and plans in the midst of doing other things and with potential distractors in sight. And here, in exercising these powers, mistakes can occur. If you fail to keep track of where you are with respect to your overall plans, whatever you intended to do might never materialize. Obviously, slips are only one among many possible illustrations here. Life is full of episodes of forgetfulness, distractions, and the like, to be all fitted in one and the same category. So, in addition to slipping, performance mistakes can happen in all sorts of ways: in the ways in which our desires and judgments connect with our decision-making processes, in the way in which what we know informs how we deliberate, and so forth. When these disconnects happen our actions will fail to reflect the relevant attitudes. So, even if they have all the right desires, beliefs, and so forth, and even if these might be rightly aligned with each other, they might fail to make a difference in what we do.

5 Attitude-based explanations The term “performance mistake” probably resonates with G.E.M. Anscombe’s followers. In a well-known passage of Intention, she marks the distinction drawn above (at least, a version of it) by contrasting mistakes in judgments and those that occur in performance. Alas, Anscombe did not spend much time explaining the contrast, nor did she offer much by way of discussion of the latter sort of mistakes. Her attitude, instead, was to casually dismiss them as the “rare exception” (Anscombe, 1963, par. 32). Other philosophers may have a similar attitude: they recognize that performance mistakes can happen, but they treat them as oddities in the life of agents. This actually might explain Bratman’s oversight above or, at least, why he hedged at various points of the passage using the word “normally.” The truth, however, is that these mistakes are more common than one might initially think, definitely at least as common, say, as the episodes of alienation or weakness of will over which so much ink has been spilled. If anything, judging by the recurrence of first-person reports, episodes of misremembering, getting distracted, and the like should not count as unexpected ( Jónsdóttir et al. 2007; Reason and Mycielska, 1982). Interestingly, even some who have recognized the pervasiveness of performance mistakes have fallen short of pressing the idea further. Their reaction, instead, has been mostly to try to explain them away, either by reducing our performance mistakes to prior errors in our attitudes or by placing them altogether outside the domain of our agency. If idealization is sometimes widespread among philosophers of action, perfectionism is rampant. The prototypical actions that our models of agency are meant to explain, almost as a matter of rule, are successful actions. To see how these attitude-based explanations work, let us return to slips. Without a doubt, the grandfather here is Sigmund Freud, who was perhaps the first to recognize the pervasiveness and psychological significance of them (1901, 1915). In his early writing, he complained about what he referred to as “psycho-physiological” explanations of the slip, 154

Agency and mistakes

in brief, explanations of them as glitches in the cerebral machinery behind our conduct. Instead, for him, slips were a window into the deeper mechanisms of the human mind. In particular, they were mistakes that resulted from the fact that people regularly held in the minds emotionally threatening motives. Nowadays, the influence of Freud and his psychodynamic approach is less gripping than it used to be (although some of it survives in the writing of some philosophers; see, for instance, Arpaly, 2003, pp. 159–162, and Velleman, 2000, intro). But this has often meant, not a retreat from attitude-based explanations but a refashioning of them in the guise of new psychological models of the human mind.4 Just to take an example, Roy Sorensen (2011) argues that popular dual-processing models in cognitive psychology underwrite a distinction between deep and shallow attitudes. According to him, slips are actions caused and rationalized by shallow beliefs and desires that happen to be out of line with our more reflective attitudes. We cannot go here into a detailed criticism of these approaches. But it is worth noting how these explanations often look very much like ad hoc reconstructions. Although the attitudes invoked to explain the slips might make sense of the behavior of the person at the time, they often fail to make sense of the overall behavior of the person. Normally, there are simpler and more comprehensive available explanations around the corner, explanations that are more transparent of the psychology behind the mistake. To illustrate, return one last time to the alternative to Bratman’s story. Let’s accept that it is possible that you developed a sudden desire to have salad. And, if that were the case, that would explain why you ended up going to the salad bar instead of the burger place. But it’s more plausible to say that what got you there was the habit of going for lunch to that place almost every day. That was, in fact, a habit that you had—you often went there for lunch there. It accords with your state of mind at the time—habits tend to capture behavior when one’s mind is somewhere else. And it explains why what you end up doing that day, despite being a mistake, was not something out of the ordinary. Unsurprisingly, this kind of criticism is one that detractors of Freud have repeatedly pointed out (see especially Grünbaum, 1984). The evidential basis for attributing agents the specific repressed motives that psychodynamic explanations involve is often thin. Except for the isolated episode of the incident, they do not show up anywhere else in the life of the person. If anything, what seems to drive Freud’s explanations is almost entirely his prior commitment to the theory. In the end, it is possible that something similar happens here: what drives attitude-based approaches to our performance mistakes are other prior philosophical commitments, plausibly a commitment that all explanations of action need to have a common form (for two very different expressions of this commitment, see Smith, 1998; Thompson, 2008). Certainly, explaining how performance mistakes happen and why they still count as exercises of our agency requires accepting a more pluralist approach to the standard of rationality that are constitutive of our actions (for a defense of this pluralism, see Amaya, 2020). It requires, for instance, understanding that habitual actions might meet the constitutive requirements of rational actions in ways that differ from actions that are rationalized by beliefs and desires.

6 Facing imperfection We began with the idea that, by dissociating constitutive and success standards, reflection on our mistakes is a window into the nature of human agency. As we have seen, most philosophical reflection has been concerned with mistakes that derive from errors in the formation of our attitudes or on how these attitudes fit together. So, in the end, most of it has 155

Santiago Amaya

focused only on conducts that qualify as mistakes in the sense that the attitudes behind them fall short of correctness standards for them. It is, no doubt, important to have an appreciation of this kind of mistakes. But concentrating on them runs the risk of painting an inaccurate picture of ourselves, one that locates that source of our imperfections as agents in the wrong place. Focusing on performance mistakes, by contrast, paints a different picture. To the extent we not only intend, plan, and decide what to do, but that it is also upon us to translate our intentions, plans, and decisions into conduct, we do not make all of our mistakes as believers, deciders, and so forth. As revealed by our performance mistakes, we also have limitations having to do directly, not derivatively, with our agency. If true, these considerations recommend that we endorse a kind of fallibilism with respect to our agency. According to it, the possibility of making mistakes is built into whichever powers make us agents. Ultimately the thought here is that the gap between the constitutive standards of our agency and our standards of success is not entirely accounted for by the limitations we have in other domains of our rational lives.

Related topics Agency, powers, and skills; Expert agency; Agency and disability; Pathologies of agency; Agency and normativity

Notes 1 The distinction is similar, but not the same, as the one that Peter Strawson (1962) draws between excuses and exemptions. Exemptions (e.g., appealing to insanity) call into question the status of the person as a morally responsible agent altogether (Watson 1987; Wallace 1994). In the distinction drawn above, instead, the exclusion is supposed to be piecemeal: the person might generally be considered an agent but not with the respect to the behavior in question. Excuses, on the other hand, need not be full exculpations. “I did not mean to do that” might show that I didn’t act of ill will, but bad will is not necessary for blame attributions. For discussion, see Amaya and Doris (2014). 2 Exceptions to this generalized trend can be found in Mele (2006), Peabody (2005), and Frost (2018), but J.L. Austin’s (1957) “Plea for Excuses” is a predecessor worth mentioning too. 3 Not all treatments of weakness of will follow this general pattern. There are ways of thinking about akratic breaks, where these involve a person acting against her considered judgment or prior resolution but where her action is the product of a habit, not so much of a conflict among her preferences (see Silver 2019 and Mylopolous and Pacherie 2020). 4 By contrast, the psycho-physiological approach to slips that Freud denounced mostly faded away. Still, off-hand remarks echoing these approaches can be sporadically found in the literature. Back in the day, for instance, Hilary Putnam (1975, p. 372) compared our thinking processes to the operations of Turing machine; he then went on to compare verbal slips with accidents in the printing device of those machines. Likewise, when it was objected that slips were not explainable from the intentional stance, Daniel Dennett dispatched the objection declaring slips “either malfunctions or outcomes of misdesign” (1982, p. 63).

Further reading Constitutive standards of action have been mostly discussed in the context of constitutivist accounts of agency (see Agency and Normativity in this volume). But a general discussion of the relation between agency and standards of evaluation, which does not presuppose the theoretical apparatus of constitutivism, can be found here:

156

Agency and mistakes Shepherd, J. (2021). The Shape of Agency. In The Shape of Agency: Control, Action, Knowledge: 57–91. Oxford University Press. Two very different approaches to performance mistakes can be found in these papers. Amaya 2020 provides a framework for thinking about some of these mistakes as exercises of agency, whereas Sripada (2019) provides a model for thinking about them as glitches of subpersonal mechanisms. Amaya, S. (2020). Out of Habit. Synthese. Sripada, C. (2019). The Fallibility Paradox. Social Philosophy and Policy, 36(1): 234–248 Although not part of the discussion of the chapter, performance mistakes raise interesting questions about moral responsibility and blame. For differente perspective about cuplability for performance mistakes, see the following two papers: Amaya, S., and Doris, J. M. (2014). No Excuses: Performance Mistakes in Morality. In J. Clausen and N. Levy (Eds.), Springer Handbook of Neuroethics: 253–272. Netherlands: Dordrecht. Rudy-Hiller, F. (2019) Give People a Break: Slips and Moral Responsibility. The Philosophical Quarterly, 69(277): 721–740.

References Amaya, S. (2013). Slips. Noûs, 47(3), 559–576. https://doi.org/10.1142/9789814569057_0006 Amaya, S. (2017). Two Kinds of Intentions: A New Defense of the Simple View. Philosophical Studies, 175(7), 1767–1786. https://doi.org/10.1007/s11098-017-0934-1 Amaya, S. (2020). Out of Habit. Synthese. https://doi.org/10.1007/s11229-020-02780-3 Amaya, S., and Doris, J. M. (2014). No Excuses: Performance Mistakes in Morality. In Springer Handbook of Neuroethics, ed. J. Clausen and N. Levy, pp. 253–272. Netherlands: Dordrecht. Andrews, K. (2013). Ape Autonomy? Social Norms and Moral Agency in Other Species. In Animal Minds and Animal Ethics, ed. K. Petrus and M. Wild. Bielefeld: Transcript. Anscombe, G. E. M. (1963). Intention. Cambridge, MA: Harvard University Press. Arpaly, N. (2003). Unprincipled Virtue. New York: Oxford University Press. Austin, M. A. (1957). A Plea for Excuses. Proceedings of the Aristotelian Society, 57, 1–30. Brand, M. (1984). Intending and Acting: Toward a Naturalized Action Theory. Cambridge, MA: MIT Press. Bratman, M. E. (1987). Intentions, Plans and Practical Reasoning. Cambridge, MA: Harvard University Press. Burge, T. (2009). Primitive Agency and Natural Norms. Philosophy and Phenomenological Research, 70(2), 251–278. Davidson, D. (1971). Agency. In Agent, Action, and Reason, ed. R.Binkley, R. Bronaugh, and A. Marras, pp. 1–37, Toronto: University of Toronto Press. Davidson, D. (1978). Intending. In Philosophy of History and Action, pp. 41–60. Dordrecht: Springer. Dennett, D. C. (1982). Making Sense of Ourselves. In Mind, Brain, and Function, ed. J. I. Biro and R. W. Shahan, pp. 63–81. Norman: University of Oklahoma Press. Ford, A. (2011). Action and Generality. In Essays on Anscombe’s Intention, ed. J. Hornsby, F. Stoutland and A. Ford, pp. 76–104. Cambridge, MA: Harvard University Press. Frankfurt, H. (1971). Freedom of the Will and the Concept of a Person. The Journal of Philosophy, 68(1), 5–20. Freud, S. (1901). The Psychopathology of Everyday Life. London: Hogarth Press. Freud, S. (1915). Introductory Lectures on Psychoanalysis. London: Hogarth Press. Frost, K. (2018). Basic Mistakes in Performance. Proceedings of the XXIII World Congress of Philosophy 44, 17–21. Gauthier, D. (1986). Morals By Agreement. Oxford: Oxford University Press. Grünbaum, A. (1984). The Foundations of Psychoanalysis: A Philosophical Critique. Berkeley, CA: University of California Press. Holton, R. (1999). Intention and Weakness of Will. The Journal of Philosophy, 96(5), 241–262. Jónsdóttir, M., Adólfsdóttir, S., Cortez, R. D., Gunnarsdóttir, M., & Gústafsdóttir, Á. H. (2007). A Diary Study of Action Slips in Healthy Individuals. The Clinical Neuropsychologist, 21, 875–883. Lavin, D. (2004). Practical Reason and the Possibility or Error. Ethics, 114(3), 424–457. https://doi. org/10.1111/1467-9213.00296 Mele, A. R. (1988). Irrationality: A Précis. Philosophical Psychology, 1(2), 171–177. Mele, A. R. (1992). Springs of Action: Understanding Intentional Behaviour. Oxford University Press.

157

Santiago Amaya Mele, A. R. (2006). Practical Mistakes and Intentional Actions. American Philosophical Quarterly, 43(3), 249–260. Mylopoulos, M., and Pacherie, E. (2020). Self-Control as Hybrid Skill. In Surrounding Self-Control, ed. A. R. Mele, pp. 81–100. Oxford: Oxford University Press. Norman, D. A. (1981). Categorization of Action Slips. Psychological Review, 88(1), 1–15. Peabody, K. (2005). Trying Slips: Can Davidson and Hornsby Account for Mistakes and Slips? Philosophia, 33, 173–216. Putnam, H. (1975). Minds and Machines. Chap. 18 in Mind, Language and Reality. Philosophical Papers, pp. 362–385. Cambridge, MA: Cambridge University Press. Radoilska, L. (2013). Addiction and Weakness of Will: An Integrated Account. In Addiction and Weakness of Will, pp. 119–139. Oxford: Oxford University Press. Reason, J. (1984). Lapses of Attention in Everyday Life. In Varieties of Attention, ed. R. Parasuraman and D. R. Davies, pp. 515–549. New York: Academic Press. Reason, J., and Mycielska, K. (1982). Absent-Minded? The Psychology of Mental Lapses and Everyday Errors. Englewood Cliffs, NJ: Prentice-Hall. Rorty, A. O. (1980). Where Does the Akratic Break Take Place? Australasian Journal of Philosophy, 58(4), 333–346. https://doi.org/10.1080/00048408012341341 Shepherd, J. (2021). The Shape of Agency. In The Shape of Agency: Control, Action, Knowledge, pp. 57– 91. Oxford: Oxford University Press. Silver, K. (2019). Habitual Weakness. Thought: A Journal of Philosophy, 8(4), 270–277. Smith, M. (1998). The Possibility of Philosophy of Action. In Human Action, Deliberation and Causation, eds. Bransen J., Cuypers S.E., pp. 17–41. Dordrecht: Springer. Sorensen, R. (2011). What Lies behind Misspeaking? American Philosophical Quarterly, 48(4), 399–409. Strawson, P. (1962). Freedom and Resentment. Proceedings of the British Academy, 48, 187–211. Thompson, M. (2008). Life and Action. Canbrdige, MA:Harvard University Press. Velleman, D. (2000). Introduction. In The Possibility of Practical Reason, pp. 1–31. Oxford: Oxford: Oxford University Press. Velleman, D. (2004). Précis of The Possibility of Practical Reason. Philosophical Studies, 121, 225–238. Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, G. (1975). Free Agency. The Journal of Philosophy, 72(8), 205–220. Watson, G. 1987. Responsibility and the Limits of Evil: Variations on a Strawsonian Theme. In Responsibility, Character, and the Emotions: New Essays in Moral Psychology, ed. Ferdinand Schoeman, pp. 256–286. Cambridge: Cambridge University Press.

158

14 AGENCY AND DISABILITY Kevin Timpe

1 Introduction There are a number of philosophical issues at play in any stipulative definition of agency. Lots of entities have agency; lots of kinds of things are able to exercise a distinctive kind of causal control that originates, at least in part, within them rather than just passes through them. In this chapter I focus only on human agency. By ‘agency,’ I mean both the complete set of capacities and abilities that humans have that enable them to do characteristic human activities (e.g., my capacity to make a double-shot of espresso) and the exercise of those capacities (e.g., my having just made and drank a double-shot of espresso). Many of these capacities and their exercise will be shared between humans and other organisms, such as the capacities for locomotion, nutrition, and various cognitive tasks. Other capacities and their exercise will be uniquely human; the exact boundaries between those that are uniquely human and those we share with other organisms need not concern us at present. Agency is typically taken to involve a broad range of capacities and abilities: volition, intention, desire, sensation, emotions, proprioception, bodily control, and the abilities to evaluate reasons and guide one’s behavior on the basis of those reasons. The use of ‘behaviors’ here should be construed broadly, including not only bodily movements but also mental acts. Bodily movements and mental acts are both behaviors in the relevant since they will be ‘agentive when … sensitive to reasons, that is, able to adjust flexibly its means and goals to varying constraints or opportunities’ (Proust 2013, 209f ). For each of these capacities or sets of capacities, there will be disabilities that impact the agent’s having or ability to exercise those capacities. Here, I look at a number of ways that disabilities can impact agency. A central claim is that looking at how agency and disability relate can tell us something about human agency more broadly. In Section 2, I clarify my approach to the nature of disability. This will make it clear why I think we must engage with the existing literature on a range of disabilities to see how disability affects human agency. Section 3 explores a number of examples. In Section 4, I draw a number of lessons about agency and argue that they apply not just to agents with disabilities, but to human agency in general.

DOI: 10.4324/9780429202131-18159

Kevin Timpe

2 The nature and approach to disability In the same way that the exact meaning and extent of ‘agency’ is contentious, so too is the exact meaning and extent of ‘disability.’ Licia Carlson refers to disability as ‘the philosopher’s nightmare’ (Carlson 2010, 1), in part, though certainly not exclusively, because of the care needed to approach the topic well. History is full of examples of how thinking poorly about the nature of disability has led to significant harm against those with disabilities. Careful reflection on disability requires us to think about what disability is (or, better, what disabilities are). There are a number of different models of the nature of disability: the medical model; forms of the social model, both strong and weak (see Kahane and Savulescu 2009 for the claim that there is ‘no single agreed formulation of the social model account of disability’ (21)); critical realist models; moral models; the Nordic model; mixed models; the welfarist model; and others. In The Minority Body, Elizabeth Barnes cautions against assuming from the outset that a particular model can apply equally well across the breadth of disabilities. In her book, she is only focusing on physical disabilities across. Even physical disability, she argues, is sufficiently complicated that we should not begin in a ‘top-down’ approach, seeking a general account or model. Doing so runs the risk of privileging an account of what disability is that may not accurately reflect the experiences of those the full range of disabilities that the account is supposed to include. Instead, Barnes suggests that we should work in a ‘ground-up’ way that begins with paradigmatic instances of disabilities and works from there toward an account of what disability in general is.1 This will also be true not just for physical disability, but for disability in general. Many treatments of disability, both within philosophy and elsewhere, fail to take seriously the diversity of disability. Like Barnes, I think we should begin our philosophical reflection by first exploring in detail specific disabilities before we seek to treat those disabilities in a unified or overarching way (see Barnes 2016, 4; for a similar approach see Kahane and Savulescu 2009). This approach begins by pointing to particular paradigmatic cases. One has first to decide which purported disabilities are in fact paradigmatic. But then two problems arise. First, if one ultimately endorses a revisionist account, what one originally took to be paradigmatic cases may fail to be paradigmatic, or even fail to be a disability altogether. Second, there’s the question of how far the boundaries of the category ‘disability’ extend beyond those paradigmatic cases. It may be, as Barnes suggests, that there is no clear boundaries for whether or not a type of physical condition is a disability (Barnes 2016, 47). Fortunately, the conclusions I am aiming for do not require a full account of what all disability is, nor that we know exactly where to draw the boundaries around the concept. One final clarification: social models of disability are especially influential in disability studies. In this field, as Ronald Berger’s writes a discussion of definitional issues typically begins with a distinction between impairment and disability, whereby impairment refers to a biological or physiological condition that entails the loss of physical, sensory, or cognitive function, and disability refers to an inability to perform a personal or socially necessary task because of that impairment or the societal reaction to it. (Berger 2013, 6) This distinction between impairment and disability is then used to argue that while impairment is biological or physiological, disability is ‘something imposed on top of our 160

Agency and disability

impairments by the way we are unnecessarily isolated and excluded from full participation in society. Disabled people are therefore an oppressed group’ (UPIAS 1976, 4). The social model is extremely influential, both academically and politically. It is now codified, for instance, in the WHO’s International Classification of Impairment, Disability, and Handicap. But this approach simply pushes the question of ‘what is disability’ back into the question of ‘what is impairment’ (Barnes 2016, section 1.4.1). Similarly, Shakespeare (2014, 22) argues that disability, understood as social, and impairment, understood as biological, are ‘always intertwined.’ Thus, we cannot extricate one from the other in the way that the social model often does. For these reasons, I only address disability and not impairment.

3 Specific disabilities Given the ‘ground-up’ approach I have endorsed, exploring the impact of disability on agency requires looking at specific disabilities. Here, I investigate Parkinson’s, a paradigmatic physical disability, a number of disabilities that impact an agent’s emotions, and then briefly address some of the wide range of intellectual disabilities.

3.1 Parkinson’s disease The primary pathophysiologies for Parkinson’s are tremor, rigidity, posture and locomotion disorders, and akinesia. While these pathophysiologies can’t be isolated from each other in many individuals with Parkinson’s, the present discussion will focus primarily on akinesia— the ‘lack of movement or slowness of initiating and maintaining movement’ primarily due to the disease’s effect on neurons in the substantial nigra region of the brain, an area ‘which is important for control and regulation of motor activity’ (Weiner et al. 2001, 5). Akinesia makes it difficult to walk or engage in other whole body movements, particularly when they involve more than one action plan. (This is one of the reasons that the pathophysiologies in Parkinson’s can’t be separated from each other.) For those with Parkinson’s, ‘normal voluntary actions may be impaired by difficulty in initiation as well as slowness of movement, both of which may be apparent during most activities. In addition, sudden freezing or involuntary cessation of ongoing activity, which is referred to as kinesia paradoxa, may be seen’ (Donaldson et al. 2012, 249). Most individuals with Parkinson’s are able to briefly overcome, or at least mitigate, their akinesia or other motor control difficulties. However, the high levels of concentration required to do so can’t be sustained, and thus the reprieve from akinesia is often short-lived. The difficulty to initiate movements isn’t consistent across contexts. The difficulty in self-initiating movement is significantly more pronounced than in response to external command by another agent, which is ‘relatively well preserved’ (Donaldson et al. 2012, 237). Furthermore, the difficulty initiating temporally disappears in certain environmental setting, such as the ringing of a fire alarm or a gunshot (Donaldson et al. 2012, 249). This suggests that whether an individual is prevented from executing a particular task depends not just on the underlying disabling condition but also upon facts concerning one’s environment. A similar conclusion can be raised from the examination of other physical disabilities.2

3.2 Emotional blunting and alexithymia Let’s turn to emotional disabilities. Numerous disabilities lead to emotional blunting or flattened affect, a decrease in the frequency or strength of emotions, both positive and 161

Kevin Timpe

negative (Kim 2015). Emotional blunting is commonly associated with schizophrenic syndrome and frontotemporal dementia, or FTD (see Berenbaum et al. 1987, 57, and Williamson and Allman 2011, 104, respectively). In some cases, FTD’s emotional and social impact may be more significant than the cognitive and neuropsychological deficits it causes. Decreased social tact and propriety, abulia, disengagement, and decreased behavioral regulation can all be rooted in FTD’s effects on the emotions. Individuals impaired by FTD can become emotionally detached; have a decrease in autonomic emotional responsiveness; lose empathy and willingness to comfort others; and more frequently fail to comfort or help others in distress, even if they are family members or close friends (Mendez et al. 2006, 242–245). Emotional blunting can impair other emotional responses, not just those involving empathy. Individuals with 2p15–16.1 microdeletion syndrome, which involves a deletion on the short (p) arm of chromosome 2, typically involves a number of physical affects (e.g., microcephaly, vision problems, kidney abnormalities); speech impairments; gross and fine motor control issues; and cognitive and developmental disabilities (2p15p16.1 microdeletion syndrome 2014, 4 and 7; see the longer discussion in Timpe 2016). It also typically involves mild to severe intellectual disability and problems with executive function, which can lead to both emotional blunting and alexithymia (Hancarova et al. 2013, 2). Though not an official diagnosis in the DSM-V, alexithymia is ‘marked by difficulties in identifying and describing feelings and difficulties in distinguishing feelings from the bodily sensations of emotional arousal’ (Bird et al. 2010, 1517; see also Ricciardi et al. 2015). Alexithymia, like emotional blunting, has been clinically associated with reduced empathy (Bird and Cook 2013). While there’s not as much evidence to be sure of this stronger claim, there’s at least anecdotal evidence suggesting that some individuals with alexithymia may have difficulties experiencing certain emotions (e.g., shame, jealousy, or self-resentment) altogether. Individuals with Down syndrome often have difficulty identifying and labeling fear, anger, and surprise, and some individuals with autism show differences in emotional self-reports and expression (Kasari et al. 2012, 240 and 244). Insofar as human agency involves emotional regulation which contributes to agency (see Carla Bagnoli’s chapter in this volume), we see here another range of influences of disabilities on agency.

3.3 Intellectual disability Intellectual disability is a particularly challenging category to address. It is defined by the WHO as: a significantly reduced ability to understand new or complex information and to learn and apply new skills (impaired intelligence). This results in a reduced ability to cope independently (impaired social functioning), and begins before adulthood, with a lasting effect on development. (WHO 2010)3 Similarly, the DSM-V points to deficits in both intellectual functions and adaptive functioning as a result of those intellectual functions as essential elements in the diagnostic criteria for intellectual disability. There are over 1,000 known etiologies of intellectual disability, and each ‘differs from all the others in meaningful ways in virtually every aspect of functioning’ (Burack 2012, 4). Furthermore, over half of individuals with intellectual disability show no known cause for their disability (Iarocci and Petrill 2012, 13). 162

Agency and disability

Intellectual disability also ranges widely in terms of the particular capacities and abilities impacted, as well as the degree to which they are affected. All of the following are included by psychologists and neurologists as belonging to the set of capacities and abilities involved with intellectual disabilities: spatial mapping; attentional abilities; memory, both short and long term and working; cognitive flexibility; linguistic skills including vocabulary comprehension, expression, lexical and syntax skills, grammatical morphology, and pragmatics; literacy; conversational skills; and comprehension. Intellectual disabilities4 can impact agential planning, learning, adaptive behavior, and social interaction. Intellectual disabilities can make it hard for an agent to exercise a range of other skills well and to plan well for the future. In some cases, those with intellectual disabilities might be ‘unable to imagine that range of alternative possible futures that are, given their social circumstances, futures that it would be realistic for them to attempt to make their own’ (MacIntyre 1999, 94). Regarding the degree of impact, intellectual disability ranges from mild learning or developmental disabilities to anencephaly, where the neural tube fails to close during neonatal development resulting in the lack of a cerebrum. Most cases of anencephaly result in miscarriage, and those that are born alive usually die from cardiorespiratory arrest within days. The DSM-V specifies four levels of severity of intellectual disability: mild, moderate, severe, profound involving evaluation across conceptual, social, and practical domains. Unlike previous classifications, the DSM-V has moved away from primary reliance on IQ tests toward a combination of clinical assessment and standardized testing for diagnosis.5 There is a danger of focusing on the agential impact of intellectual disability, particularly in severe cases of intellectual disability, given the ways that those with intellectual disabilities have been mistreated, even to the point of institutionalization and forced sterilization in even not-too-distant history. As indicated earlier, my discussion of physical, emotional, and intellectual disabilities isn’t exhaustive of the ways that disabilities can affect human agency. All sorts of other disabilities not examined here have a wide variety of implications for agency. Depression, which according to the National Institutes of Mental Health is the leading cause of disability worldwide, impacts not only the emotions but also the motivational capacities involved in agency (for work on the connection between depression and moral psychology, see Ardal 1993; Caton 1986; Hansen 2004; Roberts 2001; and Silberfeld and Checkland 1999). Lesch-Nyhan syndrome causes an overproduction of uric acid that disturbs the nervous system causing cognitive, neurological, and behavioral abnormalities such as choreoathetosis and other involuntary movements, ballismus, and self-injurious behaviors (see Lloyd 1981). The discussion in this section is intended to merely be highlight a limited cross-section of the total impact of disabilities on agency.

4 Lessons from disability for agency in general I want to draw three lessons about human agency on the basis of the disabilities canvased in Section 3, even though they don’t exhaust the importance of disability for a full understanding of human agency. Many scholars argue that agency, and moral agency in particular, should be understood as a degreed concept. For instance, Jeannett Kennett writes, Moral responsibility comes in degrees. The ordinary view implicitly recognizes both degrees of difficulty in the exercise of self-control (and indeed of judgement), and a distinction between those who are capable of synchronic self-control and those who must 163

Kevin Timpe

instead rely on diachronic techniques of control. Factors which impinge on the ease with which the capacity for self-control can be exercised mitigate responsibility. Some of those are obstacles to good judgement as well: for example, tiredness, emotional pressures, and lack of information. (Kennett 2001, 182; see also Timpe 2016, particularly section 3) If tiredness can impact self-control, then surely disabilities can as well (for further discussion, see Timpe 2016). And self-control, like many other agential abilities, is best thought of as degreed (see Smith 2017). Volitional control, rationality, emotional regulation, bodily control, sensory awareness all are degreed. The facts that disabilities evidence considerable heterogeneity and that the same disability can manifest in such a wide range of agential impact (e.g., individuals with Down syndrome exhibit significant differences in terms of their degree of intellectual impairment) give us further reason to think that agency is a degreed concept. In part because we typically think of idealized agents rather than agents with disabilities, we can fail to take seriously enough the degreed nature of the capacities and behaviors involved in human agency. (I think a similar point is true of much philosophical reflection on mental illness and childhood development too, though I don’t develop this line of inquiry here.) Second, reflection on disability and agency reminds us that our agency is socially and ecologically dependent. Much contemporary philosophy seems to endorse an atomistic and individualistic approach to human agency. But reflection on disabilities shows that the social context of our agency matters. The capacities involved in human agency themselves might depend on the agent’s social or environmental situation, so agency itself might depend, at least partially, on those social or environmental factors. On this understanding, we can change what the agent is able to do either by changing factors intrinsic to the agent or by changing the environment the agent is in. A slightly more modest claim is certainly true: the agent’s exercise of those capacities depends on the social or environmental setting the agency takes place in. On this understanding, what the agent can do is just a function of the agent, but what they will do depends on the social or environmental situation and the number of supports that it provides (see Timpe 2019). Human agency can have a social scaffolding, ‘the externalization of certain parts of the decision making process [or agency in general] that are not typically externalized,’ and ‘this externalization does not undermine the claim to autonomy’ (DeVidi 2013, 193). Certain disabilities can decrease forms of autonomy, but not all disabilities do. Furthermore, it is not true that all disabilities completely undermine autonomy. The sorts of autonomy ruled out by disabilities are plausibly forms that aren’t possessed by human agents even apart from disability. As Alasdair MacIntyre has argued, no human is a ‘fully independent practical reasoners’ (see Clifton 2018; MacIntyre 1999). A danger in defining intellectual disability in terms of ‘a reduced ability to cope independently’ is that it wrongly suggests that human agents are more independent than they really are (see Clifton 2018, 131f ). Overstating the independence of those without disabilities can increase the disenfranchisement of those with disabilities (Kittay 1998, 77). Even the independence that we do have develops in the context of our social environments. Finally, the third related lesson is that there are degrees of difficulties involved even in successful human agency. Consider again the discussion of Parkinson’s. An agent’s ability to self-initiate movement depends on the degree of their condition’s progression, how long they’ve been exerting this kind of control, and environmental variables. How likely they

164

Agency and disability

are to succeed will vary with all these factors. Disabilities may increase the fallibility of an agents’ abilities, but there are good reasons to think that abilities in general are fallible (see Smith 2017, section 3). Furthermore, it seems that not just agency itself but also morally responsible agency is degreed. Dana Nelkin convincingly argues that the degree of difficulty can affect the amount of blameworthiness or praiseworthiness: For example, we often excuse people to an extent when doing the right thing would be very, very difficult. In turn, difficulty can be understood in at least two ways: on the one hand, it can be understood as requiring a great deal of effort, and, on the other, it can be understood as requiring a great sacrifice of one’s interest. These often go together, but they might come apart. (Nelkin 2014, 357) Disabilities can impact both of these ways of understanding the relevant sense of difficulty.

5 Conclusion Much of recent philosophical work on agency focuses on instances of what might be thought of as ‘typical’ agency, concentrating on ‘clear- cut paradigm’ agents while bracketing issues related to developmental psychology, mental illness, or disability (Shoemaker 2015, 5). There is very little discussion of what Shoemaker calls ‘marginal agents’ even though these cases can teach us quite a bit. While this restriction and the related idealization of human agency might sometimes be justified, it can result in a skewed understanding of human agency. As I’ve tried to argue, we can learn about ‘typical’ agency and wander in the direction of ‘the margins’ insofar as there might be facts about human agency, such as its socially embeddedness, that we can recognize more clearly in cases involving disability. For instance, once we see that human agency is socially embedded, we can work to provide ecological structuring and social scaffolding that can lead to better expressions of human agency. (For examples of ecological structuring and social scaffolding, see Timpe 2019, particularly section 2.1.) Finally, we ought not overly associate disabilities with challenges to successful agency. As indicated above, a wide range of disabilities can have this kind of impact. But some disabilities can also make a range of agential behaviors, in the sense spelled out in section 1, easier. For instance, some individuals with certain forms of autism spectrum disorders can have heightened focus and an increased range of executive function tasks, as well as increased sensitivity with respect to vision, hearing, or touch—all of which can impact the agent’s behavior (see Crane and Goddard 2009). Similarly, an increased emotional attention and sensitivity has been found in individuals with Williams syndrome (Niccols et al. 2012). Further attention needs to be paid to these sorts of impacts as well so that we don’t reinforce problematic stereotypes of disabilities.

Related topics Agency, powers, and skills; Expert agency; Agency and emotion; Agency and responsibility; Agency and mistakes; Pathologies of agency: Agency and mistakes; Agency and autonomy.

165

Kevin Timpe

Notes 1 My own skepticism of top-down approaches can be found in Timpe (2022). See also Barnes 2016: 3. Barnes isn’t committed to the claim that there isn’t a unified category of disability. Rather, on her view the mere fact that we use the word ‘disability’ with this range of modifiers doesn’t entail that there is. 2 It is common to differentiate physical disabilities from, among other categories, intellectual disabilities. But the vast majority of the over 1,000 known etiologies associated with intellectual disability have physical consequences. Many intellectual disabilities are caused by genetic abnormalities; even with environmental causes, what they cause is changes in various bodily systems (e.g., lead ingestion, which is estimated to be the cause of approximately 10% of intellectual disability worldwide). Demarking physical disabilities from other subtypes is, in my view, significantly more vexed than often admitted. 3 According to Moss, Howlin, Oliver: ‘the diagnostic criteria outlined by the DSM-IV-TR (APA 2000) and OCD-10 (WHO 1992) manuals may not be sensitive enough to distinguish between individuals who have not yet attained the appropriate level of development required to demonstrate a particular skill and those who show a genuine impairment in those skills’ (Moss 2012, 293). For a discussion of criticisms of the WHO’s approach to defining intellectual disability, see Buntinx 2014. 4 Licia Carlson and Eva Feder Kittay (2010, 1 note 1) write that ‘We’ve chosen the term “cognitive disability” under which we include conditions like autism, dementia, Alzheimer’s, and [what has historically been called] mental retardation, rather than “intellectual disability.” The former is broader. Also, some forms of cognitive disability do not imply diminished intellectual capacity (e.g., autism)’ (see also Carlson 2010). My interest here is specifically with intellectual disability, though the exact relationship between the two is complex. Furthermore, the distinction between them is not always held. 5 The history of IQ tests in general and specifically their role in evaluating intellectual disability is contested (see Harris 2006, particularly chapters 2 and 3).

Further reading Barnes, Elizabeth. 2016. The Minority Body: A Theory of Disability. Oxford: Oxford University Press. An excellent introduction into some of the philosophical issues surrounding specifically physical disabilities. Barnes argues that physical disabilities involve ‘mere-difference,’ not ‘bad-difference.’ MacIntyre, Alasdair. 1999. Dependent Rational Animals: Why Human Beings Need the Virtues. Peru, IL: Carus Publishing Company. MacIntyre’s book focuses primarily on ‘the virtues of acknowledged dependence’ but canvasses a wide range of issues. Of particular relevance here are his arguments against overly idealized and atomistic conceptions of human agency. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. While not specifically focused on disability, Shoemaker’s book illustrates how reflection on ‘nonstandard’ or ‘marginal’ cases of agency and responsibility can help us understand important facts about human agency in general. Timpe, Kevin. 2019. “Moral Ecology, Disabilities, and Human Agency,” Res Philosophica 96.1: 17–41. This article argues that human agency is not simply a function of intrinsic properties about the agent, but rather depends on the ecology that the agent is in. In particular, it shows how, by paying deliberate attention to structuring the social environment around people with disabilities, we can mitigate some of the agential impact of those disabilities.

References 2p15p16.1 Microdeletion Syndrome. 2014. Surrey, England: Rare Chromosome Disorder Support Group. Ardal, Pall S. 1993. “Depression and Reason.” Ethics 103: 540–550. Barnes, Elizabeth. 2016. The Minority Body: A Theory of Disability. Oxford: Oxford University Press. Berenbaum, Sheri A., Richard Abrams, Samuel Rosenberg, and Michael Alan Tylor. 1987. “The Nature of Emotional Blunting: A Factor-Analytic Study.” Psychiatry Research 20: 57–67.

166

Agency and disability Berger, Ronald J. 2013. Introducing Disability Studies. Boulder, CO: Lynne Rienner Publishers. Bird, Geoffrey and Richard Cook. 2013. “Mixed Emotions: The Contribution of Alexithymia to the Emotional Symptoms of Autism.” Translational Psychiatry 23.3: 10.1038/tp.2013.61. Bird, Geoffrey, Giorgia Silani, Rachel Brindley, Sarah White, Uta Frith, and Tania Singer. 2010. “Empathic Brain Responses in Insula Are Modulated by Levels of Alexithymia but Not Autism.” Brain 133.5: 1515–1525. Buntinx, Wil H. E. 2014. “Understanding Disability: A Strengths-Based Approach,” in The Oxford Handbook of Positive Psychology and Disability, ed. Michael L. Wehmeyer. Oxford: Oxford University Press: 7–18. Burack, Jacob A., Natalie Russo, Heidi Flores, Grace Iarocci, and Edwards Zigler. 2012. “The More You Know the Less You Know, But That’s OK: Developments in the Developmental Approach to Intellectual Disability,” in The Oxford Handbook of Intellectual Disability and Development, ed. Jacob A. Burack, Robert M. Hodapp, Grace Iarocci, and Edward Zigler. Oxford: Oxford University Press: 3–10. Carlson, Licia. 2010. The Faces of Intellectual Disability: Philosophical Reflections. Bloomington, MA: Indiana University Press. Carlson, Licia and Eva Feder Kittay. 2010. “Introduction: Rethinking Philosophical Presumptions in Light of Cognitive Disability,” in Cognitive Disability and its Challenges to Moral Philosophy, ed. Eva Feder Kittay and Licia Carlson. Malden, MA: Wiley-Blackwell: 1–25. Caton, Hiram. 1986. “Pascal’s Syndrome: Positivism as a Symptom of Depression and Mania,” Zygon 21.3: 319–351. Clifton, Shane. 2018. Crippled Grace: Disability, Virtue Ethics, and the Good Life. Waco, TX: Baylor University Press. Crane, L., L. Goddard, and L. Pring. 2009. “Sensory Processing in Adults with Autism Spectrum Disorders,” Autism 13: 215–228. DeVidi, David. 2013. “Advocacy, Autism and Autonomy,” in The Philosophy of Autism, ed. Jami L. Anderson and Simon Cushing. Bloomington, IN: Indiana University Press, 187–200. Donaldson, Ivan, C. David Marsden, Susanne A. Schneider, and Kailash P. Bhatia. 2012. Marsden’s Book of Movement Disorders. Oxford: Oxford University Press. Hancarova, Miroslave, Martina Simandolva, Jana Drabova, Katrin Mannik, Ants Kurg, and Zdenek Dedlacek. 2013. “A Patient with De Novo 0.45 Mb Deletion of 2p16.1: The Role of BCL11A, PAPOLG, REL, and FLJ16341 in the 2p15-p16.1 Microdeletion Syndrome.” American Journal of Medical Genetics, Part A 9999: 1–6. Hansen, Jennifer. 2004. “Affectivity,” in The Philosophy of Psychiatry: A Companion, ed. Jennifer Radden. Oxford: Oxford University Press: 36–53. Harris, James C. 2006. Intellectual Disability: Understanding its Development, Causes, Classification, Evaluation, and Treatment. Oxford: Oxford University Press. Iarocci, Grace and Stephen A. Petrill. 2012. “Behavioral Genetics, Genomics, Intelligence, and Mental Retardation,” in The Oxford Handbook of Intellectual Disability and Development, ed. Jacob A. Burack, Robert M. Hodapp, Grace Iarocci, and Edward Zigler. Oxford: Oxford University Press: 13–29. Kahane, Guy and Julian Savulescu. 2009. “The Welfarist Account of Disability,” in Disability and Disadvantage, ed. Kimberley Brownlee and Adam Cureton. Oxford: Oxford University Press: 14–53. Kasari, Connie L., Laudan B. Jahromi, and Amanda C. Gulsrud. 2012. “Emotional Development in Children with Developmental Disabilities,” in The Oxford Handbook of Intellectual Disability and Development, ed. Jacob A. Burack, Robert M. Hodapp, Grace Iarocci, Edward Zigler. Oxford: Oxford University Press: 239–253. Kennett, Jeanette. 2001. Agency and Responsibility: A Common-Sense Moral Psychology. Oxford: Oxford University Press. Kim, Cynthia. 2015. Nerdy, Shy, and Socially Inappropriate: A User Guide to an Asperger Life. London: Jessica Kingsley Publishers. Kittay, Eva Feder. 1998. Love’s Labor: Essays on Women, Equality, and Dependency. New York: Routledge. Lloyd, Kenneth G., Oleh Hornykiewicz, Lynne Davidson, Katherine Shannak, Irene Farley, Menek Goldstein, Masato Shibuy, William N. Kelley, and Irving H. Fox. 1981. “Biochemical Evidence of Dysfunction of Brain Neurotransmitters in the Lesch-Nyhan Syndrome,” The New England Journal of Medicine 305.19: 1106–1111. MacIntyre, Alasdair. 1999. Dependent Rational Animals: Why Human Beings Need the Virtues. Peru, IL: Carus Publishing Company.

167

Kevin Timpe Mendez, Mario F., Aaron McMurtray, Eliot Licht, Jill S. Shapira, Ronald E. Saul, and Bruce L. Miller. 2006. “The Scale for Emotional Blunting in Patients with Frontotemporal Dementia.” Neurocase 12: 242–246. Moss, Joanna and Patricia Howling, and Chris Oliver. 2012. “The Assessment and Presentation of Autism Spectrum Disorder and Associated Characteristics in Individuals with Severe Intellectual Disability and Genetic Syndromes,” in The Oxford Handbook of Intellectual Disability and Development, ed. Jacob A. Burack, Robert M. Hodapp, Grace Iarocci, and Edward Zigler. Oxford: Oxford University Press: 275–299. Nelkin, Dana. 2014. “Difficulty and Degrees of Moral Praiseworthiness and Blameworthiness.” Nous 50.2: 356–378. Niccols, Alison, Karen Thomas, and Louis A. Schmidt. 2012. “Socioemotional and Brain Development in Children with Genetic Syndromes Associated with Developmental Delay,” in The Oxford Handbook of Intellectual Disability and Development, ed. Jacob A. Burack, Robert M. Hodapp, Grace Iarocci, and Edward Zigler. Oxford: Oxford University Press: 254–274. Proust, Joëlle. 2013. “Mental Acts,” in A Companion to the Philosophy of Action, ed. Timothy O’Connor and Constantine Sandis. Malden, MA: Wiley-Blackwell: 209–217. Ricciardi, Lucia, Benedetta Demartini, Aikaterini Fotopoulou, and Mark J. Edwards. 2015. “Alexithymia in Neurological Disease: A Review.” The Journal of Neuropsychiatry and Clinical Neurosciences 27.3: 179–187. Roberts, John Russell. 2001. “Mental Illness, Motivation and Moral Commitment,” Philosophical Quarterly 51.202: 1–59. Shakespeare, Tom. 2014. Disability Rights and Wrongs, 2nd ed. London: Routledge. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Silberfeld, Michel and David Checkland. 1999. “Faulty Judgment, Expert Opinion, and Decision-Making Capacity.” Theoretical Medicine and Bioethics 20: 377–393. Smith, Will. 2017. “Agency and Practical Abilities.” Royal Institute of Philosophy Supplement 80: 235–264. Timpe, Kevin. 2016. “Executive Function, Disability, and Agency.” Res Philosophica 93.4: 767–796. first Res Phil Timpe, Kevin. 2019. “Moral Ecology, Disabilities, and Human Agency,” Res Philosophica 96.1: 17–41. Timpe, Kevin. 2022. “Denying a Unified Concept of Disability.” The Journal of Medicine and Philosophy. UPIAS. 1976. The Union of the Physically Impaired against Segregation and Disability Alliance discuss Fundamental Principles of Disability. London: The Disability Alliance. Weiner, William, Lisa Shulman, and Anthony E. Lang. 2001. Parkinson’s Disease. Baltimore, MD: The Johns Hopkins University Press. WHO. 2010. “Definition: Intellectual Disability.” http://www.euro.who.int/en/health-topics/ noncommunicable-diseases/mental-health/news/news/2010/15/childrens-right-to-family-life/ definition-intellectual-disability. Williamson, Peter C. and John M. and Allman. 2011. The Human Illness: Neuropsychiatric Disorders and the Nature of the Human Brain. Oxford: Oxford University Press.

168

15 PATHOLOGIES OF AGENCY Lubomira Radoilska

Mental disorders and anomalous psychological conditions play a crucial role in defining the nature and scope of agency. The standard approach describes a particular set of symptoms as a point of contrast where some core features identified in paradigm cases are present, while others are clearly absent. For instance, the literature on psychopathy (e.g., Schramme 2014) often points to the combination of single-minded pursuit of personal goals with lack of empathy leading to casual disregard for the interests of others. On this picture, psychopathy instantiates a distinctive shape of agency where some reasons can be as integrated as in central cases of successful planning, whereas others get no purchase. Alien hand syndrome (where the unbuttoning of one’s own shirt with one hand is vigorously fought off with the other, Pacherie 2007) is a further example. It shows how intentional agency might unfold when the ability for basic bodily movements becomes fragmented. Finally, people with depression may be motivated to act in ways they manifestly disvalue (Radoilska 2013a). The experience is both alienating and disturbing. It decouples first-personal knowledge of the reasons for which one acts from the self-understanding that acting on one’s own reasons is supposed to secure. These examples show that close attention to the phenomenology of specific disorders and anomalous conditions helps differentiate between agential abilities that textbook cases of action bundle together. This is the first step of a heuristic that explores the possible interactions between facets of agency that typically come together but are functionally and conceptually separate. The payoff is twofold. First, we gain a clearer insight into manifestations of agency beyond straightforward intentional actions, including negative, omissive and second-order exercises, such as permitting, preventing, facilitating, or contravening (Alvarez 2013). Second, we can better understand and assess competing attributions of responsibility by mapping them onto different patterns of agency (Shoemaker 2015). At the same time, this heuristic might become circular and uninformative. The danger comes from sticking to the phenomenology so closely that no distinction is made between pathologies of agency and mere sources of impediments and distortion. Without such distinction, the lessons learnt about diverse facets and shapes of agency in the context of mental disorders could become unreliable, for contours of agency might be unhelpfully merged with its circumstances. To forestall this danger, we need to complement the mainstream heuristic. This chapter will try to do so by expanding on the notion of necessarily less-than-successful agency (Radoilska 2013b). A major advantage of this proposal is to do DOI: 10.4324/9780429202131-19169

Lubomira Radoilska

justice of ‘achievement’ as unifying idea that underpins all credible manifestations of agency (Bradford 2015). By focussing on how and why achievement might be affected by different mental disorders and anomalous psychological conditions, it becomes possible to avoid the circularity invited by the mainstream heuristic: falling back on a clinical diagnosis to pinpoint a pathology of, as opposed to an obstacle to, agency. To see this, the notion of necessarily less-than-successful agency will be employed to identify and explore specific pathologies within two dimensions of agential achievement: intention and responsiveness to reasons.

1 Intention, practical knowledge and planning Intention and intentional action are central manifestations of agency, to the point of obfuscating alternatives. To appreciate the attractiveness of thinking about agency through the lens of intentional action, let us consider the influential account of Anscombe (1963). On this account, intentional actions are the subject of a distinctive kind of direct first-personal knowledge—unmediated by observation, inference or reflection—which should not be equated with introspection or general awareness of one’s own mental states. For such a move would reduce the agent’s perspective to that of a well-placed, ‘inside’ observer and sever the ‘mind-world’ relation that comes with intention as stretching out into world (Moran and Stone 2009). Yet both perspective and relation are irreducible aspects of the relevant kind of fundamentally practical first-personal knowledge. The practical nature of knowing one’s intentional actions has direct implications regarding success in action. The success criteria are set out by the agent’s account of what they are doing. When this account differs from what actually happens, the ‘words impugn the facts’ as Anscombe puts it, not the other way around. For the failure to conform betrays an error in performance, not an error in judgement. Importantly, while such errors in performance are rare, agents often miss on their overall objectives. This observation shows that two kinds of practical knowledge may come apart: knowing what one is doing intentionally at present on the one hand, and generalised know-how that enables agents to plan and coordinate beyond the performance of discrete intentional actions on the other (but see Setiya 2008 on the intimate relation between these two kinds of knowledge). Satisfying the criteria of success set out with respect to the first kind of practical knowledge might not suffice overall. On these criteria, individual actions can be successful independently of whether they add up to a defensible strategy that brings them together. As a result, we cannot tell between a savvy agent who keeps an eye on longer-term objectives and one who gets bogged down in unnecessary detail. If agential achievement is to be assessed in a meaningful way, another set of criteria, linked to the second kind of practical knowledge—generalised know-how exercised over time, across different situations—becomes indispensable. This insight has been taken forward by the planning theory of intention (Bratman 1987, 2007). According to this view, an intention has an irreducibly dual function: it guides the performance of individual actions and it connects with other intentions to enable the pursuit of complex, temporarily extended projects. Looking at intentional agency from this integrative perspective, we can gauge a new set of success criteria to complement the idea of practical knowledge. Roughly, success is a matter of striking a balance between the requirements of coherence across plans and consistency over time with the need for plans to be flexible in response to changing circumstances. Let’s call this task ‘practical rationality’ to distinguish it from ‘practical knowledge’. 170

Pathologies of agency

With this distinction in mind, let us focus on a puzzling pathology of agency where the notion of necessarily less-than-successful agency comes into its own. Jill has an important presentation to make the following day early in the morning. Before heading back home after work, her colleagues invite her to join them for a drink. She decides to go out but have no more than two drinks as she knows that drinking any more will affect the quality of her performance next morning. As the evening progresses, she is offered a third drink, which, after a brief moment of hesitation, she takes. Against her better judgement, she ends up having a fourth drink as well. This kind of scenario abounds in the literature on weakness of will. There are two mainstream accounts of what goes wrong with Jill. According to Mele (1987), the problem is Jill’s acting against her better judgement at the time of action. She exhibits weakness of will on this occasion, which constitutes a failure of practical rationality. We don’t need to know anything further about her as an agent to reach this conclusion. All relevant information is already contained in this description. According to Holton (2009), instead, the scenario hints at a possible case of weakness of will if Jill’s decision to have only two drinks is a personal policy adopted in light of previous trouble with alcohol. To be deemed as weak-willed and therefore practically irrational, Jill’s failure to see through her good intentions must be representative of a pattern rather than a one-off. The two mainstream accounts derive from the two perspectives on what counts as success in action. While the first focusses exclusively on the standard set out by a specific intention, the second also considers the diachronic implications of any particular failure. These differences notwithstanding, both accounts concur on weakness of will being a failure of agency in contrast to compulsion, a pathology of agency whose distinguishing feature is loss of agential control. The thought is that a pathology of agency would fall outside the scope of criticisable irrationality, which makes failures of agency, such as weakness of will, intelligible (Davidson 2001). In other words, pathological agents do not stand a chance to succeed. And so, their irrationality is regrettable rather than criticisable. Although initially appealing, the demarcation line between failing and pathological agency that compulsion provides is ultimately misleading. For it assumes a one-dimensional notion of success in action as bringing about a desirable result—be it at a snapshot or in a temporarily extended frame. Yet on closer inspection, weakness of will and compulsion have a central common feature that helps elucidate and differentiate among these closely related phenomena: being necessarily less than successful with respect to a particular strand of one’s goal-directed conduct rather than one’s overall agency. This feature boils down to neither suboptimal behaviour, nor a straight failure. Instead, it points to a distinctive structure of agency, the manifestations of which cannot be fully successful to the extent that they arise at all (Radoilska 2013b). Unlike the two mainstream accounts, the notion of necessarily less-than-successful agency implies a more complex model of action as actualisation, where success is defined across two complementary dimensions: production (bringing about an effect) and assertion (an agent’s articulating a particular commitment of theirs). Only when these two dimensions are well aligned is an action successful on its own terms. By contrast, when they are misaligned in a distinctive and sustained way, rather than just coming apart, the ensuing actions are necessarily less than successful. Returning to compulsion and weakness of will, the misalignment that transpires in both takes the following form: each is successful as production to the extent that it is unsuccessful as assertion. To illustrate with the scenario we considered, what Jill does is successful as production (she gets to have more than two drinks) to the extent that it 171

Lubomira Radoilska

is unsuccessful as assertion (it goes against her better judgement and/or policy). There is an underlying conflict between valuing and intending with respect to drink that we may call Jill’s original akratic moment. If this conflict is not addressed but instead keeps coming back, it eventually solidifies into a necessarily less than successful strand of Jill’s agency. Is this a case of weakness of will or compulsion? Does it amount to a pathology as opposed to a failure of agency? Addressing the latter issue first, we can see that partial, yet unavoidable, failure is constitutive of pathological agency. Far from placing agents outside the space of reasons, where the charge of being practically irrational is no longer apt, pathologies of agency provide this charge with a specific focus. This explains why compulsion and the binary notion of control versus irresistibility it implies cannot give us a reliable cutting point. In fact, many cases, such as the preceding scenario, would satisfy the criteria of necessarily less-than-successful agency that underpin both phenomena. So, the differences we may find will be of degree rather than kind. When it is more illuminating to conceive the pathology manifested in terms of compulsion rather than weakness of will, the relevant actions are experienced as particularly frustrating attempts to resolve the recurring conflict between valuing and intending that effectively extend the original akratic moment to the detriment of other, potentially successful pursuits of one’s agency. This feature is at the heart of debates about the extent to which addictive behaviours can be treated as responsive to reasons (Poland and Graham 2011). In the next section, we will explore how and why responsiveness to reasons might be affected by various pathologies of agency, leading to differential appraisals of agents as members of the moral community.

2 Responsiveness to reasons, answerability and the moral community The Anscombian notion of practical knowledge sets out an immediate test of whether an action is successful. Knowing what one is doing without having to observe oneself or infer from prior experiences provides the description under which what one is doing is intentional. For instance, I don’t need to look up at what I am doing to know that I am currently opening a window, if that is an intentional action of mine rather than something I find myself doing. Nor do I need to work out what I am up to in similar cases. Thus, an agent’s performance is assessed, in the first instance at least, against such a description. What makes it special is that it designates an action as undertaken in the light of the agent’s reasons as opposed to behaviours understood in causal terms. Knowing what one is doing in the relevant practical sense rests on knowing why one is doing it. My giving account of what I am doing is unlike my reporting on what is happening around me—with or without my causal involvement. This is the guiding idea behind the Anscombian question ‘Why?’. It is helpful to contrast the responses this question is meant to elicit with action explanations afforded by the so-called ‘reasons why’ (Dancy 2000), where the language of reasons is used without referring to any considerations in the light of which an agent acts. Examples include my failing to turn up to a party because I forgot about it or because I am too shy. Neither figures in my reasoning whether to go the party and yet either can rationalise my not being there. In this respect, they both could present bona fide ‘reasons why’ for my staying away from the party. In so doing, however, they would mark out what I did as behaviour, where the Anscombian question ‘Why?’ is denied application. ‘Reasons why’ do not warrant the kind of knowledge agents have of actions they perform in the light of reasons. What is lost is the immediacy and certainty of practical knowledge: knowing what one is doing, as Hornsby puts it, ‘without recourse to further knowledge’ or ‘just like that’ (2013, p. 16). 172

Pathologies of agency

The expectation of an immediate, unbreakable bond between actions done for a reason and actions of which agents have practical knowledge is put to the test by apathy or auto-activation deficit (AAD) syndrome, a neurological disorder where self-generated voluntary and purposive actions are virtually absent, while externally driven behaviours are normally executed. As Levy (2012, p. 590) explains: This syndrome consists in a loss of spontaneous activation in three different domains: behaviour, cognition and emotion. Patients tend to remain quietly in the same place or position all day long, without speaking or taking any spontaneous initiative. When questioned, patients express the feeling that their mind is empty when they are not stimulated… [however] when solicited, patients can produce relevant answers and behaviours. The following case studies illustrate the range of behaviours where apathy is at work: •

•

•

Patient A spent 45 minutes with his hands on a lawn mower, totally unable to initiate the act of mowing. The block disappeared instantaneously when his son told him to move (Laplane and Dubois 2001). Patient B developed a hobby of collecting broken TVs. His collection filled up the family flat, spilt over into the common premises of the building, risking eviction. B showed awareness of the consequences of his new hobby, denied interest in his collection and could not explain why he nevertheless kept on bringing more and more items (Levy 2015). Research subject C saw the keys for the researcher’s car on his desk, took the keys, got into the car, did a few rounds and returns to the room. When asked what she did, C was able to describe accurately her actions but could not provide any reason for them other than seeing the researcher’s car keys on the desk (ibid).

What makes these cases unsettling is the contrast between the lucidity with which people with apathy are able to report on what they are doing without having any inkling as to the reasons motivating their actions. They seem to be responding to cues or incentives in the environment as opposed to reasons they can recognise as their own. The difference between these two kinds of responses has significant implications for attributions of responsibility and moral appraisal. To appreciate this, consider the central place of notions, such as answerability, in mainstream conceptions of responsibility. In Hieronymi (2014), the practice of responsibility is essentially about the way in which an agent settles questions like whether to undertake a particular course of action, maintain an attitude or a relationship, or revisit an existing commitment. In all relevant cases, the agent is answerable to the Anscombian question ‘Why?’. This question tracks the reasons in the light of which the agent acted or refrained from acting, maintained or revisited attitudes, commitments or relationships of hers. Answerability provides an attractive model able to account for the interpersonal significance of allocating responsibility as opposed to a disinterested study of why people do the things they do. For it speaks directly to the idea of a counterfactual conversation with those affected by one’s actions or an internal dialogue that mainstream conceptions of responsibility build upon (e.g. McKenna 2012, Wallace 1994). In such a communicative setting, giving reasons for one’s actions plays a key role in being and holding responsible (Smith 2007). It involves anticipating challenges, but also showing due concern for others’ interests 173

Lubomira Radoilska

and perspectives. In this way, reason-giving demonstrates an agent’s secure belonging to the moral community where the so-called reactive attitudes, such as resentment or gratitude, the backbone of being and holding responsible, can be appropriately addressed (Strawson 1962). As seen earlier in cases of apathy, competently executed behaviours may not be responsive to reasons but only sensitive to incentives. Arguably, such behaviours would not qualify as appropriate ground for reactive attitudes, nor would reflect negatively on the moral standing of apathic agents: Patient B’s family did not reproach him for risking eviction, nor asked him to abandon his collection of broken TVs. Instead, they sought advice from an AAD specialist to see how best to manage the situation. Although in other cases the divide between reasons and incentives might not be as clearcut, the same logic would apply to separating out responsible conduct from conduct that should be exempt from responsibility. For instance, according to an influential account of addiction as incentive-sensitisation (Berridge and Robinson 2011), blame is an unsuitable response to addictive behaviours since they are not reasons-responsive. Instead, social effort should be directed at restructuring everyday environments so that people with addictions are not readily exposed to the incentives they have been sensitised to. In this respect, addiction is seen as an obstacle to personal agency rather than a pathology of it: there are situations where a person with a particular addiction cannot be expected to fare well. So, we better make it easy for her to avoid such situations rather than call on her to demonstrate uncommon strength of will and then berate her for not doing so. Importantly, when mental disorders and anomalous psychological conditions are conceived as obstacles to responsiveness to reasons within a particular sphere or context, they do not impact on the perception of the agents affected as members of the moral community. Research on the moral commitments and agency of people with autism (Kennett 2002) and intellectual disabilities (Shoemaker 2009) is a case in point. While in both instances, access to morally relevant reasons can be seen as impeded, there remain alternative pathways to reasons for caring for others and respecting their interests. This enables full participation in the practice of being and holding responsible in stark contrast with agent appraisal in cases of psychopathy. As Kennett (2002) and Shoemaker (2009) highlight, there is a qualitative difference in the way morally relevant reasons are present in practical reasoning in these contrasting cases. For instance, a person with a mild-to-moderate intellectual disability might be initially unable to see a particular course of action as morally required. However, once the rationale for it is brought to her attention, she would appreciate its binding force. By contrast, psychopaths are effectively able to contemplate morally relevant reasons. In this sense, they have unimpeded access to what morality requires of them. Yet, they do not treat these requirements as binding. Their response to moral reasons is warped rather than obstructed. As evidenced by the popular perception of psychopaths as amoral, or permanent outsiders to the moral community, indifference to the action-guiding aspect of morality grounds a particularly robust negative moral appraisal. Thus, psychopaths are treated as appropriate target for blame and resentment even though, ex hypothesi, they are not expected to engage in the kind of meaningful, reciprocal exchange that instantiates responsiveness to reasons proper. In this respect, psychopathic indifference to reasons for action that are nevertheless successfully identified bears the hallmarks of necessarily less-than-successful agency we discussed in the previous section. As in akratic action, the response to other-regarding reasons afforded by psychopathy cannot be fully successful to the extent that it takes place at all: for misrecognition is the mode of their cognition. The agency here is pathological not because psychopaths struggle with tasks, whose performance is typically taken for granted by others, as was the case with incentive sensitisation. Instead, the crux of the matter is that the stable, 174

Pathologies of agency

self-fulfilling mechanism underpinning necessarily less-than-successful agency makes alternative approaches to agential achievement, to which success and failure are equally open, riskier and less attractive as a result. Recent work on rationalisation shows how responsiveness to reasons in non-clinical populations might also exhibit a relevantly similar mechanism maintaining necessarily less-than-successful projects over time. As Schwitzgebel and Ellis (2017, p. 170) point out: Rationalisation occurs when a person favours a particular conclusion as a result of some factor (such as self-interest) that is of little justificatory epistemic relevance. The thinker then seeks an adequate justification for that conclusion but the very factor responsible for her favouring it now biases how the research for justification unfolds. As a result of an epistemically illegitimate investigation, the person identifies and endorses a justification that makes no mention of the distorting factor that has helped guide her search. There is a dissonance here between three categories of reasons: actual motivating reasons, ‘reasons why’ which explain the process of rationalisation and reasons which the agent ultimately professes as those in the light of which she made up her mind. Rationalisation is especially pernicious as it echoes the self-fulfilling mechanism we observed in other necessarily less-than-successful manifestations of agency. It effectively papers over the dissonance between different categories of reasons, making its detection and eventual resolution extremely unlikely. The upshot has direct implications for the practice of being and holding responsible. As Schwitzgebel and Ellis (2017, p. 171) observe, rationalisation not only ‘obstructs the critical evaluation of one’s own reasoning’ but also ‘impedes the productive exchange of reasons and ideas among well-meaning interlocutors’.

3 Implications for decisional capacity and autonomy in the context of mental disorder The preceding discussion showed that neither mental disorders nor abnormal psychological conditions necessarily imply a pathology of agency. Sometimes, their impact on agency is best understood in terms of an obstacle to sidestep or a constraint to overcome via targeted strategies. This conclusion finds further support in the literature on decisional capacity and personal autonomy in psychiatric contexts. For instance, many authors draw attention to the harmful nature of implicit assumptions about rationality, according to which a diagnosed psychiatric condition is enough to put a question mark on a person’s decisional capacity (Culvert and Gert 2004; Bolton and Banner 2012; Bortolotti 2013). In response to this, international psychiatric practice is developing towards robust and clear policies embedding the first-personal perspective and experiences of service-users at the heart of treatment (Widdershoven and Abma 2012; Potter 2013). The ambition is to prevent failures to recognise as reasons-responsive projects and commitments that are of great personal significance merely because they might seem unusual or unappealing from the perspective of a clinician. Yet, in the absence of a positive notion of what constitutes a pathology of agency, this mainstream approach can backfire in clinical settings. As Jamison (1995) and Radden (2012) poignantly illustrate in the context of bipolar disorder, it is equally important to address threats to personal autonomy that derive from having one’s behaviours that are not reasons-responsive treated as though they were. The unsympathetic imposition of penalties for such behaviours is an immediate issue. Examples include crippling loans undertaken during a manic episode, to fund out-of-character hobbies. A deeper concern, however, is 175

Lubomira Radoilska

that misallocating responsibility in this way would generalise any pathological strands of agency already present (Radoilska 2015). The concept of a necessarily less-than-successful structure could be of help here. For it allows us to pinpoint, clinical diagnosis notwithstanding, the kind of ambivalent, ultimately self-defeating achievements that may not be protected in the name of agent autonomy.

Related entries Intentional agency; Agency and disability; Agency and practical knowledge; Planning agency; Agency and responsibility; Agency and identification; Agency and reasons.

Further reading Fulford, K.W.M., Davies, M., Gipps, R.G.T., Graham, G., Sandler, J.Z., Stranghellini, G. and T. Thornton (eds.) 2013. The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press. A comprehensive overview of how and why mental disorder might impact different aspects of agency. It brings together conceptual analysis, empirical research, and clinical studies to shed light on philosophical and normative issues arising from psychiatric practice. Radoilska, L. (ed.) 2012. Autonomy and Mental Disorder. Oxford: Oxford University Press. The first book-length multidisciplinary inquiry into the nature and scope of personal autonomy in the context of mental disorder. It helps challenge tacit assumptions in the philosophy literature, according to which mental disorders are primarily threats to autonomous agency. Radoilska, L. 2013. Addiction and Weakness of Will. Oxford: Oxford University Press. Introduces the notion of necessarily less-than-successful action to qualify the kind of criticisable irrationality that underpins both addiction and weakness of will. It argues that being necessarily less-than-successful rather than a symptom or clinical diagnosis is what distinguishes pathologies from inner obstacles to agency. Schramme, T. (ed.) 2014. Being Amoral: Psychopathy and Moral Incapacity. MIT Press. The collection addresses the question of what makes the moral agency of psychopaths wanting. Against the grain, the contributors show that psychopathy is best understood as a cluster of deficits affecting different moral capacities rather than a uniform lack of empathy towards others. Shoemaker, D. 2015. Responsibility from the Margins. Oxford: Oxford University Press. The monograph develops a pluralist conception of responsible agency which can be broken down into three main facets: answerability, accountability and attributability. Different mental disorders and abnormal psychological conditions are shown to preclude some facets of responsible agency while at the same time leaving others intact.

References Alvarez, M. 2013. Agency and Two-Way Powers. Proceedings of the Aristotelian Society 113(1): 101–121. Anscombe, G.E.M. 1963. Intention, 2nd ed. Oxford: Blackwell. Berridge, K. and Robinson, T. 2011. Drug Addiction and Incentive Sensitization. In Poland, J. and Graham, G. (eds.) Addiction and Responsibility. Cambridge, MA: MIT Press; 21–53. Bolton, D. and Banner, N. 2012. Does Mental Disorder Involve Loss of Personal Autonomy? In Radoilska, L. (ed.) Autonomy and Mental Disorder. Oxford: Oxford University Press; 77–99. Bortolotti, L. 2013. Rationality and Sanity: The Role of Rationality Judgments in Understanding Psychiatric Disorders. In Fulford, K.W.M., Davies, M. et al. (eds.) The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press, 480–496. Bradford, G. 2015. Achievement. Oxford: Oxford University Press. Bratman, M. 1987. Intention, Plans and Practical Reason. Cambridge, MA: Harvard University Press. Bratman, M. 2007. Structures of Agency. Oxford: Oxford University Press. Culvert, C.M. and Gert, B. 2004. Competence. In Radden, J. (ed.) The Philosophy of Psychiatry: A Companion. Oxford: Oxford University Press; 258–270.

176

Pathologies of agency Dancy, J. 2000. Practical Reality. Oxford: Clarendon Press. Davidson, D. 2001. How Is Weakness of the Will Possible? In Davidson, D. (ed.) Essays on Actions and Events. Oxford: Clarendon Press; 21–42. Hieronymi, P. 2014. Reflection and Responsibility. Philosophy & Public Affairs 42: 3–41. Holton, R. 2009. Willing, Wanting, Waiting. Oxford: Oxford University Press. Hornsby, J. 2013. Basic Activity. Proceedings of the Aristotelian Society, Suppl. Vol. 87:1–18. Kennett, J. 2002. Autism, Empathy and Moral Agency. Philosophical Quarterly 52 (208): 340–357. Jamison, K. R. 1995. An Unquiet Mind. New York: Vintage Books Random House. Laplane, D. and Dubois, B. 2001. Auto-Activation Deficit: A Basal Ganglia Related Syndrome. Movement Disorders 16: 810–814. Levy, R. 2012. Apathy: A Pathology of Goal-Directed Behaviour. A New Concept of the Clinic and Pathophysiology of Apathy. Revue Neurologique 168: 585–597. Levy, R. 2015. Apathy: Anonymised Case-Studies. Neuroscience and Philosophy. Paris-Sorbonne Universités, 27 Nov. McKenna, M. 2012. Conversation and Responsibility. Oxford: Oxford University Press. Mele, A.R. 1987. Irrationality: An Essay on Akrasia, Self-Deception and Self-Control. New York: Oxford University Press. Moran, R. and Stone, M. 2009. Anscombe on the Expression of Intention. In Sandis, C. (ed.) New Essays on the Explanation of Action. London: Palgrave Macmillan, 132–168. Pacherie, E. 2007. The Anarchic Hand Syndrome and Utilization Behaviour: A Window onto Agentive Self-Awareness. Functional Neurology 22 (4): 211–217. Poland, J. and Graham, G. (eds.) 2011. Addiction and Responsibility. Cambridge, MA: MIT Press. Potter, N. N. 2013. Empathic Foundations of Clinical Knowledge. In Fulford, K.W.M., Davies, M. et al. (eds.) The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press; 293–306. Radden, J. 2012. Privacy and Patient Autonomy in Mental Healthcare. In Radoilska, L. (ed.) Autonomy and Mental Disorder. Oxford: Oxford University Press; 123–142. Radoilska, L. 2013a. Depression, Decisional Capacity, and Personal Autonomy. In Fulford, K.W.M., Davies, M. et al. (eds.) The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press, 1155–1170. Radoilska, L. 2013b. Addiction and Weakness of Will. Oxford: Oxford University Press. Radoilska, L. 2015. Autonomy in Psychiatric Ethics. In Sadler, J., van Staden, G.W. and Fulford, K.W.M. (eds.) The Oxford Handbook of Psychiatric Ethics. Oxford: Oxford University Press; 354–371. Schramme, T. (ed.) 2014. Being Amoral: Psychopathy and Moral Incapacity. Cambridge, MA: MIT Press. Schwitzgebel, E. and Ellis, J. 2017. Rationalization in Moral and Philosophical Thought. In Bonnefon, J.F. and Trémolière, B. (eds.) Moral Inferences. New York Routledge: 170–190. Setiya, K. 2008. Practical Knowledge. Ethics 118 (3): 388–409. Shoemaker, D. 2009.Responsibility and Disability. Metaphilosophy 40(3/4): 438–461. Shoemaker, D. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Smith, A. 2007. On Being Responsible and Holding Responsible. The Journal of Ethics 11: 465–484. Strawson, P. 1962. Freedom and Resentment. Proceedings of the British Academy 48: 1–25. Wallace, R.J. 1994. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Widdershoven, G.A.M. and Abma, T.A. 2012. Autonomy, Dialogue and Practical Rationality. In Radoilska, L. (ed.) Autonomy and Mental Disorder. Oxford: Oxford University Press; 217–232.

177

PART 4

Agency: mind, body, and world Introduction to Part 4 Luca Ferrero

Where is agency located? Where is agency primarily exercised and realized? As the distinctive capacity of agents, agency is located where the agents are. But the location of agents, it turns out, is not independent of the issue of the ‘realization’ of agency: of the places and modes in which agency is directly and primarily manifested and exercised. When we start reflecting on the nature of agency, it is quite common to offer as standard examples cases of ordinary bodily actions, such as walking or raising our arms. There is something intuitive about the suggestion that voluntary or intentional movements of our bodies are the ways in which we directly exercise our agency. For at least two reasons: first, we normally do not need to do anything else in order to move our bodies; second, even if our goals are quite distinct from our bodily movements, it seems that we can bring these goals about only by starting with some bodily movement of ours. These intuitively plausible observations can then be made part of a philosophical view about the realization of agency, a view according to which, to use Donald Davidson’s famous saying, “We never do more than move our bodies: the rest is up to nature” (Davidson D. 2001. Essays on Actions and Events. New York: Clarendon Press: 59.) However, the idea that our agential capacities, powers, and skills are primarily and directly bodily or corporeal can be challenged from two opposite sides. First, much of our mental life seems to be agential. Consider out talk about making choices, directing our attention or imagination, guiding our reasoning and deliberation, and so forth. Many mental episodes seem to have at least some of the distinctive hallmarks of agency: they can be said to be active, intentional, reason-responsive, self-conscious, the object of non- observational knowledge, and so forth. At the very least, it seems that there could be genuine mental agency that is independent of any bodily agency. But one might even push for a more radical claim: what if all of agency is primarily mental? Even when we consider bodily action, it seems that one could always fail to move one’s body due to a sudden and unknown paralysis. When so, it seems that one is still exercising one’s agency in initiating that (failed) movement: one initiates it by one’s decision, act of will, or ‘trying’—which seem to be purely mental phenomena. If so, one might conclude, paraphrasing Davidson, that all we do is ‘move’ our mind, the rest if left to nature.

DOI: 10.4324/9780429202131-20179

Luca Ferrero

On the opposite side, one might argue not only that agency is necessarily embodied but that agency is primarily realized in the physical (and social) world at large. The thought here is that the biologically given body is just a technology that helps with the realization of our ends. One might argue that many portions of the world function like our bodies, and in that sense, they extend what counts as the body to literally ‘incorporate’ some technologies, namely, those that are functionally equivalent to the biological body, starting with prosthetics. But an even more radical view is possible, one that inverts the relation: it is not that technologies can be incorporated in the body, but that the body is just an instance of technology. If so, agency should be characterized as the capacity to pursue our ends via the use of material and socially determined technologies in the world (which include but are not limited to natural and prosthetic bodies)—as Matthew Noah Smith argues in the last chapter of this section, “Material agency.”

Mental agency The question about the relationship between bodily and mental agency (and the philosophical relevance of this distinction) is the topic of ‘Mental agency’ by Matthew Soteriou. Soteriou opens by claiming that there does not seem to be an obvious and uncontroversial way to draw the distinction, and it is unclear whether the categories of mental and bodily action are exclusive. Soteriou then considers a specific proposal: what is distinctive of mental agency is the possibility of exercising it without performing any overt bodily action. This conception of mental agency has been used (for instance by Descartes) in arguments that try to establish the conceivability and thereby the metaphysical possibility of disembodied agents: agents who have a capacity for mental agency (as essential to them) while lacking any capacity for overt bodily agency. In response, anti-Cartesian views of the mind (such as behaviorism, some versions of functionalism, and interpretationism) argue that overt bodily agency is necessary to mentality, that there is a necessary connection between being the subject of mental states and being an agent capable of overt bodily action. These views raise the concern that a Cartesian understanding of mental agency induces a picture in which overt bodily actions must be conceived as the effects of distinct mental acts, which, in turn, might lead to an interiorization of all exercises of mental agency. Soteriou argues, however, that it might still be possible to think that the capacity for bodily agency is distinct from mental agency if it is conceivable that one might lose the former without losing the latter, as shown by the possibility of suffering a bodily paralysis without suffering mental paralysis. Thus, in the last portion of the chapter, Soteriou compares and contrasts the notions of bodily paralysis and mental paralysis to explore (a) whether there any significant epistemological and metaphysical differences between the exercise of mental agency and the exercise of bodily agency, and (b) the more general significance of the exercise of mental agency in our conscious mental lives.

Agency and the body The role of the body in agency is the topic of Hong Yu Wong’s ‘Agency and the body,’ in which he compares two pictures of the role of the body in bodily action. On the standard causal theory of action, the exercise of the capacity for bodily action consists in the acquisition of intentions, which, in turn, causes the intended bodily movements. Here the body is merely an enabling condition of bodily action, but it is not part of what makes such action an 180

Agency: mind, body, and world

exercise of agency. Agency rests in the operation of psychological attitudes; the body is only there to execute and implement. Wong defends an alternative picture, in which the body is partly constitutive of what makes bodily action agentive. We are embodied agents, who sense and act directly with our (animal) bodies. Support for this picture comes from considering pathologies of embodiment, which fundamentally alter the character of our agentive capacities. Wong presents in detail two such pathologies. First, deafferented agents, as a result of the loss of proprioception, lose a core capacity for embodied agency: the non-observational knowledge of their bodily actions (even with basic bodily actions, deafferented agents have to observe what is happening in order to know what they are doing). Second, patients who suffer from vertigo attacks have their egocentric perception and action seriously compromised: their core behavioral capacities for movement are disabled and immediate possibilities for action abolished. If the body were a mere enabling condition, these pathologies would only compromise the patients’ ability to act. But what we observe, instead, are fundamental alterations of the character of their agentive capacities. Wong then discusses several ways in which the body structures agency. First, the body as the effector: we act directly with our bodies, we perform teleologically basic actions with our bodies. That is, we do not normally need to target our body as we would do, instead, with an external tool. Second, no conscious attention to the bodily effectors is required in fluid everyday action, whereas they are required by deafferented patients. Although deafferented patients show that conscious proprioceptive awareness is not necessary for action, but Wong argues that the capacity for proprioception still shapes ordinary bodily agency. Finally, there is a complex form of bodily awareness that is the most vital aspect of embodiment: the sense of body ownership. This sense seems to be a condition—what Wong calls the ‘ownership constraint’—on acting directly with any parts of our body and having a sense of agency over that body when acting with it.

Agency, consciousness, and attention A nature place to raise questions about the intersection between bodily and mental agency is the role of consciousness and attention in our bodily agency. This role is the topic of Wayne Wu’s chapter on ‘Agency, consciousness, and attention.’ This chapter focuses on visually guided bodily action, which is one of the areas where much detailed empirical research has been conducted. Wu argues in support of the commonsense picture that conscious vision guides action and contributes to the phenomenology of visually guided agency. Wu begins by distinguishing pure reflexive behavior, where the outcome is entirely determined by the input, and genuine action, where the input is mapped in a multiplicity of potential behavior outcomes. In action, the task is thus to select one path among the many possible ones (what Wu calls the “Many-Many” problem). Not everything in action is under the agent’s control, however. Most features of action are actually automatic. In intentional action, then, the intention specifies which features of the action are controlled and which ones are automatic. Attention is the process by which the mind takes possession of one of the main possible targets in order to deal effectively with it, while withdrawing from the other ones. Wu then considers the specific operations of visually guided action, surveying recent empirical work. Much of this work suggests that attention restricts phenomenal consciousness (by having a kind of tunnel vision fixed by visual attention) and that some attention in visually guided action is unconscious. Contrary to these views, the author argues that visual attention fixes 181

 Luca Ferrero

visual phenomenology but does not exhaust it. Hence, our phenomenology overflows both attention and access. Wu also disputes the evidence in favor of actions guided by unconscious visual states (so-called ‘zombie actions’). He defends a commonsense view in which visual attention provides an anchor for visual experience in action but without limiting experience to what one attends. He then concludes by suggesting that a similar structure might be present in other perceptual modalities that feed into bodily action.

Material agency The effects of a specific exercise of agency can in principle extend as far as the causal chain initiated by the action extends. But how far does the realization of agency extend? Historically, the realization of agency has been located either in the mind or in the agent’s body. In ‘Material agency,’ Matthew Noah Smith argues instead that human agency is realized in the world: agency, he argues, is to be conceived primarily as the capacity to use material and socially determined technologies (including the biologically given body) for the pursuit of ends. In the philosophical literature, attention is usually given only to a narrow menu of agential capacities, from which one might conclude that all we ever do when we act is to move our bodies. But careful observation reveals that almost all of our capacities are realized via material technologies. It is not just that technologies are functionally equivalent to the biological body. Smith argues for the far more radical claim that there is no privileged material for the realization of human agency. According to Smith, it follows that there are no general practical abilities, only specific ones. Abilities are specified by the contexts of their realizations. This can be seen by considering that some inabilities are transient, whereas others are lasting. Some inabilities are due to temporary losses in cognitive capacities or technological deficiencies. But there are never general capacities to act; there are no generic capacities to perform an act-type. Act-types are neither natural kinds nor social kinds, but abstracta. Ascriptions of agential capacities are contextually dependent and usually presuppose standards for their normal realizations. Each human capacity is realized in the technological moment in which the agent lives, conditional on the technologies to which she has access. Human agency is not the capacity to intentionally move the body but the capacity to use specific material (including the natural body, the prosthetic body, and non-bodily technologies) for the sake of some end. Smith concludes that his account of agency is attractive for at least three reasons. First, its plausibility does not rest on the resolution of debates about the nature of basic actions. Second, individuating capacities by technologies allows for a socio-historically precise and flexible characterization of the heterogeneity of human capabilities. Third, his account highlights how all agential capacities are realized in systems and that agency is not naturally occurring but socially produced. The understanding of human agency should not be individualistic and body-focused but rather social and world-focused. A conclusion that, as Smith writes, “bring philosophy of action directly into contact with political theory. As it should be.”

182

16 MENTAL AGENCY Matthew Soteriou

Philosophical discussions of action typically focus on the agency we exercise by moving our bodies. However, a significant amount of our agential behaviour appears not to fit straightforwardly into that category. For instance, we can choose what to consider, what to imagine, and we can exercise agency over the direction taken by our thinking when we deliberate about what to do or when we try to work out some puzzle. Such examples indicate that some of the events that constitute our conscious mental lives can amount to mental behaviour over which we exercise agency, which we can exercise without moving our bodies. Hence, it might be said that we don’t just have a capacity for bodily agency but also for mental agency. One response to that way of categorising our agential capacities is to say that while it isn’t objectionable in itself, we should not be misled into thinking that it indicates anything of philosophical significance. If a subject has the capacity to exercise agency over her arms as well as her legs, it might be said that she has a capacity for ‘arm agency’ in addition to having a capacity for ‘leg agency’, but we tend not to think that sort of distinction between capacities for agency gives rise to interesting philosophical questions. So is there any good reason to think that there is a distinction between mental and bodily agency that tracks, or gives rise to, deeper philosophical concerns?1 In what follows I shall explain why some have thought there is. I will consider the following questions: What is distinctive of our capacity for mental agency? How does it compare with our capacity for bodily agency? What is the relation between mental agency and bodily agency? How might answers to such questions add to our understanding of issues in the metaphysics and epistemology of mind, and the philosophy of action?

1 Drawing a clear and stable distinction between mental agency and bodily agency is not straightforward. One might try to mark a distinction by appealing to the idea that body agency involves exercising agency over parts of one’s body in a way that mental agency does not. However, there are difficulties with this suggestion; the exercise of mental agency involves brain events, and the brain is part of one’s body. One might instead try to distinguish the two agencies by attempting to say what is distinctive of mental rather than bodily action. However, specifying what makes an action a mental one is also not a straightforward matter: DOI: 10.4324/9780429202131-21183

Matthew Soteriou

there seems to be no obvious and uncontroversial way of classifying every action as either mental or bodily, and it is unclear whether these categories are exclusive. For example, if one is thinking out loud or working something out on paper, it could be argued that one is thereby performing an action that is both bodily and mental. Furthermore, some physicalists may suggest that if all mental events are in fact identical with bodily events, we should dispense with the assumption that there is a distinction between mental action and bodily action, for if all mental events are bodily events, then all so-called mental actions are really bodily actions, insofar as they are bodily events that are intentional under some description.2 A response worth exploring is this: What is distinctive of our capacity for mental agency is that we can exercise it without performing any overt bodily action. Note that this isn’t the proposal that there aren’t any exercises of our capacity for mental agency that involve overt bodily actions. It is, rather, the modal claim that it is possible to exercise one’s capacity for mental agency without performing an overt bodily action. This proposal obviously depends on a cogent distinction between overt and covert actions. So how should that distinction be drawn? Epistemological considerations are relevant to the distinction, for, roughly speaking, overt actions are those actions for which a certain kind of third-person epistemic access is possible. One might try to capture the relevant notion of access by saying that overt actions are observable and covert actions are not. However, for a physicalist who holds that all mental events are identical with bodily events, that may simply amount to the matter of how difficult it happens to be for other subjects to observe the bodily events that constitute the agent’s action, and as Yair Levy (2019) notes, a problem with this way of marking a distinction is that it appears to depend on a contingent feature of those actions, given that technological advances can allow us to perceive bodily events that were previously inaccessible. In response, one might offer a more refined characterisation of the relevant kind of third-person epistemic access. Suppose that we have the brain imaging technology to observe the neural events that occur when a subject is exercising agency over her thinking—for example, when she is attempting to solve some arithmetical problem. Although we may be able to observe the bodily events that constitute that exercise of her agency, there is a respect in which that doesn’t thereby provide us with epistemic access to her action. For we are likely to be none the wiser when it comes to the question of what she is thinking about and what her aim is in so thinking; this might just be because the mental events that constitute such actions cannot be observed as such. One might still come to know what a subject is doing when they are exercising their mental agency in this way. One can, for example, be told by the agent. The overt action the agent performs in communicating that information can be a route to knowing which mental events are tokened by the brain events one perceives. But the obstacle to knowing what an agent is thinking independently of her performance of overt bodily actions is that one cannot perceive as such the mental events that constitute her thinking. Sometimes one can tell that someone is exercising their mental agency covertly without being told by that agent, and one may know someone so well that one can make a good guess as to what they are thinking. However, in such cases, one’s knowledge and conjectures on such matters cannot be grounded by one’s perception of the mental events that constitute their mental behaviour, insofar as one cannot observe as such the mental events that constitute their conscious thinking. These points about the third-person epistemic access that one might have to an agent’s overt actions, in contrast to their covert mental actions, are relevant to a distinctive feature of the first-person epistemic access that an agent has to her own covert mental actions. The respect in which overt actions are observable makes the belief that one is performing an overt 184

Mental agency

action potentially susceptible to perceptual counterevidence. If Φ-ing is some overt action, then one’s belief that one is Φ-ing is something that perceptual evidence can speak against. This enables us to make sense of the possibility of there being evidence that speaks against your belief that you are Φ-ing but which is inaccessible to you because your senses fail to provide you with that counterevidence. It is not clear that one’s belief that one is performing a covert mental action is susceptible to the same kind of perceptual counterevidence. For instance, if you believe that you are working out an arithmetical problem in your head and you are told that the brain imaging reveals that you cannot be working out an arithmetical problem, you are likely to treat that as evidence that undermines the neuroscientific theory on which that pronouncement was made, rather than counterevidence against your belief that you are calculating. In light of these points, let us return to the proposal that what is distinctive of our capacity for mental agency is that we can exercise it without performing any overt bodily action. What issues of philosophical concern does it raise?

2 This proposal about what is distinctive of our capacity for mental agency appeals to a modal claim that invokes epistemological considerations. Claims of that form are often used in arguments that appeal to what is conceivable to ground claims about what is metaphysically possible.3 In this case, such an argument might attempt to establish that it is conceivable that one possess a capacity for mental agency without possessing a capacity for overt bodily agency, and then on that basis attempt to establish that it is metaphysically possible for one to possess a capacity for mental agency without possessing a capacity for overt bodily action. One finds something like this line of thought in Descartes’ Meditations. There Descartes attempts to argue for the conceivability and metaphysical possibility of disembodied existence. In his attempt to protect the coherence, and hence conceivability, of disembodied existence, Descartes is concerned to protect the coherence of the notion of a disembodied agent—an agent capable of action but incapable of bodily action. On one understanding of Descartes’ strategy, he seeks to show that sceptical arguments that target the putative knowledge we acquire via the senses can be used to undermine our putative knowledge of our own bodily actions, and in a way in which they cannot be used to undermine our putative knowledge of our own mental actions. He exploits this line of thought in support of the proposal that it is conceivable for one to exist as a subject capable of exercising mental agency without having a capacity for bodily agency. A further concern of Descartes’ Meditations is to argue, in first-person guise, that the capacity for mental agency is essential to me, whereas the capacity for bodily agency isn’t, and to appeal to that idea in support of substance dualism. However, we can envisage arguments that are structurally similar to Descartes’, but which eschew any commitment to substance dualism, and which may be consistent with some version of physicalism. For instance, a physicalist might try to exploit sceptical arguments that appeal to the possibility of being a brain-in-a-vat in order to support the claim that it is conceivable, and hence metaphysically possible, for one to exist as a subject capable of exercising mental agency without having a capacity for overt bodily action. What one makes of such arguments will depend, in part, on the stances one takes on the following questions: what are the relations between mind and behaviour, mind and brain, and mind and body? And one’s answers won’t simply be settled by whether one is committed to physicalism, for there are versions of physicalism that can be more, or less, ‘Cartesian’. 185

Matthew Soteriou

Descartes’ attempt to argue for the conceivability of disembodied existence respects the idea that there is a non-contingent connection between mind and behaviour, for it is consistent with the idea that there is a necessary connection between being the bearer of mental states and being an agent of action. However, Descartes’ position is inconsistent with the idea that there is a necessary connection between being a bearer of mental states and being an agent capable of overt bodily action. Certain anti-Cartesian views of mind seek to establish such a necessary connection. This disagreement between advocates of Cartesian and anti-Cartesian views won’t simply be settled by the question of whether physicalism is correct.4 However, this dispute is relevant to the question of the relation between mind and brain, and the question of the relation between mind and body, on the assumption that one needs a body, and not just a brain, in order to engage in overt bodily action. There are a variety of different anti-Cartesian views of the mind—for example, behaviourism, some versions of functionalism, and interpretationism. These views make overt bodily agency central to their understanding of mentality, for an idea common to them is that it is a mistake to regard as merely contingent the connections between one’s mental life and its publicly observable manifestations in overt action. Some of those arguing against the Cartesian view raise the concern that the Cartesian stance invites a flawed conception of the relation between exercises of mental agency and exercises of bodily agency. In particular, the concern is that the Cartesian view invites the idea that one’s overt bodily actions are to be understood as bodily effects that are preceded by, and caused by, distinct mental acts that are exercises of one’s mental agency. According to one version of the target view, what makes one’s overt bodily actions voluntary is ultimately to be explained by the fact that they involve bodily events that are preceded by, and caused by, exercises of one’s mental agency, such as volitions. An objection to that approach is that it cannot explain what makes the relevant exercises of mental agency voluntary. A further objection is that the view leads to the interiorisation of all exercises of mental agency, insofar as it suggests that exercises of mental agency are those occurrences that would be the remainder if we were to subtract any overt bodily events that are their effects. Some think that that conception of mental agency is unsatisfactory, because it fails to accommodate the idea that some exercises of mental agency are themselves overt bodily actions. For example, it leads to the view that the agency one exercises in thinking out loud is to be factored into two distinct components: (a) one’s thinking, which is covert, and (b) the publicly observable bodily movements which are the merely separate effects of one’s thinking.5 However, not all anti-Cartesian views need call into question the distinction between mental and bodily agency. Suppose one grants that it is a mistake to regard as merely contingent the connections between one’s mental life and publicly observable manifestations of one’s mental life in overt action. And suppose too one grants that some mental actions are overt bodily actions. Still, one might think that our capacity for mental agency is not identical to our capacity for bodily agency, insofar as one can lose the one without losing the other, as indicated by the idea that one can suffer bodily paralysis without suffering mental paralysis. The notion that it is possible to suffer bodily paralysis without suffering mental paralysis is connected with the proposal that what is distinctive of our capacity for mental agency is that we can exercise it without performing any overt bodily action. Later, I shall compare the notions of bodily and mental paralysis to explore further whether there are any significant epistemological and metaphysical differences between the exercise of mental and bodily agency.

186

Mental agency

3 As previously mentioned, the third-person epistemic access that one might have to an agent’s overt bodily actions, in contrast to her covert mental actions, might lead one to think that there are differences in the first-person epistemic access one has to her own action in each case. Recall that the belief that one is performing an overt action is potentially susceptible to a perceptual counterevidence that may be inaccessible to her. That susceptibility allows us to envisage the following case of bodily paralysis: One attempts to Φ (where Φ-ing is some overt bodily action), one seems to be Φ-ing but no Φ-ing occurs, because the requisite overt bodily movement does not occur. This may suggest that there might be a potential for a gap between the obtaining of the sort of mental state that is involved in practical self-awareness (i.e. the one involved in one’s awareness of what one is doing) and the exercise of agency over one’s body. That, in turn, may suggest that in the exercise of bodily agency, the mental state involved in one’s practical self-awareness does not constitutively depend for its obtaining on the bodily events required for bodily agency. It is often suggested that when one knows what one is doing when performing an action, one has a form of practical self-awareness that isn’t perceptual. Some hold that this practical self-awareness is embodied in an intention-in-action.6 Given the above points, one might think that in the case of an exercise of overt bodily agency, the obtaining of any such intention-in-action cannot be constitutively dependent on the exercise of agency over one’s body. Either there is dependence in the other direction, so the bodily movements that are required for an exercise of bodily agency depend causally on that intention-in-action, or the intention-in-action and the bodily movements are joint effects of a common cause.7 If one adds to this the idea that any such bodily movements can in principle occur without being an exercise of agency, one might then be led to suppose that what makes a bodily movement a bodily action is something ultimately determined by the causal pedigree of that bodily movement—for example, whether it is caused by a prior intention or an intention-in action.8 When we turn to mental agency, it is not obvious that the same considerations apply, because there might not be an analogous case of mental paralysis, that is, a case in which the following is true: One attempts to Φ (where Φ-ing is some covert mental action), one seems to be Φ-ing but no Φ-ing occurs, because the requisite mental events do not occur. A factor relevant to the intuition that the situations are not analogous is that in the case of covert mental action, one’s belief that one is exercising one’s agency is not susceptible to the same kind of counterevidence, that is, perceptual evidence that is inaccessible to one. As a result, in the case of an exercise of mental agency, it is not so obvious that there is the potential for a gap to open up between the obtaining of the sort of mental state that is involved in practical self-awareness and the exercise of one’s mental agency. That is, in the case of an exercise of mental agency, we may not have the same reason to deny that the sort of mental state involved in one’s practical self-awareness constitutively depends for its obtaining on the occurrence of the mental events required for an exercise of mental agency. 187

Matthew Soteriou

One consequence of such considerations is this: in mental agency, we may not have the same reason to hold that an intention-in-action can obtain in the absence of any exercise of one’s mental agency. This may lead us to question whether we are obliged to regard the mental events involved in an exercise of mental agency as the causal products of a metaphysically distinct mental state—the intention-in-action. That, in turn, may lead us to question whether we should accept that what makes the occurrence of mental events an exercise of one’s mental agency is something ultimately determined by the causal pedigree of those mental events, for example, whether they are caused by a prior intention or an intention-in action.9 These points may be relevant to debates about the scope of our mental agency.10 According to one line of thought, thinking about something involves the occurrence of mental events individuated, in part, by their propositional contents, and these mental events can be mental actions only if the particular contents that individuate them are ones that one intends to think. However, in the case of many such mental events, it seems that the content of the mental event cannot figure in the content of one’s prior intention. Strawson (2003) has argued that no thinking of a particular thought-content is ever an action. According to Strawson, “Mental action in thinking is restricted to the fostering of conditions hospitable to contents coming to mind” (2003: 234). This is because one’s thinking of a particular content can only amount to an action if the content is already there, “available for consideration and adoption for intentional production”, in which case, “it must already have ‘just come’ at some previous time in order to be so available” (2003: 235). This line of thought may be undermined if we have reason to reject the aforementioned ‘causal pedigree’ view of mental agency, that is, the view according to which the fact that the occurrence of mental events amounts to an exercise of one’s mental agency is something that is ultimately determined by the causal pedigree of those mental events (e.g. whether they are caused by a prior intention or an intention-in action).11 An alternative to that ‘causal pedigree’ view is that to exercise one’s mental agency is to engage in mental behaviour that realises practical self-awareness—knowledge of one’s aim in so acting. On such a view it could be said that exercising mental agency (e.g. exercising agency over the occupation of conscious attention and the direction taken by one’s thinking) is exercising a capacity for a certain kind of self-knowledge, for example, practical awareness of one’ s aim in so acting.12 Let us return to the earlier example of bodily paralysis: one attempts to Φ (where Φ-ing is some overt bodily action), one seems to be Φ-ing, but no Φ-ing occurs, because the requisite overt bodily movement does not occur. Some have been tempted to hold that in this case, even though one’s intention-in-action cannot be constitutively dependent on the kind of bodily movement that is required for an exercise of agency over one’s body, its obtaining is nonetheless constitutively dependent on the occurrence of some kind of active event— namely, one’s attempt to move one’s body; for even though one doesn’t know that one is moving one’s body, one at least knows that one is trying to do so.13 That, in turn, may lead to the view that the epistemic residue of one’s failed attempt to perform the bodily action must be knowledge of an active mental event, for example, a volition, or act of will. One version of that view might lead to a conception of the relation between exercises of mental agency and exercises of bodily agency that I previously said has been criticised by some of those who advocate an anti-Cartesian view of action. There are several possible objections and alternatives to that conception of the relation between exercises of mental agency and exercises of bodily agency, which I don’t have the space to discuss here. But the point I want to emphasise is that these are lines of thought 188

Mental agency

that can be generated by the idea that it is possible to suffer bodily paralysis without suffering mental paralysis, and that the idea that is connected with the proposal that what is distinctive of our capacity to exercise mental agency is that we can exercise that capacity without performing any overt bodily action. So that way of drawing a distinction between mental and bodily agency can potentially give rise to significant philosophical questions and concerns. A consideration of the differences between bodily and mental paralysis may also be potentially helpful in trying to uncover the more general significance of the exercise of mental agency in our conscious mental lives. That is a line of thought that has been developed in detail by Brian O’Shaughnessy (2000). One of O’Shaughnessy’s central concerns is “to discover the contribution made to consciousness, not by the bodily will (for we can be fully conscious though supine in a hammock, and even if totally paralysed), but by the mental will” (2000: 226). In that work, O’Shaughnessy attempts to argue that if one were to suffer mental paralysis, one would cease to be awake. If O’Shaughnessy is right, then we can envisage what it would be like for one to undergo conscious mental events while suffering mental paralysis. One would be dreaming. And so if O’Shaughnessy is right, the capacity to exercise agency over our mental lives may turn out to be of central importance to an account of consciousness, insofar as it is of central importance to an account of what it is to be conscious in the sense of being awake. Whether or not O’Shaughnessy’s proposal is ultimately correct, there appears to be ground for thinking that the idea that we have a capacity for mental agency distinct from our capacity for bodily agency gives rise to philosophical questions and concerns that are far deeper than those raised by, for instance, a distinction between ‘arm agency’ and ‘leg agency’. The answers we give to questions about the nature of mental agency and its relation to bodily agency have significant implications for our understanding of some central issues in the metaphysics of mind, the epistemology of mind, and the philosophy of action.

Related entries Agency and the body; Agency, consciousness, and attention; Epistemic agency; Agency and self-knowledge.

Notes 1 For scepticism about the explanatory value of trying to draw a distinction between mental action and bodily action, see Levy (2019). 2 Compare Davidson’s (1971) discussion of cerebral events involved in action. 3 For various discussions of arguments of this form, see Gendler and Hawthorne (2002). 4 For discussion of these ideas, see Shoemaker (1976). 5 Both of these lines of objection play a significant role in Ryle’s (1949) critique of the Cartesian view. 6 For example, see Anscombe (2000), Velleman (1989), and Falvey (2000). 7 For the latter view, see Peacocke (2009). 8 For examples of causalist approaches to action, see Davidson (1963), Goldman (1970), Searle (1983), Brand (1984), and Bishop (1989). For objections to such approaches, see, for example, Haddock (2005), Steward (2012), Lavin (2013), and Levy (2016). 9 Compare Levy (2016), where Levy appeals to certain examples of mental agency when objecting to causalist views of action. 10 For example, there is considerable disagreement as to whether our judgements and decisions can be regarded as mental actions. For discussions of such debates, see, for example, Strawson (2003), Mele (2009), Dorsch (2009), Gibbons (2009), Hieronymi (2009), and Pink (2009).

189

Matthew Soteriou 11 For discussion and criticism of Strawson’s argument, see Buckareff (2005), Proust (2013), Wu (2013), and Levy (2016; 2019), and Upton and Brent (2018). 12 This view is developed in Soteriou (2013). 13 For discussion of the role of trying in an account of the metaphysics of action, see, for example, O’Shaughnessy (1973), Hornsby (1997), essays 5 and 6, McCann (1975), and Grünbaum (2008).

Further reading Strawson, Galen (2003). ‘Mental Ballistics or the Involuntariness of Spontaneity’, Proceedings of the Aristotelian Society 103(3): 227–257. This paper argues that the role of mental action in our thinking can at best be indirect. It has been very influential in discussions of mental action, and many have since proposed alternative conceptions of mental agency, which in one way or other try to challenge Strawson’s argument. L. O’Brien and M. Soteriou (eds), (2009). Mental Actions. Oxford: Oxford University Press. This volume contains 12 specially written papers covering such topics as the scope of mental action, the epistemology of mental action, the phenomenology of mental action, the role of mental agency in perception, and the relationship between the voluntary and the active in the mental sphere. O’Shaughnessy, B. (2000). Consciousness and the World. Oxford: Oxford University Press. This book contains a sophisticated and detailed exploration of the role that mental agency plays in our conscious mental lives. It includes proposals about the connections between the exercise of mental agency and self-knowledge, our temporal awareness and states of consciousness, and it also discusses the different structures that mental agency can take. Levy, Y. (2019), ‘What Is Mental Action?’, Philosophical Psychology 32(6): 969–991. This paper questions the theoretical significance of distinguishing between mental and bodily agencies, and discusses the proposal that a distinction between covert and overt action should supplant the one between mental and bodily actions.

References Bishop, J. (1989). Natural Agency: An Essay on the Causal Theory of Action, Cambridge: Cambridge University Press. Brand, M. (1984). Intending and Acting, Cambridge, MA: MIT Press. Buckareff, A. (2005). ‘How (Not) to Think about Mental Action’, Philosophical Explorations: An International Journal for the Philosophy of Mind and Action 8(1): 83–9. Davidson, D. (1963). ‘Actions, Reasons, and Causes’, Journal of Philosophy 60(23): 685–700. Davidson, D. (1971). ‘Agency’, in his Essays on Actions and Events, Oxford: Oxford University Press, 1980. Descartes, R. (1996). Meditations on First Philosophy, trans. by J. Cottingham, Cambridge: Cambridge University Press. Dorsch, F. (2009), ‘Judging and the Scope of Mental Agency’, in L. O’Brien and M. Soteriou (eds), Mental Actions, Oxford: Oxford University Press. Gibbons, J. (2009). ‘Reason in Action’, in L. O’Brien and M. Soteriou (eds), Mental Actions, Oxford: Oxford University Press. Goldman, A. (1970). A Theory of Human Action. Englewood Cliffs, NJ: Prentice-Hall. Grünbaum, T. (2008). ‘Trying and the Arguments from Total Failure’, Philosophia 36: 67–86. Haddock, A. (2005). ‘At One with Our Actions But at Two with Our Bodies: Hornsby’s Account of Action’, Philosophical Explorations 8(2): 157–172. Hieronymi, P. (2009). ‘Two Kinds of Agency’, in L. O’Brien and M. Soteriou (eds), Mental Actions. Oxford: Oxford University Press. Hornsby, J. (1997). Simple Mindedness: In Defence of Naïve Naturalism in the Philosophy of Mind, Cambridge, MA: Harvard University Press. Lavin, D. (2013). ‘Must There be Basic Action?’, Noûs 47: 273–301. Levy, Y. (2016). ‘Action Unified’, Philosophical Quarterly 66 (262): 65–83. Levy, Y. (2019). ‘What Is Mental Action?’, Philosophical Psychology 32(6): 969–991 McCann, H. J. (1975). ‘Trying, Paralysis, and Volition’, in H.J. McCann (ed.), The Works of Agency: On Human Action, Will and Freedom (pp. 94–109). Ithaca, NY: Cornell University Press.

190

Mental agency Mele, A. (2009). ‘Mental Action: A Case Study’, in L. O’Brien and M. Soteriou (eds), Mental Actions. Oxford: Oxford University Press. O’Shaughnessy, B. (1973). ‘Trying (as the Mental “pineal gland”)’, The Journal of Philosophy 70(13): 365–386. O’Shaughnessy, B. (2000). Consciousness and the World, Oxford: Oxford University Press. Peacocke, C. (2009). ‘Mental Action and Self-Awareness (II): Epistemology, in L. O’Brien and M. Soteriou (eds), Mental Actions. Oxford: Oxford University Press. Pink, T. (2009). ‘Reason, Voluntariness and Moral Responsibility’, in L. O’Brien and M. Soteriou (eds), Mental Actions. Oxford: Oxford University Press. Proust, J. (2013), ‘Mental Acts as Natural Kinds’, in Till Vierkant, Julian Kieverstein and Andy Clark (eds), Decomposing the Will (pp. 262–282). Oxford: Oxford University Press. Ryle, G. (1949), The Concept of Mind, London: Hutchinson. Searle, J. (1983). Intentionality, Cambridge: Cambridge University Press. Soteriou, M. (2013). The Mind’s Construction: The Ontology of Mind and Mental Action, Oxford: Oxford University Press. Steward, H. (2012). A Metaphysics for Freedom, Oxford: Oxford University Press. Strawson, Galen (2003). ‘Mental Ballistics or the Involuntariness of Spontaneity’, Proceedings of the Aristotelian Society 103/3: 227–257. Upton, C. and Brent, M. (2018). ‘Meditation and the Scope of Mental Action’, Philosophical Psychology 32(1): 52–71 (2019) Wu, W. (2013). ‘Mental Action and the Threat of Automaticity’, in Andy Clark, Julian Kiverstein and Tillman Vierkant (eds), Decomposing the Will (pp. 244–261). Oxford: Oxford University Press.

191

17 AGENCY AND THE BODY Hong Yu Wong

1 Two pictures of the body in bodily action We are rational self-movers. The idea of self-movement is central to our understanding of agency. What do we move? We move our bodies. In doing so, we are moving ourselves. Even though embodied creatures, like ourselves, are capable of purely mental agency, which does not involve moving one’s body, our paradigm of agency remains bodily action. But just what is the role of the body in bodily action? This is not a question that is often addressed in the philosophy of agency, since philosophers have been fixated on the conative and motivational springs of agency. Orthodox action theory begins with the question: what is left over when I ‘subtract’ my hand’s going up from my raising my hand? (e.g. Velleman 2000). The thought is that the structures which remain are those that make for agency – and that this is what matters for understanding agency. Is this correct? I will argue that orthodox picture cannot make sense of the role of body in action and that an account of embodied agency is called for. Let us consider two pictures of the body in bodily action: I

Body as an enabling condition for agency. On the standard picture, the body is merely an enabling condition for bodily action. That is, the body doesn’t enter into the nature of what makes bodily action agentive. Though this is not explicit in the work of most orthodox action theorists, it is a consequence of the way they conceive of agency. On the standard causal theory of action, the exercise of one’s capacity for bodily action consists in the formation of intentions, which then bring about the intended bodily movements (e.g. Davidson 1980; Searle 1983; Velleman 2000). There is nothing more to agency than the appropriate causation of bodily movements by prior psychological states. Guidance and control of action consists, then, entirely in terms of the control one has over one’s attitudes. The body is there just for execution. (This is often allied with the thought that action execution is of no philosophical interest, something best left for the psychologists or physiologists.) It is no wonder that standard causal theorists consider everything bodily – motor control, perception, a functioning body, and all that – as simply enabling conditions for action.

DOI: 10.4324/9780429202131-22192

Agency and the body

II

Body as structuring agency. On an alternative picture, the body is partly constitutive of what makes bodily action the kind of action it is. It is in part because of the kind of body we have – and because we sense with and through our bodies and act directly with this body we sense that our bodily action has the agentive character it has (O’Shaughnessy 2008; Wong 2018). This is because we are not only agents but also embodied agents. The picture of bodily action that comes into view is one of a fundamental animal capacity (Steward 2009). We can then see the human capacity for rational agency as built on this foundation of embodied animal agency.

Which picture of the role of the body in bodily action is preferable? Seeing embodiment as an enabling condition for agency fails to do justice to human agency as we know it. Why? When we consider cases of subjects with problems with embodiment – such as numbness, chronic pain, deafferentation (loss of proprioception and touch), somatoparaphrenia (disturbances in the sense of ownership of one’s body), or vestibular disturbances (problems with the sense of balance) – these deficits drastically curtail the capacity for bodily action and fundamentally alter the character of their agentive capacities. Consider three examples. The first two come from a loss of proprioception, whilst the third comes from vestibular disturbance.

1.1 Evidence for the body as structuring agency A

B

Deafferentation and the automaticity of everyday agency. Peripherally deafferented subjects have no proprioception and touch in those parts of their bodies where their afferent nerves have been destroyed. Most such patients are severely impaired in their ability to act. However, in a few exceptional cases, despite loss of proprioception and touch in much of their body, patients relearn how to act. Deafferented agents use visual attention to substitute for proprioception and touch (Cole 1992; Cole and Paillard 1995 and Cole 2016). If deprived of vision, subjects may find it impossible even to stand. However, deafferented agency lacks the fluidity we take for granted in ordinary bodily action. Their bodily action requires considerable attention and degrades dramatically if attentional resources are depleted (Ingram et al. 2000). Deafferented agency demands constant visual vigilance even for ordinary everyday activities. Because deafferented visual attentional control is deliberate and effortful, it lacks the spontaneous automaticity of afferented agency. The contrast here is not merely between clumsy and fluid action, but about the automatic character of the core of everyday agency. This would appear to show that deafferented agency differs in its nature from afferented or embodied agency. Deafferentation and non-observational knowledge of action. Anscombe stressed that it is key to our understanding of intentional action that agents know what they are doing when they are acting intentionally without observing themselves (Anscombe 1963). Whilst Anscombe’s claim about non-observational knowledge of action is problematic as a claim about all intentional agency, the observation that agents normally know what they are doing without observing themselves when they are intentionally performing basic bodily actions seems to be correct. In the ensuing discussion of Anscombe’s claim, there has been much controversy about what ‘observation’ means. We do not need to settle this. For our purposes, it is enough to lean on the observation that in doing things that we know how to do and can perform well, in normal circumstances at least, it does not appear that we need to attend to what is happening as we are acting and it does not appear that we need to wait and see in order to know what we are doing. Arguably,

193

Hong Yu Wong

C

knowing what one is doing with one’s body without observation is a core aspect of the capacity for embodied agency – at least in central cases (O’Brien 2007). However, this is not the case with deafferented agents, who, even with basic bodily actions, have to observe what is happening in order to know what they are doing (Cole 1992; Balslev et al. 2007; and Cole 2016). The contrast here is not simply between actions controlled by different sources of sensory feedback, but an alteration in the character of the core of basic everyday striving: whether knowledge of what one is doing is non-observational. Once again, this would appear to support the claim that deafferented agency differs in its nature from embodied agency (Wong 2015 and 2018). Vestibular disturbance and egocentric perception and action. Our sense of balance is underpinned by the vestibular system, which provides a sense of which way is up and a sense of self-motion. The vestibular system tracks one’s orientation in the ambient gravitational field (which is a relational property of one’s body). What happens when we have vestibular problems? There are a wide range of vestibular diseases, with varying symptoms, depending on the underlying pathology. Symptoms include dizziness, vertigo, disorientation, feeling imbalanced, blurry or double vision, and behavioural consequences such as loss of balance, loss of postural control, nausea, vomiting, and falls. I want to focus on vertigo because it illustrates the foundational character of the vestibular system. Vertigo is characterised clinically as sudden episodes involving “an illusion of movement, usually that of rotation, although patients occasionally describe a sensation of linear displacement or tilt” (Baloh et al. 2011). During bouts of vertigo, the sensations of whirling and imbalance are exacerbated by movement. And if patients try to get up and move about, head movements may result in them falling. Egocentric perception and action are compromised. Patients very quickly learn to stay motionless during vertigo attacks. In suffering vertigo, one lacks a sense of orientation, and the consequent confusion cripples one’s ability to act. Core behavioural capacities for movement are disabled and immediate possibilities for action are abolished. In this way, vestibular malfunction brings the agent to a standstill; the possibility of egocentric perception and action is jeopardised. Once again, a deficit in bodily awareness appears to alter the character of the capacity for bodily action. I have argued elsewhere that the foundational role of vestibular functioning for egocentric perception and action is best explained by the vestibular system providing a master frame of reference that coordinates all spatial frames of reference (Wong 2017a, building on Berthoz 1991 and 2000). If this is correct, then the vestibular system even provides a condition of possibility for action because the master frame of reference makes egocentric spatial perception and action possible.

These three cases are not idiosyncratic. Further examples can be found when we consider other pathologies of embodiment. How do cases like these bear on the choice between the two pictures of the role of the body in action? If the body were an enabling condition, we would expect that subjects would be unable to act if embodiment were compromised, since it is an enabling condition. But we would not expect fundamental alterations of the character of agentive capacities. However, this is what we find in these cases of pathologies of embodiment. The nature of the agentive capacity has been fundamentally changed. This is not bodily agency as we know it. Thus, the first picture of the body as an enabling condition for agency cannot capture how significant embodiment is for shaping agency as we know it.

194

Agency and the body

2 How is embodied agency possible? How does the body structure agency? There are two ways in which embodiment is significant (Wong 2018). First, the body is the effector – it is what we act through. Second, in acting, we need to be aware of the effector. Let us take each point in turn.

2.1 The body as the effector through which we act In recognising the body as the effector, we are recognising it as the primary thing we are moving when we act. This has both functional and phenomenological consequences. In acting with one’s body one doesn’t act with it by acting with something else. Rather one has the capacity to act directly – to perform teleologically basic actions (Hornsby 1980) – with one’s body. For example, one can directly raise one’s right arm, rather than having to raise it using one’s left arm. Furthermore, in exercising the fundamental capacity of acting with one’s body, we don’t have to target one’s body first in order to act directly with it (pace O’Shaughnessy 2008). (Of course, the body is sometimes a target object, as when I scratch an itch.) We can bring this out by thinking of the contrast between acting with our bodies only and acting with tools with which we are very familiar, such as a pen or a fork. Ordinary bodily action in both cases can be skilled and fluid, but with tools, one always has to target the tool initially – it has to be picked up – before one can act with it. Acting with one’s body is not like that. There is no such targeting needed as the body is the effector itself. Does our ordinary capacity for bodily action require conditions on embodiment that go beyond the body being the effector? The denial that one needs to target one’s body in acting with it comes together with a certain structure of awareness of one’s body as one acts with it. The kind of awareness of one’s body required for acting with it isn’t the kind of awareness required of an object that one intends to act on. Unlike objects we act on, the object we act with – one’s body – doesn’t (typically) require conscious attention. This is evident in many everyday activities that we engage in, where our attention is focused on the external objects that we are acting on. Most of us do not reflect much on our bodies as we act. Fluid everyday action does not require conscious attention to the bodily effectors of the action (Cole and Paillard 1995; Ingram et al. 2000). For example, when I am walking down the street, I am looking at whatever catches my attention. I am often not attending to my walking and certainly not attending to the joint angle displacements and muscular forces involved in walking. Conscious attention to one’s effectors is required only in cases where one learns a new movement or skill, or in cases of injury, recovery, and rehabilitation. Clearly, in some cases, we act with conscious attentional control, but this is not the norm. The absence of conscious attentional control is a generic characterisation of a central feature of much of ordinary everyday action. How is this possible? It appears to be possible because we are aware of the effector in some way. To see this, let us compare the ordinary bodily action of healthy afferented agents with those of deafferented agents. As noted above, the character of deafferented agency differs from that of afferented agents. The ordinary bodily action of deafferented agents is effortful and controlled via deliberate visual attention. In contrast, the ordinary bodily action of afferented agents has a kind of effortless automaticity. Furthermore, deafferented agents have severe problems in online control of bodily action, whereas afferented agents have fluid online control of bodily action. Since the key difference between afferented and deafferented 195

Hong Yu Wong

agents is intact proprioception, this suggests that the capacity of proprioception is required for the capacity of ordinary bodily action as we know it (Wong 2015).

2.2 Proprioception as a condition on embodied agency What account can we give of how proprioception is a condition on ordinary bodily action as we know it? A classic philosophical account is due to O’Shaughnessy (1995, 2008), who claims that conscious proprioception is required for all bodily action because it is used in the online control of bodily actions. His argument is based on the apparent inconceivability of acting on some body part in the absence of conscious proprioception. His idea is that conscious proprioceptive awareness is what allows the bodily will to engage the body in action. O’Shaughnessy’s view may seem initially attractive on phenomenological grounds, as bodily awareness and bodily action appear intimately connected in everyday lived experience, but it is subject to counterexamples. We have already encountered deafferented agents who can act directly with parts of their bodies that they have no proprioceptive awareness in (Cole and Paillard 1995; Cole 2016). Hence, O’Shaughnessy’s inconceivability claim is incorrect. Furthermore, there is evidence that conscious proprioceptive awareness is not necessary for the online control of bodily action. There is strong experimental evidence that normal, healthy subjects can successfully perform various actions they are tasked with whilst under proprioceptive illusion (Dijkerman and de Haan 2007). If conscious proprioceptive awareness is providing wrong parameters for online control of bodily action, it cannot be necessary for successful online control (see Wong 2015 for more detailed argument). The failure of O’Shaughnessy’s account teaches us that conscious proprioceptive awareness is not necessary for action, as deafferented agency demonstrates. However, the presence of the capacity of proprioception shapes the capacity for ordinary bodily action, as the argument above from contrasting the difference between afferented and deafferented agency indicates. Thus, there is a sense in which embodiment is key for embodied agency as we know it. This suggests that we need an alternate account of how proprioception shapes ordinary bodily action. One alternative account is in terms of the capacity of proprioception being required for the capacity of ordinary bodily agency in normal agents (Wong 2018). There are two key differences from O’Shaughnessy’s account. First, conscious proprioceptive awareness is not required for ordinary bodily action, but the capacity for proprioception is required. Second, the dependence between bodily awareness and bodily action is at the general level of capacities rather than at the level of the specific deliverances of the capacities on each occasion. The capacity for proprioception in healthy, neurologically normal agents functions to provide proprioceptive information, which is sometimes manifest in episodes of conscious proprioceptive awareness of one’s body ‘from the inside’. (It is important to note that while proprioception is a sensory modality with a dedicated receptor system and a proprietary phenomenology, we should not think of the body representations underlying proprioception as unimodal but rather as multimodal; see de Vignemont 2014 and Wong 2017b.) The experiments above where there is successful action despite proprioceptive illusion show that proprioceptive information for action control does not have to be provided by proprioceptive awareness and that proprioceptive representations for conscious proprioceptive awareness can be distinct from those that are used in the control of bodily action (Dijkerman and de Haan 2007). But such dissociations do not challenge the claim that the capacity of proprioception is required for the capacity of ordinary bodily agency in normal agents – as the contrast between afferented and deafferented agency attests to (for a direct argument see Wong 2018). 196

Agency and the body

The capacity for proprioception is a condition on non-observational knowledge of intentional action as well. The sense of agency is the experience of acting, including the sense that you are the author of your actions and some sense of control over the action as it unfolds. As I noted earlier, part of the normal sense of agency is that when we are intentionally performing basic bodily actions, we do not need to look and see before we know what we are doing. We have non-observational knowledge of intentional basic bodily action. However, deafferented agents do not have such non-observational knowledge of what they are doing even in the case of intentional basic bodily action (Cole 1992, 2016; Balslev et al. 2007). They may have knowledge of what they are trying to do, but they do not have non-observational knowledge of what they are doing. Thus, the capacity for proprioception is also vital to a normal sense of bodily agency.

3 A sense of body ownership Moving beyond proprioception, what other aspects of awareness of one’s body have an impact on bodily agency? Other forms of bodily awareness include pain, touch, the sense of balance, and interoception (the perception of one’s visceral states, e.g. heartbeat, hunger, thirst, need to micturate). These – including proprioception – are all what we might call first-order sensory sources about one’s body, each involving its own receptor systems. Instead of surveying how each can shape agency, I want to end by considering a more complex form of bodily awareness that is probably the most vital aspect of embodiment and examine how it shapes agency. One not only feels one’s body ‘from the inside’, but one’s body feels to be one’s own – one has a sense of ownership over one’s body (Martin 1995; Bermúdez 2018 and de Vignemont 2018). Why think that there is such a phenomenon as a sense of body ownership? First, there are cases of the loss of a sense of ownership of parts of one’s body. In somatoparaphrenia, which is associated with extensive lesions in the right hemisphere, patients report that a limb feels alien and deny ownership of it, often ascribing it to someone else (Vallar and Ronchi 2009). Second, we can artificially induce an illusory sense of ownership over a fake limb or even an avatar under specific conditions of stimulation, such as visuo-tactile synchrony (Blanke 2012; Botvinick and Cohen 1998). Finally, in everyday experience of one’s body, located bodily sensations (e.g. itches and pains) feel to be within the boundaries of one’s body (Martin 1995). Though in normal healthy human beings this sense of body ownership may be anchored in first-order forms of bodily awareness (e.g. Martin 1995; Bermúdez 2019), it does not appear to require any fixed first-order anchor. For example, the deafferented agent IW reported losing a sense of ownership over his body when he first lost proprioception and touch but regaining it when he learned to control his body (Cole 1992 and 2016). Thus, the sense of body ownership appears to be something more complex than first-order forms of bodily awareness, which results from integrating multiple sensory and motor sources. What constitutes the sense of body ownership is a topic of current controversy. But we do not need to go into the many competing accounts for our purposes. How does the sense of body ownership bear on action? Patients with problems with the sense of ownership of parts of their body typically also have problems with acting with those limbs and have a defective sense of agency relating to acting with those limbs (de Vignemont 2018). This would appear to suggest that a sense of ownership of a limb is a condition on acting directly with that limb and having a sense of agency in so doing. Call this the ownership constraint. Brain lesion data also point to a connection between the sense of body ownership and the sense of agency (Baier and Karnath 2008). The ownership constraint is perhaps 197

Hong Yu Wong

what we would expect a priori. Why? It makes no (rational) sense to try to act directly with a body that feels not to be yours. And if it is not yours why would you be able to directly control it? If we think the sense of ownership marks bodily boundaries (delineated by the interplay of bodily awareness and touch à la Martin 1995 or by defensive boundaries à la de Vignemont 2018), it would be magical to think you can act directly with something outside your bodily boundaries under normal circumstances in the absence of any linking mechanism. That would be telekinesis. So it is not implausible to think that a sense of ownership over a body is a condition on acting directly with that body and having a sense of agency over that body when acting with it. Recent neuropsychological studies in patients with disturbed body ownership also provide support for the ownership constraint (Garbarini et al. 2017). These patients experienced a delusory sense of agency over movements made by a hand they saw only when they had a delusory sense of ownership over that hand. There is a challenge to the ownership constraint though. We act directly and skilfully with tools that we know how to use, like cutlery. Yet we do not have a sense of ownership over these extracorporeal objects. One response is to claim that the sense of ownership is required only for directly striving with one’s body. The ownership constraint does not apply to extracorporeal objects. What are the grounds for this restriction? Acting with the body is special. It is striving through the corporeal object that one is. Something like this distinction would seem to be reflected in the empirical finding that a sense of ownership can only be induced in body-like objects, but not in other extracorporeal objects, such as tools (Tsakiris 2010). If this is correct, it means that the capacity to act directly with something does not require that the agent must feel it ‘from the inside’ or have a sense of ownership over it (Wong 2015). This would suggest that these constraints of embodiment on agency that I have been discussing are not constraints on agency tout court, but constraints on the kind of agency we have, namely, embodied agency. This restriction on the ownership constraint illustrates just how special the body is in our agentive life. But even this more restricted claim might be threatened by empirical cases. In xenomelia, also known as apotemnophilia or body identity integrity disorder, we have subjects who have a desire to amputate some part of their limbs in order to feel whole (Brugger et al. 2013, Sedda and Bottini 2014). In some of these cases, there may be grounds for thinking that these subjects feel some part of their limb as alien, yet they still use these limbs. However, we do not yet understand the pathology of xenomelia – and it is as yet unclear whether the alien affect and negative attitudes towards the limb which a xenomelic wishes to amputate are really comparable to the cases of disownership in somatoparaphrenia. So the threat of such cases to the restricted ownership constraint is still uncertain. One might even speculate that the a priori link codified in the ownership constraint may figure in an explanation of a xenomelic’s desire for amputation. Amputation would take the limb out of action and out of their bodies. They do not feel they should act with the limb in question, for they don’t feel right about it. The case of xenomelia calls for further research. To conclude, our question concerned the role of the body in bodily action. I have argued that we must see the body as structuring our agency, and not just as an enabling condition. In considering how the body and bodily awareness structure our capacity for acting directly with our bodies and our sense of agency in so striving, I have been tracing the outlines of an account of embodied agency.1

Related topics Agency, consciousness, and attention; Animal agency; Mental agency; Agency and practical knowledge; Agency, powers, and skills; Expert agency; Agency and disability; Pathologies of agency; Material agency. 198

Agency and the body

Note 1 I am indebted to Stefan Brandt, Chiara Brozzo, Malte Hendrickx, Erasmus Mayr, Matt Nudds, and especially Stephen Butterfill, Denis Bühler, Gregor Hochstetter, and Krisztina Orbán for their feedback. This publication was made possible through the support of an ACT Fellowship from the John Templeton Foundation. The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the John Templeton Foundation.

Recommended reading 1 Bermudez, J. L., Marcel, A. J., and Eilan, N. (1995) The Body and the Self. MIT Press. This is the volume from which the contemporary discussion of body, self, and agency sprang. 2 O’Shaughnessy, B. (2008), The Will, 2nd ed. Cambridge University Press. This is the 2nd edition of O’Shaughnessy’s classic work (1st ed. 1980), which includes pioneering discussions of volitionism (or ‘trying theory’), the body image, the phenomenology of embodied agency, sub-intentional actions, and a dual aspect theory of action. 3 De Vignemont, F. and Alsmith, A. (2017) The Subject’s Matter. MIT Press. The latest collection of essays exploring themes inspired by the discussion of the body, self, and agency inaugurated by Bermudez et al.’s The Body and the Self (1995). 4 De Vignemont, F. (2018) Mind the Body. Oxford University Press. This is a cutting-edge discussion of philosophical issues in bodily self-consciousness arising from reflecting on the cognitive science of body representations, which culminates in a theory of body ownership relating to body representations for action (body schema) for self-defense. 5 Wong, H. Y. (forthcoming) Embodied Agency. Oxford University Press. This is a systematic discussion of the relation between embodiment and agency, bringing together the recent philosophy and science in an account of embodied agency.

References Anscombe, G. E. M. (1963). Intention, 2nd ed. Oxford: Blackwell. Baier, B., and Karnath, H. O. (2008). Tight link between our sense of limb ownership and self-awareness of actions. Stroke, 39(2): 486–488. Baloh, R. W., Honrubia, V., and Kerber, K. A. (2011). Baloh and Honrubia’s Clinical Neurophysiology of the Vestibular System, 4th ed. Oxford: Oxford University Press. Balslev, D., Cole, J., and Miall, R. C. (2007). Proprioception contributes to the sense of agency during visual observation of hand movements: Evidence from temporal judgments of action. Journal of Cognitive Neuroscience, 19(9): 1535–1541. Bermúdez, J. L. (2019). Bodily ownership, psychological ownership, and psychopathology. Review of Philosophy and Psychology, 10(2): 263–280. Berthoz, A. (1991). Reference frames for the perception and control of movement. In J. Paillard (Ed.), Brain and Space (pp. 81–111). Oxford: Oxford University Press. Berthoz, A. (2000). The Brain’s Sense of Movement (Vol. 10). Cambridge, MA: Harvard University Press. Blanke, O. (2012). Multisensory brain mechanisms of bodily self-consciousness. Nature Reviews Neuroscience, 13(8): 556–571. Botvinick, M., and Cohen, J. (1998). Rubber hands ‘feel’ touch that eyes see. Nature, 391(6669): 756–756. Brugger, P., Lenggenhager, B., and Giummarra, M. J. (2013). Xenomelia: A social neuroscience view of altered bodily self-consciousness. Frontiers in Psychology, 4: 204. Cole, J. (1992). Pride and a Daily Marathon. Cambridge, MA: MIT Press. Cole, J. (2016). Losing Touch: A Man without His Body. Oxford: Oxford University Press. Cole, J., and Paillard, J. (1995). Living without touch and peripheral information about body position and movement: Studies with deafferented subjects. In J. Bermúdez, A. Marcel, and N. Eilan (eds.) The Body and the Self. Cambridge, MA: MIT Press: 245–266. Davidson, D. (1980). Essays on Actions and Events. Oxford: Oxford University Press. Dijkerman, H. C., and De Haan, E. H. (2007). Somatosensory processing subserving perception and action: Dissociations, interactions, and integration. Behavioral and Brain Sciences, 30(2): 224–230.

199

Hong Yu Wong Garbarini, F., Pia, L., Fossataro, C., and Berti, A. (2017). From pathological embodiment to a model for body awareness. In F. de Vignemont & A. Alsmith (Eds.), The Subject’s Matter: Self-consciousness and the Body. Cambridge, MA: MIT Press: 371–388. Hornsby, J. (1980). Actions. London: Routledge. Ingram, H. A., Van Donkelaar, P., Cole, J., Vercher, J. L., Gauthier, G. M., and Miall, R. C. (2000). The role of proprioception and attention in a visuomotor adaptation task. Experimental Brain Research, 132(1): 114–126. Martin M. G. F. (1995). Bodily awareness: A sense of ownership. In J. Bermúdez, A. Marcel, and N. Eilan (eds.) The Body and the Self. Cambridge, MA: MIT Press, 267–289. O’Brien, L. (2007). Self-knowing Agents. Oxford: Oxford University Press. O’Shaughnessy, B. (1995). Proprioception and the body image. In J. Bermúdez, A. Marcel, and N. Eilan (eds.) The Body and the Self. Cambridge, MA: MIT Press, 175–203. O’Shaughnessy, B. (2008). The Will: A Dual Aspect Theory, 2 vols., 2nd ed. Cambridge: Cambridge University Press. Searle, J. R. (1983). Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Sedda, A., and Bottini, G. (2014). Apotemnophilia, body integrity identity disorder or xenomelia? Psychiatric and neurologic etiologies face each other. Neuropsychiatric Disease and Treatment, 10: 1255. Steward, H. (2009). Animal agency. Inquiry, 52(3): 217–231. Tsakiris, M. (2010). My body in the brain: A neurocognitive model of body ownership. Neuropsychologia, 48: 703–712. Vallar, G., and Ronchi, R. (2009). Somatoparaphrenia: A body delusion. A review of the neuropsychological literature. Experimental Brain Research, 192(3): 533–551. Velleman, D. J. (2000). The Possibility of Practical Reason. Oxford: Oxford University Press. De Vignemont, F. (2014). A multimodal conception of bodily awareness. Mind, 123: 989–1020. De Vignemont, F. (2018). Mind the Body: An Exploration of Bodily Self-awareness. Oxford: Oxford University Press. Wong, H. Y. (2015). On the significance of bodily awareness for bodily action. Philosophical Quarterly, 65: 790–812. Wong, H. Y. (2017a). In and out of balance. In F. de Vignemont & A. Alsmith (Eds.), The Subject’s Matter: Self-consciousness and the Body. Cambridge, MA: MIT Press: pp. 311–334. Wong, H. Y. (2017b). On proprioception in action: Multimodality versus deafferentation. Mind & Language, 32(3): 259–282. Wong, H. Y. (2018). Embodied agency. Philosophy and Phenomenological Research, 97(3): 584–612.

200

18 AGENCY, CONSCIOUSNESS, AND ATTENTION Wayne Wu

How are agency and consciousness intertwined? This entry focuses on visually guided bodily action where visual attention serves as an anchor for visual phenomenology in action. The emphasis is not to denigrate mental actions (see Soteriou, this volume) nor is it due to a coarse visuo-centrism. Rather, much of the detailed empirical work relevant to our topic has been done on vision and the visual guidance of motor movement. I focus on this literature to help us understand conscious intentional agency in an empirically informed way. Against empirical arguments to the contrary, I contend that a common sense picture of consciousness in action is tenable: conscious vision guides action and thus contributes centrally to the phenomenology of visually guided agency.

1 Consciousness I draw on a familiar distinction between phenomenal consciousness and access consciousness (Block, 1995). Phenomenal visual consciousness involves those aspects of visual experience that we cite when we describe what that experience is like. Such descriptions often draw on what one sees: If you were to say what your current visual experience is like, you would speak about what you see, say that your experience is like typical experiences of reading printed text. You can report the specific text that you are currently reading and can go on to describe the features of the text to further specify what it is like, say that the experience is low in contrast between printed word and page. Specifying what it is like can draw on descriptions of the visible world revealed from one’s visual point of view, and in that sense, introspection draws in a natural way on visual attention (Harman, 1990). Access consciousness in this context is exemplified by the fact that in describing what experience is visually like, your cognitive-conceptual system draws on visual content to inform the introspective report. In general, we can treat the form of access consciousness as involving one system drawing on the content of another. In visually guided intentional action, the action system draws on visual content to ensure appropriate visual guidance. Noting that introspection in reporting what it is like to see is itself an intentional action, we can opt for a broad notion of access consciousness that focuses on the visual guidance of action that includes bodily action and introspective action. Strikingly, theorists have restricted access consciousness to introspection in part because of doubts about the guiding role of visual consciousness in bodily action. DOI: 10.4324/9780429202131-23201

Wayne Wu

2 A structure for action I derive a basic structure for action beginning with a simple idea that there is a category of behaviour constituted by a response, the output, being driven by a subjective state, the input. Psychologists speak about behaviours as stimulus-response mappings, but in actions, this mapping is mediated by mental state-response mappings where the mental state registers the relevant stimulus. So, actions can be understood as among the category of mental state-response mappings and visually guided actions as visual state-response mappings. The subject’s seeing the world guides the produced response. Thus, we can think of behaviours generically as in Figure 18.1.

Figure 18.1

Where I is the input mental state that represents the stimulus in respect of which the output O is generated. The point of the coupling is that the input state informs the output response, and we can understand this in part by treating the output system as accessing the content of the input state. Now assume that reflexes are not actions. A reflex is a rigid form of behaviour where specific inputs necessitate the output, what I call a pure reflex to separate it from biological reflexes. The significance of necessitation is that there is no alternative mapping of mental state input to response output available. Canonical reflexes are certainly not actions (say the flexor reflex that moves a body away from an extremely hot surface) and a fortiori pure reflexes are not either. We can represent a pure reflex as follows:

Figure 18.2 The “N” indicates that necessarily, if I occurs then O occurs. This means that no other responses to I are available, and there is only one possible behaviour.

Purely reflexive behaviour is not an action since the outcome is driven entirely by the input. Where we have action, pure reflex must be absent. This entails the absence of the necessity that defines the pure reflex, but eliminating necessity yields additional behavioural possibilities: the input I can be mapped to other outputs. Minimally, we have a mapping between I and O1 and O2, so the mapping of possible behaviours branches. Call this representation of behavioural possibilities an action space. This branched space fixes the possibilities for action for an agent at a time. What we have shown is that if actions rule out pure reflex and thus open up behavioural options, then actions entail a branched action space. The branched action space forms the basis of what I call the Many-Many Problem (Wu 2011). Once additional behavioural possibilities are on the table, a question arises: which behaviour will be performed? If actions as philosophically conceived involve agentive control, then the selection of one behaviour rather than another, given the Many-Many Problem, cannot be arbitrary. As intentional actions often arise from an agent’s intention and yet must emerge from a branched action space, for intentions to bring about action, they must solve the Many-Many Problem. Specifically, the path taken in a action space satisfies the content of the intention precisely because the intention’s having that content explains why that path was taken. The term selection describes the fact that one path among many possible paths in a action space is executed, and for intentional action, intention explains why specific selection occurs. This explicates the psychology of an intention causing an action. 202

Agency, consciousness, and attention

Two notions are central to understanding agency, namely the contrast between automaticity and control. Actions are expressions of agentive control and yet actions are also automatic. While automaticity and control have been discussed in psychology, empirical theorists have despaired of providing a clear analysis. Here, I provide the analysis that I have argued for elsewhere that renders three seemingly inconsistent propositions consistent (Wu, 2013a): (1) Intentional actions express agentive control. (2) Intentional actions involve substantial automaticity (e.g. the muscle contractions during reach). (3) Automaticity rules out control and vice versa. Point (1) expresses the core idea that intentional actions express control, while (2) is a biological fact in that there are many facets of action that we do not control. Yet control and automaticity look to be mutually exclusive given (3). To make (1)-(3) consistent, we must not focus on processing kinds as automatic or controlled. Rather, every process can have automatic and controlled features. If so, an action as a unfolding process can exemplify control, as every intentional action must (1), but also be imbued with automaticity, as every action must (2), in a way where every property that is automatic is not one that is controlled and vice versa (3). If we understand intentions as a source of agentive control, then we can explain control in respect of a feature relative to an intention to act in a way that exemplifies that feature. More formally, we capture (3) as: An action A is controlled relative to its property F iff A’s having F occurs because the agent intended to A in an F-way. An action A is automatic relative to its property F iff A’s having F is not controlled. Every intentional action then involves control in that some of its features are intended. Most features, however, are automatic. As skills develop, there will be a shift in the properties of the action that are automatic versus controlled (see Montero this volume). At the extreme, when no intention is involved, a behaviour is fully automatic, that is passive, in not involving any control. Reflexes are fully passive in this sense, but there are arguably pathologies of agency that are passive as well such as thought insertion in schizophrenia (see Radoilska this volume on pathologies of agency). Passive actions involve the fully automatic exercise of agentive capacities. The structure of intentional action begins with an action space where many mental state inputs can inform different responses. Action arises when a specific input-response coupling occurs, and intentional action occurs when this coupling is the result of an intention, one that specifies which features of the action are automatic and which are controlled.

3 Attention and action William James characterized attention as follows: Everyone knows what attention is. It is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought. Focalization, concentration, of consciousness are of its essence. It implies withdrawal from some things in order to deal effectively with others. ( James, 1890) 203

Wayne Wu

James’ characterization of attention echoes the dynamics of selection in the Many-Many Problem. James speaks of the mind’s taking possession of one among many objects or trains of thought, and this entails withdrawing from some things to deal effectively with the object taken possession of. This is precisely what occurs in the selection that solves the Many-Many Problem: a specific mental input among many comes to inform a response with the subsequent withdrawal from other behavioural possibilities. Since James identifies attention with the mind’s taking possession of one among many objects to deal effectively with it, the simplest realization of his characterization in an action space is to identify the state of attention with the mental state that is the input to the selected path. We can also identify the process of attending with the process of the input guiding the response output. Thus, if we are considering a visually guided action, then the visual state that represents the target of action constitutes the subject’s visual attention to that target when that state informs the response. When this input selection occurs due to a subject’s intention, we have an intentional action in that a subject responds to a target given how she sees it in light of her intention. When the action occurs, the mind has taken possession of one target among many to deal effectively with it, withdrawing from other targets and the responses they afford. That is, the subject has attended. One complexity that we will set aside is that many theorists of attention hold a spotlight model of it. Many things are meant by that metaphor, but a core idea is that attention causes the selection in the Many-Many Problem. In contrast, I hold that attention is not what causes selection, but is part of selection (Desimone and Duncan, 1995; Anderson, 2011). Attention’s causal role is that given selection, it guides response. Guidance here is to be understood in terms of the content of the input state informing the generation of the response. In that way, attention explains the notion of guidance deployed in philosophy of action (Frankfurt, 1988; Wu, 2015). That said, for present purposes, either model will do in terms of understanding attention, action, and consciousness. I will, however, continue to speak of attention as the input state that guides action and not as a spotlight (for other arguments against the spotlight conception, see Allport, 2011; Mole, 2011).

4 Attention and blindness: is visual phenomenology in action like tunnel vision? The set of input visual states that structure the agent’s action space at a given moment in time are the set of visual states that the agent is in at that time which are available to guide response capacities. Yet when action occurs, given selection, only a subset of those states will constitute attention, namely, those that guide response. Some of these visual states will be phenomenal states (we will consider unconscious visual states below). The visual phenomenology inherent in agency will be captured in part by delineating the total number of input visual states that structure an action space for an agent at a time. Some of those visual states, namely those that guide response, will fix the states which constitute visual attention. So, the visual phenomenology that occurs during visually guided action is set first by the total phenomenal visual states that provide inputs to the action space within which action emerges. Some visual states constitute attention, while others do not. We thus begin with an overflow picture of visual phenomenology (Block, 2008). What it is like to be an agent in visually guided action is fixed by what the agent sees, given the action space. Yet given the structure of action, only some of what the agent sees will be a target of attention. With attention in view, however, we confront a gatekeeping view about attention and consciousness: visual 204

Agency, consciousness, and attention

phenomenology is restricted by attention in that conscious vision is limited to what one visually attends to. For all other visual states, one is inattentionally blind (Mack and Rock, 1998). Note that this makes vivid a counterintuitive consequence of the gatekeeping view: in visually guided action, agents have a form of tunnel vision fixed by visual attention. We can expand from attention to access. The visual states that do not engage action are not access conscious in that they are not accessed, and for many theorists, this means that those states are not phenomenally conscious, access consciousness being a necessary condition of phenomenal consciousness. For example, on the Global Workspace theory of consciousness, access consciousness of a visual state requires entry into the Global Workspace which allows its content to be broadcast, that is accessed, by action control systems (Baars, 1988; Dehaene and Naccache, 2001). Entry into the workspace is gated by attention, so attention is necessary for access consciousness as well. Yet if the relevant selection to guide performance is restricted to the object of visual attention, then prima facie, no further visual contents enter the Workspace and, by hypothesis, no further visual contents are phenomenally conscious. Thus, inattention to X abolishes access consciousness of X (entry into the global workspace) which then abolishes phenomenal consciousness of X. Most scientists and philosophers writing on visual consciousness deny overflow. From an empirical point of view, most maintain that there is substantial evidence in favour of attention and access as gating phenomenal consciousness. It is worth recalling, however, a Jamesian constraint in empirical work on attention. When studying attention, scientists must control how their experimental subjects attend, and they do so by designing experimental tasks where the desired target of attention is made relevant to correct performance of the task. The subject must select the target to guide their performance. This is the Jamesian idea of taking possession of the stimulus to deal effectively with it and can be expressed as the following sufficient condition: If a subject S visually selects X to guide task performance, S visually attends to X. Once this condition is made clear, however, we discover an immediate problem with the evidence that is purported to support gatekeeping. If one holds that visual attention to X is a necessary condition for visual consciousness of X, then we test such conditions by eliminating what is purportedly necessary, attention, and showing that consciousness disappears. Eliminating visual attention should eliminate visual consciousness. At the same time, our evidence for visual consciousness requires access in introspective report, in this case, report that we see or do not see X. Access of X, however, requires attention to X, but now the experimental methodology undermines itself. To test whether attention to X is necessary for consciousness of X, we must eliminate attention to X, and yet this eliminates the basis for evidence for consciousness of X. Since the consciousness at issue concerns whether we are phenomenally conscious of X and introspective access relies on access to X, then the experimental methodology guarantees that we will not have the required introspective access. In short, if the experiment works, then it doesn’t work. That is, if we eliminate attention to test gatekeeping, we eliminate the very basis of access needed to assess whether consciousness is eliminated. This is true of all attempts to experimentally demonstrate that attention is necessary by eliminating attention, say in inattentional blindness paradigms, change blindness, the attentional blink and related neuropsychological cases like hemispatial neglect. The upshot is that despite widespread endorsement of inattentional blindness, the empirical evidence does not support the gatekeeping view. For our purposes, this leaves the appearances intact as a working hypothesis, namely that phenomenology overflows attention 205

Wayne Wu

and access. Visual attention fixes visual phenomenology but it does not exhaust it (for more discussion on options weaker than inattentional blindness such as inattentional agnosia (see Wu, 2017).

5 Unconscious guidance: is there zombie action? Some theorists have argued that even where a target of action is attended to, this vision in action is unconscious in the phenomenal sense. This claim rests on work on the functional division in the primate cortical visual system. At a coarse level of analysis, early cortical vision in the occipital lobe (back of the head) bifurcates into two anatomically distinct (though not fully segregated) streams, a dorsal stream that projects into the parietal lobe and a ventral visual stream that projects into the temporal lobe. Leslie Ungerleider and Mortimer Mishkin (Mishkin, Ungerleider and Macko, 1983) identified this division in non-human primates and assigned distinct functional roles to the streams: the dorsal stream is a where stream that plays a central role in spatially guided movements, while the ventral stream is a what stream that contributes to identification and categorization. Later, Melvyn Goodale and David Milner (Milner and Goodale, 1995) reconceptualized the functional characterizations of the two streams, treating the dorsal stream as the action stream and the ventral stream as the perception stream. Andy Clark (2001) drew on these findings to question the following idea: experience-based control: Conscious visual experience presents the world to the subject in a richly textured way, a way that presents fine detail… [and] is, in virtue of this richness, especially apt for, and typically utilized in, the control and guidance of fine-tuned, real-world activity. (2001, p. 496) The functional division in cortical vision is neutral on phenomenology. Why then do many think that the dorsal stream’s role in guiding motor movement involves unconscious guidance? In earlier work, I argued that certain dissociations between the dorsal and ventral streams allow one to argue that some visual information that guides behaviour is unconscious (Wu, 2013b). The experimental work done both with the neuropsychological patient DF, who shows degraded conscious vision with strikingly preserved visuomotor capacity, and also with normal subjects suggests that the ventral stream, as evidence by verbal report, is subject to visual illusions, while the dorsal stream, given motor output, is resistant to them. For example, using a form of the Ebbinghaus illusion that allowed the geometrical shapes to be grasped, Haffenden and Goodale (1998) showed that reach and grasp movements (dorsal stream guided) appeared to be resistant to the illusion, while verbal report (ventral stream guided) was not. They had subjects report on the sizes of the central disk as well reach for them. In this example, the central disks appear to be of different sizes even though they are the same. Subjects’ reach seemed resistant to the illusion, while their report did not (for details, see Wu, 2013b). While there is controversy regarding the data, the general tendency of the study seems to me robust enough to license the following argument focusing on the one inner disk that is both grasped in a reaching movement and whose size is reported on. Given that the ventral stream is under an illusion as to the disk’s size, given report, but the dorsal stream is not, given demonstrated grasp, premises (1) and (2) indicate different sizes as registered by

206

Agency, consciousness, and attention

Figure 18.3

Ebbinghaus Illusion.

those streams. Hence, under conditions of illusion but where a reach and grasp movement is initiated: (1) At t, the visual representation controlling reach (dorsal stream) represents the disk as size x. (2) At t, the conscious visual representation controlling report (ventral stream) represents the disk as size y where y≠x. (3) Assume for reductio that the representation in (1) is conscious. (4) So at t, there is a conscious visual representation of the disk as size x and one representing the same disk as size y. (5) So, there is a conscious visual representation of the disk as size x and size y. (6) If the subject visually consciously represents o as F and G, then o looks F and G. (7) So, the disk looks both size x and y (y≠x). (8) But the disk does not look that way. (9) So, the visual representation guiding reach in this task is not conscious. (Adapted from Wu 2013b, 223–234) The first two premises are derived from experimental observation. If the argument is sound, then we reject the claim that dorsal stream vision is conscious. If that stream realizes subject-level visual guidance of motor behaviour, then that behaviour is guided by unconscious visual states. Since this guidance involves visual selection in light of the Many-Many Problem (Cisek and Kalaska, 2010), then visual attention in the dorsal stream is also unconscious. Here, visual attention is not sufficient for visual consciousness in agency. Given that the reach to grasp movement in the experimental task likely draws on typical reach to grasp capacity in normal behaviour, the conclusion that I drew was more general: in mundane reach to grasp behaviour in normal subjects, the visually guided component is unconscious. Hence, many mundane bodily actions are zombie actions in the sense that visual guidance is unconscious. We can absorb the possibility of unconscious vision playing a pervasive role in reaching actions if we deny that the dorsal stream yields subject-level visual states. Rather, such states are subpersonal. There is a cost, however, to denying that the dorsal stream generates subject-level states. The fine-grained spatial processing in visually guided action relies on dorsal stream

207

Wayne Wu

processing. Consider then the skills that develop when one is learning a movement that requires sensitivity to spatial detail, one that is mediated by visual processing of spatial targets. Over time, one can see the improvement of a subject who practises the behaviour in question and who learns to generate more accurate and efficient responses. Her action demonstrates a shift from control to greater automaticity. This learning involves both changes in visual processing in perceptual learning and the fine-tuning of motor response. This means that learning affects dorsal stream visual processing and how it engages with motor behaviour. If we deny that the relevant dorsal stream visual states are subject level, then the picture of skilled behaviour is one that does not involve the subject at a critical stage. We are forced to say that the learning is all in the subpersonal withholding this achievement from the subject. But then the subject has not learned anything, and the increase in skill has nothing directly to do with her. She can only be pleased that her subpersonal underlings have taken her instruction well, given her intention to learn. This strikes me as giving up on the idea of a skilled agent in a basic arena of action. A crucial move in my earlier argument is the contrast between (7) and (8), that in reaching for an object there is an introspectible clash between the dorsal and ventral stream representations of the disk to be grasped. If we have subject-level states that differ in the represented metrical size of an object, I assume the subject would surely notice a contradictory content. The expectation would be that there would be something visually odd about how the size of the disk appears, and yet it seems that when we grasp objects, no such odd content is apparent. My current view is that for (8) to do the work required, the way things look in dorsal vision must be accessible to conceptual capacities (see also Mole, 2009, 2013). If the functional division of cortical vision is correct, on either the Milner-Goodale or the Ungerleider-Mishkin interpretation, then the dorsal stream content is not typically accessed by the conceptual system served by the ventral stream. For that reason, the subject is not in a position to jointly conceptualize the ventral and dorsal stream representations of size, and so her reports of her visual contents will not reflect both spatial representations. We should not have the expectation that the dorsal stream contents will inform our reports about how things look, as required in premise (8). So, the argument fails. This does not show that the dorsal stream is conscious, but it undercuts what seems to me the best argument to demonstrate that, in normal subjects, dorsal stream visual states are unconscious. This is consistent with Ungerledier and Mishkin’s emphasis on spatial awareness in the dorsal stream. If so, we have yet to show that experience-based control is false. For the sake of the current discussion, I am going to take on board experience-based control (in Wu, 2018). I support the idea that given the role of the dorsal stream in guiding intentional action, we should treat it as conscious. It is enough to note that the arguments for unconscious vision or for restrictions on vision via attention have not compelled us to deny this principle (Wallhagen, 2007).

6 What it is like to act, perceptually speaking The hypothesis that is currently on the table is that attention is a component of action explaining the notion of agentive guidance. In visually guided actions, it is visual attention that guides action. Given that vision is understood in the phenomenal sense, visual attention provides a type of subjective anchor to the world. This anchor can be spoken of as the agent’s focus. Outside of the agent’s focus remains a form of non-attentional visual awareness. In this way, we have returned to a view of the phenomenology of visual experience in agency as fixed by the visual structure that defines the action space of the subject. I have argued that we 208

Agency, consciousness, and attention

have no reason yet to endorse the claims that attention restricts the contours of the phenomenal visual field and that visual guidance is often unconscious, despite widespread agreement to the contrary on the basis of empirical work. If there are various visual inputs available to the subject when she acts, then these inputs can contribute in different ways to the visual phenomenal field. If this holds for vision, it also likely holds for the other perceptual modalities that feed into action. Perceptually guided action is structured by the guidance via perceptual attention which anchors perceptual awareness in action.

Related topics Mental agency; Agency and the body; Intentional agency; Expert agency.

Further reading Block, N. 2007. Consciousness, Accessibility, and the Mesh between Psychology and Neuroscience. The Behavioural and Brain Sciences, 30(5–6): 481–499. This work, with associated commentaries, provides an empirically engaged perspective on the issues about access and consciousness. A more philosophical presentation of the issues can be found in (Block, 2008). Goodale, M. A. and A. D. Milner. 2004. Sight Unseen: An Exploration of Conscious and Unconscious Vision. Oxford: Oxford University Press. An informal presentation of Milner and Goodale’s important work on human cortical vision, especially their collaborative work with the patient DF. Mack, A. and I. Rock. 1998. Inattentional Blindness. Cambridge, MA: MIT Press. A classic work on the limits of attentional capture. Mack and Rock coined the term inattentional blindness. Wu, W. 2014. Attention. New York: Routledge. This book introduces the topic of attention including two chapters that summarize work in psychology and neuroscience. Wu, W. 2017. Attention and Perception: A Necessary Connection? In B. Nanay, ed. Current Controversies in Philosophy of Perception. Routledge: New York: 148–62. This work provides a critical discussion on whether attention is necessary for perceptual consciousness.

References Allport, A. (2011) ‘Attention and integration’, in Mole, C., Smithies, D., and Wu, W. (eds) Attention: Philosophical and Psychological Essays. Oxford University Press US: 24–59. Anderson, B. (2011) ‘There is no such thing as attention’, Frontiers in Theoretical and Philosophical Psychology, 2: 246. doi: 10.3389/fpsyg.2011.00246. Baars, B. J. (1988) A Cognitive Theory of Consciousness. Cambridge: Cambridge University Press. Block, N. (1995) ‘On a confusion about the function of consciousness’, Behavioral and Brain Sciences, 18: 227–247. Block, N. (2008) ‘Consciousness and cognitive access’, Proceedings of the Aristotelian Society, 108: 289–317. Cisek, P. and Kalaska, J. F. (2010) ‘Neural mechanisms for interacting with a world full of action choices’, Annual Review of Neuroscience, 33(1): 269–298. Clark, A. (2001) ‘Visual experience and motor action: Are the bonds too tight?’ Philosophical Review, 110(4): 495–520. Dehaene, S. and Naccache, L. (2001) ‘Towards a cognitive neuroscience of consciousness: Basic evidence and a workspace framework’, Cognition, 79(1): 1–37. Desimone, R. and Duncan, J. (1995) ‘Neural mechanisms of selective visual attention’, Annual Review of Neuroscience, 18: 193–222. Frankfurt, H. (1988) ‘The problem of action’, in The Importance of What We Care About. Cambridge: Cambridge University Press: 70–84.

209

Wayne Wu Haffenden, A. and Goodale, M. (1998) ‘The effect of pictorial illusion on prehension and perception’, Journal of Cognitive Neuroscience, 10(1): 122–136. Harman, G. (1990) ‘The intrinsic quality of experience’, in J. Tomberlin (ed) Philosophical Perspectives, 4 Action Theory and Philosophy of Mind. Atascadero, CA: Ridgeview Publishing Company: 31–52. James, W. (1890) The Principles of Psychology, Volume 1. Boston, MA: Henry Holt and Co. Mack, A. and Rock, I. (1998) Inattentional Blindness. Cambridge, MA: MIT Press. Milner, A. D. and Goodale, M. A. (1995) The Visual Brain in Action. Oxford: Oxford University Press. Mishkin, M., Ungerleider, L. G. and Macko, K. A. (1983) ‘Object vision and spatial vision: Two cortical pathways’, Trends in Neurosciences, 6: 414–417. Mole, C. (2009) ‘Illusions, demonstratives and the zombie action hypothesis’, Mind, 118(472): 995–1011. Mole, C. (2011) Attention Is Cognitive Unison: An Essay in Philosophical Psychology. New York: Oxford University Press. Mole, C. (2013) ‘Embodied demonstratives: A reply to Wu’, Mind, 122(485): 231–239. doi: 10.1093/ mind/fzt031. Wallhagen, M. (2007) ‘Consciousness and action: Does cognitive science support (mild) epiphenomenalism?’ The British Journal for the Philosophy of Science, 58(3): 539. Wu, W. (2011) ‘Confronting many-many problems: Attention and agentive control’, Noûs, 45(1): 50–76. Wu, W. (2013a) ‘Mental action and the threat of automaticity’, in Clark, A., Kiverstein, J., and Vierkant, T. (eds) Decomposing the Will. Oxford: Oxford University Press: 244–261. Wu, W. (2013b) ‘The case for zombie agency’, Mind, 122(485): 217–230. Wu, W. (2015) ‘Experts and deviants: The story of agentive control’, Philosophy and Phenomenological Research, p. n/a-n/a. doi: 10.1111/phpr.12170. Wu, W. (2017) ‘Attention and perception: A necessary connection?’ in Nanay, B. (ed.) Current Controversies in Philosophy of Perception. New York: Routledge: 148–162. Wu, W. (2018) ‘The neuroscience of consciousness’, in The Stanford Encyclopedia of Philosophy ( Winter 2018 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/archives/win2018/entries/ consciousness-neuroscience/.

210

19 MATERIAL AGENCY Matthew Noah Smith

1 This essay explores how human agency is realized beyond the mind. To do this, we must begin with a few stipulative definitions. For the purposes of this essay, agents are beings that can act (i.e., agents can take actions). The term agency therefore refers to the capacity to act. (I am explicitly side-stepping questions of free will, autonomy, and responsibility: for all I say here, one can be an agent – one can have the capacity to act – but not be free, or be autonomous, or be responsible.) Saying someone is an agent tells us nothing about what specific actions she can take. I therefore use the term agential capacity to refer to the capacity to take a specific action. We can say that someone has the capacity for agency, and so she is an agent, and then further specify what agential capacities she has (she can walk, she can write, she can climb, etc.). With this out of the way, I now identify my starting point, which is perhaps slightly unusual for philosophical inquiry into agency. That starting point is the observation that as a general rule, agents vary in their agential capacities. The philosophical literature on human agency typically analyzes agency entirely by way of generic cases that presumably apply to all agents, typically treating borderline cases, such as the senile or the developing child or nonhuman animals, as special puzzles. Common examples are raising one’s arm, walking across the street, and other activities performed with limbs. Thus, we find in central texts such as Frankfurt (1978), Velleman (1992), Bratman (2003), and Korsgaard (2009), construction of generic agency-realizing psychological architectures by way of attention to a narrow menu of agential capacities, with difference being an explanatory afterthought (that sameness in specific capacities is the unquestioned starting point perhaps reveals an insensitivity in the literature to disability). I aim to avoid that by centering my account on differences in agency. Attention to these differences reveals something important about the nature of human agency. Thus, the question that provokes this essay is as follows: How is agency differentiated into different agential capacities? The typical approach to answering this question is to explain differences in agential capacities by appeal to psychological difference. This, I shall argue, misses what is interesting about how material differences contribute to the explanation of differences in agential capacities.

DOI: 10.4324/9780429202131-24211

Matthew Noah Smith

2 Non-mental agential capacities exist only when psychological elements of agency can be materially realized. In what material must these psychological elements be realized? If one generalized from the examples in the literature, one might conclude that all we do when we act is move our bodies (we raise our arms, we move our hands, we walk, and so on). So, the material component of our agential capacities amounts to the human body. Thus, differences in agential capacities are just bodily differences (e.g., John lacks an arm, so John cannot raise his arm like Jack can; Jack is physically exhausted, so Jack cannot run like John can). But a closer look even at bodily differences reveals complications. Human bodies that lack limbs can be extended or ‘fixed’ via prosthetics, which, in turn, may become part of the bodily schema of the prosthetic wearer. There is no principled reason to deny that good prosthetics cannot, in principle, be functional equivalents of the biologically given body parts for which they are replacements. On the most basic functionalist view, prosthetically given and biologically given body parts could be identical. This observation invites a distinction between (i) a functional-parity view according to which agency is understood as the intentional movement of bodies, where our ‘agential bodies’ can include technologies that make up for absences in our biological bodies and (ii) a more radical view according to which agency is understood as the intentional employment of technology, where the body is as much a technology as a hammer. As just suggested, any view of agency built on a widely acceptable functionalism should accept (i), namely, that the intentional movement of prostheses making up for fleshly absences can be treated on par with the intentional movement of non-prosthetic or ‘natural’ fingers. This is an underappreciated point in philosophy agency. Differences in our capacities are not biological differences but are instead differences in what we might call the functional body, which is to say the body functionally conceived as opposed to the body biologically conceived. But what of the more radical view, according to which there is no principled distinction between body and tool? Let us state this view as follows: Human agency is the capacity to use technologies for the sake of some end. The human body counts as a technology as much as a hammer or a computer. The body, on this view, is a tool that is produced, reproduced, transformed, and so on, through the exercise of human agency. But this has the consequence of eliminating any obvious boundaries between what one acts with and what one affects: the operation of the entire car one drives is on a par, at least from the understanding of human agency, as is the bare shifting of one’s arm as one steers that car. Aside from the just identified boundary problem, this view of technology as bodily extensions as opposed to fundamentally non-bodily, is not new. We can see it in Freud (1962, 42): With every tool man is perfecting his own organs, whether motor or sensory, or is removing the limits to their functioning. Motor power places gigantic forces at his disposal, which, like his muscles, he can employ in any direction; thanks to ships and aircraft neither water nor air can hinder his movements; by means of spectacles he corrects defects in the lens of his own eye; by means of the telescope he sees into the far distance; and by means of the microscope he overcomes the limits of visibility set by the structure of his retina. Notice how seeing through spectacles, telescopes, and microscopes, on the one hand, and seeing using only the retina, on the other hand, are treated as members of the same genus of activities. Notice further how using one’s muscles to act and using motor power to act 212

Material agency

are similarly treated as members of the same genus of activity. Notice how the movement of ships and aircraft, on the one hand, and the movement of the body, on the other, are treated as members of the same genus of activity. This, I argue, is the correct view of agency. It’s not primarily about bodily movement, or movement at all. It is the capacity to intentionally do things with the world, whatever happens to be in that world. According to this approach, the body should not be understood as biologically given but instead as artificial. Haraway (1990) defends a version of this view, which treats the body as a ‘Deleuzian assemblage’ seamlessly meshing the organic and the inorganic.1 Marc Wigley (1991, 8) extends the body to include the built environment in this way: …the body depends on foreign elements that transform it. It is reconstituted and propped up on ‘supporting limbs’ that extend it… the prosthesis reconstructs the body, transforming its limits, at once extending it and convoluting its borders. The body itself becomes artifice. Haraway’s and Wigley’s claims do not rest on an appeal to functional parity between the biologically given and the technologically given. Rather, they are claiming that bodies are transformed and extended through interactions with their technological environments. It is in this vein, then, we should read Seltzer (1992) on the characterization in Ford (1923) of the many different ways in which individual workers physically mesh with Ford’s new assembly line system of production. Seltzer writes that Ford’s account ‘projects a transcendence of the natural body and the extension of human agency through the forms of technology that supplement it’ (Seltzer 1992, 157).2 When we act, we act through our bodies but also technologically. From the perspective of understanding human agency, in short, there is no privileged ‘material’ for human agency. There are just technologies we intentionally use for the sake of some end.

3 Agency, on this view, is not a natural power of the body (see the list of examples of agency so understood in Small 2017, 250), but is instead the capacity to use the world for the sake of some end. Consider, for example, the capacity to write. This is not a capacity to move one’s body in a writing fashion, where the world takes over from there. In addition to a system of written language, the capacity to write requires a functioning writing implement. If one lacks such an implement, one does not have the capacity to write. Surely, though, if you’ve learned how to write, then you have the capacity to write even if you lack the implement. No, you haven’t. For one learns how to write with certain implements. One can have the capacity to write with a crayon but not with a ball point pen. One can learn to write with a ball point pen but not with a fountain pen. One can learn to write with a fountain pen but not with a quill. One can learn to write with a quill but not with a chisel. The capacity to write is not a generic capacity realized in a mind and a biologically given body, but instead a capacity to use certain tools for writing. Drawing again on Wigley, I am conjecturing that the discipline of writing – of learning how to use writing implements in order to make the marks one intends to make – ‘reconstructs the body, transforming its limits, at once extending it and convoluting its borders. The body itself becomes artifice….’ Here the artifice is realized in the specific character that the writing implement–hand interface must take on in order for a person to be able to write. Our agential capacities, therefore, are developed in relation to certain material affordances designed for making marks (crayons, big bulky 213

Matthew Noah Smith

chalk, pens, pencils, paintbrushes, fountain pens, and so on). The specific capacity to write is therefore a capacity to write with a certain thing. As people who have taught young children to write can attest, many can write with one implement but can barely make a mark with another, seemingly very similar implement. To say of the child who can only write with bulky writing implements like large-size crayons or fat felt-tip pens that she can write, therefore, undercharacterizes her capacity. In a world of fountain pens, she cannot (yet) write at all. This approach is counter to the view that our practical abilities come in general and specific forms (see, e.g., Mele 2003; Glick 2012; Maier 2015). On these views, if you can write, then you have the general ability to write. Then there is the specific ability to write with whatever implement which happens to be around. Small (2017, 243) has offered the following example as an illustration of this view: Serena Williams has the [general] ability to hit a forehand winner down the line, on the run. It is the same ability she exercises whenever she hits one, but the opportunities come and go, and some she does not take. When inebriated or injured, Williams does not lack a physical ability to exercise her ability to play tennis; rather she has a condition that prevents her from exercising her tennis-playing abilities. But presumably if it is too windy, Williams cannot hit a forehand winner down the line. Williams, while perhaps the greatest tennis player ever, has had at least more difficulty on clay courts like the ones at Roland Garros Stadium than she has had on other surfaces. Imagine now that she is playing the greatest clay court tennis player ever, who is also a man and so probably able to occasionally hit balls harder and faster than she can: Rafael Nadal. Perhaps Williams has met her match: she cannot hit a forehand winner down the line, on the run when she is playing Rafael Nadal at his best, on a windy day at the French Open. Perhaps Williams is unique: she can hit forehand winners on the run anywhere any time against anyone. But this is only because she is the greatest of all time. This is the exception that proves the rule: every other person can hit forehand winners on the run only under certain conditions. The claim that Williams does not lose her abilities when she is drunk or injured is also odd. She is, let us stipulate, unable to play top-tier tennis when she is drunk. The only reason to believe that she retains her ability to play top-tier tennis is that we expect her to play toptier tennis once she becomes sober. But what if that takes ten years? It may be appropriate to say that she lost the relevant ability during that time span, especially if we are unsure of when she will regain the ability. The tacit appeal to the temporal character of certain limitations on our capacities must be made explicit. Once it is, we can see that it is unmotivated to say that just because an ‘inability’ is temporary it is not an actual inability. Being unable to do something for whatever reason just is an inability. That is all. Some inabilities are transient, and some are lasting. Some are due to temporary or extended losses in cognitive capacity (as in the case of drunkenness), and others are due to technological deficiencies. What follows are a series of examples to illustrate this line of thought. First, suppose one has a neurological deficit that negatively impacts the visual-motor integration typically used for writing. Now suppose that a pen is invented that can make up for this deficit – it’s a prosthetic for the part of the brain too damaged to provide normal visual-motor integration. One suddenly has the capacity to write. If one lacks the capacity to write when one needs a writing implement to make up for a neurological deficit, then one also lacks the capacity to write when one needs a writing implement to make up for a deficit in one’s tools. Compare, for example, the case in which one knows how to write with a ball point pen but does not 214

Material agency

know how to write with a fountain pen, but the latter is the only writing implement around. The lesson to draw from this is that the agential capacity to write is the capacity to use some specific writing technology, or, better yet, the capacity to use some specific technology for the sake of writing. Now, consider the capacity to drive a car. Is this just the capacity to move one’s body in a car-driving fashion? Suppose that one has a car but one has lost one’s legs below the knee. One can no longer move one’s body in a car-driving fashion (even if one knows how to do this, one no longer can do this). But suppose that the car is designed to operate for people who have lost limbs below the knee. The person can move their body in a car-driving fashion but only because the car is designed for that. Does one regain the bodily capacity once the car is designed? Or did one always have that bodily capacity but just had to wait for the new technology to appear? But suppose that we designed a car that an ancient Roman charioteer could drive simply by using typical charioteer skills. Does this mean that 2,000 years ago, these charioteers all had the agential capacity to drive a car that was just awaiting being matched with 21st-century technology? This leads to an unacceptable explosion of an infinity of unknown (and, for the bearer, rationally unknowable) abilities. This may be a bullet someone wants to bite, but it cuts against the grain of how we typically conceive of abilities. We might say of the ancient Romans, ‘They would be able to drive had they had automobiles.’ This subjunctive characterization directly reveals an absence of a capacity (they are not able to drive a car but would be if…) and the condition necessary to realize that capacity (… they had automobiles). As a final example, consider the capacity to walk. Suppose the pathway on which I walk is covered in frictionless black ice. I cannot walk on it. For, I will slip and fall. Even if someone is the world’s foremost expert in walking, she still cannot walk on the frictionless black ice. In this case, the walking expert lacks the capacity to walk on black ice but has the capacity to walk on un-iced ground. This walking expert might actually have a very focused expertise, because she lacks the capacity to walk on a sailboat rolling on sea swells. She has the ability to walk… on dry, un-iced ground. But, the scurvy captain, who has a bad limp on dry land, is as graceful as a ballerina when she walks from the stern to the bow to adjust some line. The agent who can walk gracefully in all these conditions, it turns out, is quite unusual.

4 As is widely accepted in the knowledge-how literature, it is an error to insist that ability persists entirely because whatever internal structures constituting know-how persist. Ginet (1975, 8) and Stanley and Williamson (2001, 416) both offer counterexamples of expert musicians who have lost relevant body parts and so can no longer play their instrument: these people know how to play their instruments but lack the ability to do so. Curiously, no one takes further lesson from this: the abilities are typically realized in structures beyond both the mind and body. For example, one may know how to write, but in the absence of any writing implements, one lacks the ability to write. To insist otherwise by treating knowledge how to write as a ‘general ability’ to write reduces agency to a mental capacity and is therefore subject to the familiar objections to intellectualist versions of know-how. After all, if one knows how to move one’s hand in a writing fashion but cannot write (because one lacks a pen) that is no different than knowing how to move one’s arm in a violin bowing fashion but not being able to play the violin (because one lacks a hand). This paper is not concerned with knowledge how, though, but instead with the possession of a capacity. So, how arguments around intellectualism resolve do not immediately affect the thesis defended here, other than 215

Matthew Noah Smith

that if intellectualism is broadly false, then the account of material agency on offer here is far more plausible. What matters for our purposes is that the anti-intellectualist approach in the knowledge how literature emphasizes the importance of the material component of a capacity for understanding what a capacity is. Even if access to this material component is not a necessary condition for knowing how, it is a necessary condition for the possession of an ability. This raises the specter that there is no such thing as a general ability to act. There is no generic capacity to perform an act-type. Indeed, this is a bullet that I bite. Act-types are neither natural kinds nor social kinds, but are instead abstracta deployed in discourse for a variety of purposes. In particular, when we assert that someone has the capacity to do something, we are making a claim whose content is highly dependent on conversational context. The context provides at least the following relevant information: the range of both normal conditions and tools available. In the tennis case, for example, we might assume normal conditions include being awake, sober, of average health, in decent weather, on an acceptable court, and against an opponent of roughly equal skill; and the normal tools include a normal racket, normal clothing, and normal shoes. If I hit a winner against Serena Williams when she is playing on a court where her side is three times larger than my side, she has a broken racket, and is wearing very painful shoes, we might reasonably refuse to assert that I thereby have demonstrated the capacity to hit winners against Serena Williams. That capacity is realized only on normal courts against opponents wearing normal shoes and playing with normal rackets. In this toy case, these assumptions are quite clear. But it suggests that all ascriptions of an agential capacity presuppose standards for normal realizations of that capacity. These presuppositions, while not explicit, show that there is no generic capacity to perform some action, but instead merely a capacity to do something under certain conditions. Human agency is therefore not the capacity to intentionally move the body (and/or its functional equivalents). Human agency is instead the capacity to use specific material for the sake of some end. Typically, that material is bodily – natural or prosthetic – but it just as often today involves non-bodily technologies: a pen, a car, shoes, a pathway, drugs, and so forth.

5 This account of agency is attractive for at least three reasons. Considering those reasons provides further dialectical support for the view that agency is the capacity to use material or technologies for the sake of some end. First, this account’s plausibility does not rest on the resolution of debates about the nature of basic action. Whatever action is, agency is not, strictly speaking, a capacity to engage in basic action. Rather, agency is the capacity to use material or technologies for some activity. We can be agnostic about the nature of basic action and just plug our best detailed account of using technologies in this theory of agency. That we lack a detailed account of using technologies cashed out in familiar terms of philosophy of action is not evidence that this account of agency as the capacity to use technologies for the sake of some end is not a good account. It is simply evidence of a debit in philosophical ledger. Second, individuation of capacities by technologies and activities allows us to characterize people’s capacities in a socio-historically precise way. This is at least prima facie preferable over characterizations that are insensitive to the significant transformations history brings to how we lead our lives. Any theory of agency that treats such transformations as literally unremarkable is impoverished. This point receives support when we see how the retreat to trans-historically applicable act-types as the objects of agential capacities has unwelcome 216

Material agency

consequences. For example, philosophers of action might treat the specificities of the acttype of driving a car as historically formed when automobiles were invented and then unchanged since then. But driving a contemporary high-end electric car is quite different from driving an average car produced in the 1920s. This is not a point about the phenomenological differences – of which there are surely many – but instead a point about the myriad of other differences in these activities. For example, driving a car without power steering is quite different from driving a car with active steering (this is the increasingly common technology in which the car autonomously corrects for driver steering error). These two driving modalities involve different forms of attention, different bodily movements, different systemic engagements with braking, accelerating, and so on (and so even if we focus only on bodily movements, the bodily mechanics of the two activities are quite different from one another). It’s not even clear that being able to drive a 21st-century car guarantees that one can drive a 1920s-era car, and vice versa. This same point applies, mutatis mutandis, to walking with a prosthesis compared with walking with biologically given limbs, or walking on a narrow slippery, icy mountain path in sleeting rain compared with walking on a wellmaintained, bone-dry city sidewalk on a clear summer day. These are all different capacities to walk. Suppose one group has three of these capacities, and another group has only the one remaining capacity. In some sense, all members of both groups have the capacity to walk, but in another sense they do not all have the same capacity to walk. The theory of human agency on offer here therefore is flexible enough to capture the tremendous heterogeneity of human capabilities in its very characterization of human agency. In particular, built in to any account of human capacities is the theoretical requirement to specify at least some of the conditions of the realization of that capacity. This forces us away from eliding important differences. It also forces us toward the incorporation of importantly relevant features of the human condition into any theory of agency. As such, it avoids what might be an objectionably rationalistic universalism in a theory of human agency, which is after all a material phenomenon, without sacrificing much analytical power. As an illustration of how to do this, let us consider a favorite example of philosophers of action: raising one’s arm. On my account, the agential capacity to raise one’s arm is the capacity to use one’s body for the sake of raising an arm. This may seem pointlessly baroque when compared to the ability to raise one’s arm, but it’s not. For, it immediately gives us the conceptual resources for understanding the ways in which the capacity to use a prosthetic arm for the sake of raising an arm is similar to and different from the capacity to use a biologically given arm for the sake of arm. If the capacity to raise an arm was just the capacity to raise one’s biologically given arm, then simply by definition there is very little similar in and very, very much different between what a person with a prosthetic arm can do and what a person with a biologically given arm can do, namely, the one with the prosthetic can’t do anything arm-centered that a person with a biologically given arm can do. This brings me to the final theoretically attractive feature of this account. Seeing agency as a capacity realized in heterogeneous systems extending beyond the body opens up fruitful lines of inquiry into the provisioning of human agency. This transforms our understanding of agency from something naturally occurring to something that is socially produced, meaning that agency exists because of the collective exercise of human agency. For, the material recruited to realize the end (e.g., the arm that is raised, the car that is driven, tennis racket that is swung) all must be produced and made accessible for a person to have the capacity to use that material for the sake of that end. I cannot have the capacity to play tennis if tennis rackets, tennis balls, tennis courts, and so on have not been provisioned. (I am agnostic here 217

Matthew Noah Smith

about whether one knows how to play tennis in the absence of these materials, but the point here is that the absence of the material makes for the absence of the ability.) By highlighting how all agential capacities are realized in systems, our understanding of human agency ceases to be individualistic and body-focused and instead becomes social and ‘world-focused’. This, in turn, helps us to understand how a person’s relative position in various geographies (spatial, social, economic, etc.) can determine to a great degree what it is she can do. This brings philosophy of action directly into contact with political theory. This is as it should be. For, if we view the absence of extra-bodily means not as a threat to agency but instead as a mere question of means, we might view the absence of such means as a secondary matter in political theory. But the view I’ve offered makes the uneven distribution of technologies a first-order threat to perhaps the most basic commitment of (at least) liberal political theory, namely, a principle guaranteeing equal freedom. For, lacking technologies necessary for some activity just is a first-order inability, that is, a restriction on a person’s freedom. It is not merely an absence of ‘means’. It is akin to being shackled or, perhaps more vividly, to having a limb cut off.

6 Human agency should not be understood as the capacity to act, but instead as the capacity to use technology for ends. Technologies are socially produced material – from the human body to modern infrastructure. Using a technology for an end is not a primitive – we can say a fair bit more about it than has been said here. For example, presumably, a necessary condition for possession of the capacity to use a technology for some end is knowledge how to use that technology for that end. But mere possession of know-how is not sufficient for possession of an agential capacity. Furthermore, we need some theory of access, that is, when does someone have access to a technology and when do they lack it (e.g., does Serena Williams have access to tennis gear and a tennis court if they are all located ten feet from her? What if they are behind a locked gate? what if she is also an expert lockpicker?). These are challenges for future discussions of agency.3

Related entries Agency, skills, and powers; Agency and the body; Agency and disability; Agency and expertise; Agency and mistakes; Bounded agency; Agency and games; Agency and causation.

Notes 1 For recent defenses of this view, see Shildrick (2009) and Shildrick (2015). For the concept of the assemblage see Deleuze and Guattari (1987). 2 The passage Seltzer is commenting on is found in Ford (1923, 108). 3 I am extremely grateful to Luca Ferrero for his generous comments and patience. The ideas in this essay benefited from a long gestation while I was an associate professor in the University of Leeds School of Philosophy, Religion, and History of Science. I am grateful to all my wonderful former colleagues there for valuable conversations about this material, with special appreciation for Pekka Vayrynen, Gerald Lang, Helen Steward, Daniel Elstein, Heather Logue, and Robbie Williams.

Further reading 1 Andy Clark (2008) Supersizing the Mind: Embodiment, Action, and Cognitive Extension (Oxford: Oxford University Press). This book, which provides a sustained argument for decentering our understanding of the mind from biology without abandoning materialism, could be seen as a model for the account of material agency offered in this essay.

218

Material agency 2 Donald Davidson (1980) “Agency,” Chap. 3 in Essays on Actions and Events (Oxford: Clarendon Press). Even if many disagree with the account here, it articulates an understanding of agency that retains its grip on our intuitions. 3 Anton Ford (2018) “The Province of Human Agency,” Noûs 52(3): 697–720. Much of the ingenious and profound argument in Ford’s sophisticated defense of an understanding of human agency as realized in the relation between person and world can be deployed as ground clearing for the account of material agency offered in this essay. 4 Margrit Shildrick (2015) “‘Why Should Our Bodies End at the Skin?’: Embodiment, Boundaries, and Somatechnics,” Hypatia 30(1): 13–29. Shildrick’s essay defends an account of the human body realized in across biological and nonbiological systems, and so invites this essay’s understanding of our agential capacities as realized across those same systems.

References Bratman, Michael. (2003) “Two Problems about Human Agency,” Proceedings of the Aristotelian Society 101(1): 309–326. Deleuze, Gilles and Guattari, Felix. (1987) A Thousand Plateaus: Capitalism and Schizophrenia, translated by Brian Massumi (Minneapolis: University of Minnesota Press). Ford, Henry (1923) My Life and Work (NYC: Doubleday). Frankfurt, Harry (1978) “The Problem of Action,” American Philosophical Quarterly 15(1): 157–162. Freud, Sigmund (1930, 1962) Civilization and Its Discontents, translated by James Strachey (NYC: W.W. Norton). Ginet, Carl (1975) Knowledge, Perception, and Memory (Dordrecht: D. Reidel Publishing Company). Glick, Ephraim (2012) “Abilities and Know-How Attributions,” in Knowledge Ascriptions (Oxford: Oxford University Press): 120–139. Haraway, Donna (1990) “A Cyborg Manifesto: Science, Technology, and Socialist-Feminism in the 1980s,” in Feminism/Postmodernism, ed. Linda Nicholson (NYC: Routledge): 190–233. Korsgaard, Christine (2009) Self-Constitution: Agency, Identity and Integrity (NYC: Oxford University Press). Maier, John (2015) ‘The Agentive Modalities’, Philosophy and Phenomenological Research 90(1): 113–134. Mele, Alfred R. (2003) Motivation and Agency (Oxford: Oxford University Press). Seltzer, Mark (1992) Bodies and Machines (NYC: Routledge). Shildrick, Margrit (2009) “Prosthetic Performativity: Deleuzian Connections and Queer Corporealities,” in Deleuze and Queer Theory, ed. C. Nigianni (Edinburgh: Edinburgh University Press). Shildrick, Margrit (2015) ‘‘Why Should Our Bodies End at the Skin?’: Embodiment, Boundaries, and Somatechnics,” Hypatia 30(1): 13–29. Small, Will (2017) “Agency and Practical Abilities,” Royal Institute of Philosophy Suppl. 80: 235–264. Stanley, Jason and Williamson, Timothy (2001) “Knowing How,” The Journal of Philosophy 98(8): 411–444. Velleman, J. David (1992) “What Happens When Someone Acts?” Mind 101(403): 461–481. Wigley, Mark (1991) “The Disciplining of Architecture,” Assemblage 15: 6–29.

219

PART 5

Agency and knowledge Introduction to Part 5 Luca Ferrero

The practical realm is the pre-eminent domain of agency. The ‘practical realm’ comprises not only actions and activities but also conative attitudes (such as desires and intentions), certain forms of reasoning (such as deliberation), and such notions as accountability, responsibility, voluntariness, intentionality, freedom, deontic concepts, and the good. By contrast, the theoretical domain comprises such things as cognitive attitudes (especially beliefs) knowledge, theoretical reasoning, receptivity, contemplation, and the truth. Does agency have anything to do with the theoretical/cognitive domain? At first, it might seem that it shouldn’t, especially if this domain is conceived in terms of pure passive receptivity or contemplation. But upon reflection, such stark distinction appears unwarranted: there are several places where agency might bear on the theoretical. As a start, consider that agency plays a role in the acquisition, retention, and revision of our beliefs, at least to the extent that beliefs are sensitive to the outcomes of the activity of investigation. Likewise, it is uncontroversial that agents exercise some control and guidance over mental activities that affect cognitive states, attitudes, and processes (if anything, we exercise mental agency in directing our attention and in initiating and sustaining our theoretical reasoning). This does not entail, however, that our believing is something that is or should be under our voluntary control. The central concern is how best to conceive of our responsiveness to reasons (as paradigmatically exhibited in our judgment-sensitive attitudes, such as beliefs and intentions). On the one hand, this responsiveness appears to be a hallmark of full-blooded agency (especially in its rational dimension), on the other hand, the kind of response commanded by reasons appears to be a matter of registering and yielding to them in a passive fashion, to be contrasted with the active character of attitudes and actions that agents can acquire or undertake ‘at will’ by the exercise of a seemingly unfettered freedom. What I have just sketched is a basic tension that underlies the topics of this section (and other related discussions, see especially the chapters on ‘Rational agency,’ ‘Agency and normativity,’ ‘Aesthetic agency,’ and ‘Agency and responsibility’). To begin with, there is the large topic of whether we can legitimately talk of a form of ‘epistemic’ agency. But even those who might be skeptical about the possibility of epistemic agency at large should acknowledge that there are three specific forms of first-person knowledge where agential considerations might prove germane. First, there is the distinctive form of first-person knowledge that the agent seemingly has about what she is intentionally doing as she is doing it. This is the DOI: 10.4324/9780429202131-25221

Luca Ferrero

knowledge that Anscombe famously called ‘practical knowledge’ and that in her view has peculiar epistemic features (including its non-observational and non-inferential character) and that many claim to be at the core of full-blooded intentional agency. Second, there is the issue of the distinctive epistemic status that an agent has over her confidence that she will act as she now intends to do, despite the objective difficulty of the task ahead and the external observer’s epistemically justified pessimism about her eventual success. Finally, there is the question of the distinctive kind of authoritative self-knowledge that a subject has about her own judgment-sensitive attitudes. Does the subject have some kind of authority or agency over these attitudes—an agency that might account for her epistemically privileged position concerning these attitudes? These are the main questions discussed in the following chapters.

Epistemic agency Some philosophers argue that a particular kind of mental agency is exercised in the acquisition and revision of beliefs, given the existence of important similarities between acting intentionally and believing. In ‘Epistemic agency,’ David Hunter provides a survey of the different ways in which the idea of epistemic agency has been defended in the literature and of the problems that arise in connection with each of these proposals. As Hunter argues, it has proven difficult to specify where and how agency is involved in believing. This is not to deny that there are important similarities between believing and acting, but the jury is still out on whether there is such a thing as a genuine epistemic agency. There are several ways to suggest similarities between believing and acting. First, one might argue that believing is an active phenomenon. This, however, requires conceiving of believing as a kind of process or activity rather than a state. Even so, believing does not seem to be an activity that can be done intentionally, instrumentally, or in different modes. These differences suggest that beliefs might belong to a different ontological category than actions and activities. In addition, it is unclear whether we could have the kind of self-knowledge about our beliefs that we appear to have about our own intentional actions. Others have suggested that epistemic agency is a matter of how one makes up or changes one’s mind, especially if one comes to believe something by way of the mental acts of judging or inferring. Yet, this alleged kind of epistemic agency still lacks the ‘at will’ character characteristic of standard instances of intentional agency (which is not to deny that, in cases where evidence is not conclusive, one might acquire a belief voluntarily). Finally, it might be suggested that genuine epistemic agency is exercised prior to judgment and inference, that is, in investigating a subject matter. This is uncontroversial, but it is unclear how this might bear on the agency of believing itself rather than of the activities that lead to the acquisition of beliefs. Hunter claims that what prompts the search for epistemic agency is the idea that there is an ethics of belief that parallels in relevant ways the ethics of action: believing is fundamentally normative in a way that can be vindicated only if believing is an exercise of agency. However, Hunter doubts that the properties of beliefs are goodness fixing. This is not to deny that the goodness of a person partly depends on what they believe. But the same is true about other mental attitudes and emotions, even if they do not involve distinctive kinds of agency.

Agency and practical knowledge When we act intentionally, we seem to have a distinctive kind of knowledge of what we are doing and of why we are doing it. We seem to have this knowledge without observing our conduct or inferring from evidence. Elizabeth Anscombe famously argued that this 222

 Agency and knowledge

knowledge, which she called ‘practical knowledge,’ is constitutive of our full-blooded rational intentional agency (and it is missing in simpler purposive agency). This knowledge is the topic of Kim Frost’s chapter, ‘Agency and practical knowledge.’ In response to some skepticism about the existence of such knowledge, Kim argues that intentional action does necessarily involve some kind of knowledge of what is done, a knowledge that is nonobservational, non-inferential, and spontaneous. There are several alternative accounts of practical knowledge. Empiricist views claim that this knowledge is ultimately observational. Inferentialist accounts, instead, claim that the knowledge is inferred from our knowledge of intention (where the latter is indeed noninferential and non-observational). Kim argues that both empiricist and inferentialist accounts seem to equivocate on the definition of ‘epistemic basis.’ What they offer (whether it be proprioception, a special sense of agency, inference from intention) might well be necessary for practical knowledge, but this does not entail that this basis is available to the agent’s selfconscious awareness and that the agent ordinarily appeals to it in justifying one’s claim to know what one is doing. In addition, both empiricist and inferentialist accounts have trouble explaining the necessity of the connection between intentional action and practical knowledge. The latter is not a problem for doxastic (or cognitivist) accounts, which claim that to intend to do x is, by its very nature, a belief that one is doing x or one is going to do x. Doxastic accounts, therefore, have no trouble accounting for the non-inferential and non-observational character of practical knowledge. However, doxastic accounts (like empiricist and inferentialist views) still take practical knowledge as a kind of theoretical or contemplative knowledge, which is exactly what Anscombe wants to deny. Frost defends the viability of Anscombean views of practical knowledge, in which this knowledge cannot be contemplative, given that its object (the intentional action) depends on its being known (by way of practical knowledge). In the closing section of the entry, Frost briefly considers some important questions about how reason and knowledge relate: how practical knowledge can be both of what one does and of why one does it; whether and how practical knowledge is constrained by theoretical epistemic standards; and whether there is a genus to which both practical and theoretical or contemplative knowledge belong as species.

Agency and evidence Anscombean practical knowledge is the distinctive first-person knowledge that the agent has about what she is intentionally doing as she is doing it. The agent also seems to be in a distinctive epistemic status with respect to the question of whether she will eventually act as she intends to do. As Berislav Marušić and John Schwenkler discuss in ‘Agency and evidence,’ the fact that it is up to the agent to act as she intends to act makes a difference to how the agent should think of the evidence concerning whether she will actually do as she intends to. When the question is considered by a third party, this party estimates the likelihood of future action, taking into account, among other things, how difficult it is for the agent to follow through. The same evidence is also relevant to the agent, who would be mistaken in ignoring it. Nonetheless, given that the action is up to the agent, she is not simply in the position to make a prediction about her future conduct, as a third party would do instead. Marušić and Schwenkler argue that a decision-theoretic approach fails to explain how the agent can take evidence into account in her practical reasoning without making predictions. Additionally, this approach gives rise to an incoherence between our practical and our theoretical conclusions: on practical grounds, the agent reaches a conclusion that she will x, but on purely theoretical grounds, she will reach a conclusion that there is only a certain chance that she will x. 223

Luca Ferrero

Marušić and Schwenkler argue that the proper account requires recognizing that beliefs about matters that are up to the agent can be grounded in practical reasoning. Beliefs of this sort are special in that they are not subject to an evidential norm: they are not meant to reflect a reality that is independent of them. These beliefs are rather meant to bring about the reality they represent. The question of what one is up to is not a question that the agent treats as having an answer to be discovered by one’s reasoning. In other words, it is a matter to be treated as fixed by an observer but not by the agent. When the agent reasons about what she is going to do, the answer is supposed to be settled by the agent’s reasoning, not by some independent considerations that are simply to be reflected in one’s reasoning. This is why it is an error to predict matters that are up to oneself. In predicting them one treats the subject matter as something to be reflected in reasoning rather than to be made the case by one’s reasoning. The fact that the conclusion of one’s practical reasoning is not under a requirement to be supported by adequate evidence does not mean that the agent’s practical reasoning always gives her practical knowledge that she will do what she intends to do. The fact that it might be difficult to act as one intends to act may be evidence that defeats one’s practical knowledge, preventing the agent from knowing that she will act as she intends to. It is usually impossible for an agent to eliminate entirely the possibility of failing to do what she intends. In particular, agents cannot eliminate the need for resolve in sticking to an intention in the face of temptations to abandon it. Hence, as Marušić and John Schwenkler conclude, “The moral is… that, in making up our mind, we sometimes must be prepared to be resolute in the face of difficulty—even when we know that no bookie will think in the way that we do.”

Agency and self-knowledge We often are in a special position to know about our own mental states: to know what we think, feel, believe, desire, or intend in ways that seem inaccessible to other subjects. Is this kind of self-knowledge just a matter of passive receptivity of independently given objects and phenomena? As Brie Gertler discusses in ‘Agency and Self-Knowledge,’ some philosophers argue that the self-knowledge of beliefs and intentions has distinctive agential features. The agentialists, as Gertler calls them, claim that the standard model of knowledge applies to passive mental states, such as sensations and desires, but not to ‘committal’ attitudes, such as beliefs and intentions. The latter attitudes are active mental states that involve exercises of agency in the mode of responsiveness to reasons. In acquiring and revising beliefs and intentions, the weighing of the evidence and the drawing of conclusions are things that the subject does. The acquisition of beliefs and intentions involves effort, and we regard each other as responsible for this effort: we praise or blame people according to whether their beliefs and intentions are well-supported by reasons. Hence, knowing about our beliefs and intentions is not a matter of mere passive observation of our cognitive life. Or so the agentialists argue. We have a distinctive first-personal authority about self-attributions of beliefs and desires. Agentialists argue that this authority should not be accounted for in purely epistemic terms, as a kind of privileged access to passively acquired attitudes. Rather, our expressions of beliefs and intentions are commitment avowals. They convey that we can be held accountable for the avowed attitude. The authority to avow the attitude belongs exclusively to its subject because these attitudes are the results of acts of first-person agency: we have first-personal agential authority relative to beliefs and intentions. According to some prominent agentialists, the distinctive epistemic features of self-knowledge of committal attitudes derive from the subject’s agential authority over them, specifically our authority as rational thinkers. According to other agentialist views, instead, judging that one has a particular belief or intention is 224

Agency and knowledge 

a creative act, which simultaneously expresses self-knowledge and constitutes the commitment to the attitudes in question. In the closing portion of her chapter, Gertler presents some of the main criticisms of agentialism. The main concern is that agentialism is based on an overly idealized picture of human cognition—a picture that does not fully acknowledge that the acquisition of beliefs and intentions is affected by many non-rational factors (such as emotions, mood, and arbitrary features of the context). In addition, the critics of agentialism point out that we often manufacture or selectively attend to evidence that confirms our beliefs. If this is indeed the case, the worry is that, contrary to agentialism, our epistemic authority over some of the contents of our minds might not always be grounded in agential authority.

225

20 EPISTEMIC AGENCY David Hunter

Similarities to action seem to indicate that believing is an exercise of agency. But it has proven difficult to specify precisely where and how agency is involved in believing. And the relevant similarities seem to be present in phenomena where there is no question of agency at all. So it may be best to think of ‘epistemic agency’ as a title in search of a topic more than a name for a widely acknowledged phenomena. Nonetheless, there is much to be learned about belief, action, and ethics by studying these similarities and by asking whether there is any sense in which believing is an exercise of agency.

1 Overview A fan of epistemic agency might hope to locate it in three places. First, she might hold that believing something is itself an active phenomenon. This would contrast with the standard view that to believe something is to be in a certain sort of state, as opposed to performing an action. But one might hold that believing something is more like an activity than a state or an action. And one might question the underlying metaphysical assumptions that frame these alternatives. I will consider this in Section 2. A more promising approach might be to look for agency in how a person makes up and changes her mind. One venerable view is that judging or inferring is a mental act that generates or yields belief. But most also hold that we cannot judge or infer something at will, at least not in the way that we can raise an arm or blink an eye at will. But, somewhat curiously, there is little agreement on why. And some have held that judgment and inference can be voluntary, even if one cannot believe at will. I will discuss this in Section 3. Finally, a fan of epistemic agency might also look for it in the epistemic activities that precede judgment and inference. All sides can agree that investigating something is an activity or an action, though they should also agree that investigating whether, say, the cat is on the mat is one thing and believing that the cat is on the mat is another. But perhaps these various epistemic activities can help account for the similarities between believing and acting. I will discuss this in Section 4. The similarities between action and belief that prompt the search for epistemic agency are clear. There are things a person ought to believe just as there are things she ought to do. People can be assessed based on what they believe just as much as on what they do. And people 226

DOI: 10.4324/9780429202131-26

Epistemic agency

usually know what they believe without having to rely on evidence, just as they usually know what they are doing without having to rely on evidence. These similarities have led some to claim that believing is fundamentally normative. And some say that this normativity can be vindicated only if believing is an exercise of agency. In my view, more work is needed to clarify and support these claims. I will discuss this in Section 5.

2 The ontology of believing What is it for a person to believe something? What ontological category does believing belong to? Everyone can agree that for Sarah to believe that the cat is on the mat is for her to have a certain property, that of believing that the cat is on the mat. But this does not yet tell us much about what ontological category believing belongs to. For one thing, there are different kinds of properties. We need to distinguish sortals from attributes.1 A sortal property is one whose exemplification is an individual substance or stuff. The property of being a mat, of being feline, and of being water is called sortal. We can count the mats and the cats, and we can measure water. An attribute, by contrast, is a property that is instantiated in a substance or stuff but whose instantiations are not themselves substances or stuffs. The properties of being tall, being rich, and being visible are attributes. We cannot count or measure the talls, the riches, or the visibles, though we can count and measure the things and stuffs that are tall, rich, or visible. Are belief properties sortals or attributes? One difficulty is that the English term ‘belief ’ is ambiguous between believing and what is believed. When Sarah believes that the cat is on the mat, there is, on the one hand, the fact that she believes it, and, on the other, the thing or object that she believes. Everyone can agree that the objects of belief are quantifiable. Most hold that they are abstract objects of one kind or another. Arguably, when we count a person’s beliefs, we are counting the things she believes, not her believings of them. Similarly, when we count a person’s possessions, we are counting the things she owns, not her ownings of them. Our question is what ontological category believing fits into, not what ontological category the objects of believing fit into. But keeping this distinction firmly in mind is not easy. It is very common for theorists to say that there are belief ‘tokens.’ This certainly suggests that belief properties are sortals. After all, tokens can be counted. And this is encouraged by (and perhaps encourages) the idea that a person’s beliefs are states inside her central nervous system, states that play a causal role in generating her behavior. Examining this view of belief is beyond my present scope, but it seems to rely on and perhaps encourage treating belief properties as sortals. The distinction between sortals and attributes matters to a study of epistemic agency because actions are countable. On most view, actions are events, and events are exemplifications of sortal properties. So a fan of epistemic agency might say that belief properties are action properties. She might say that believing that the cat is on the mat belongs in the same ontological category as the action of putting the cat on the mat. But there are familiar and serious problems with this suggestion. For one thing, an action usually has a beginning, a middle, and an end. Sarah can be halfway through or finished putting the cat on the mat, but she cannot be halfway through or finished believing that it is on the mat. The believing is fully present as soon as it starts. But a fan of epistemic agency might respond that believing is an activity, not an action. She might say that believing something is like petting the cat. One can start or stop petting the cat, but one cannot be halfway finished petting it, but the petting is fully present as soon as it starts, just like the believing. 227

David Hunter

Still, there are other differences between activities and believing. One can pet the cat quickly or carefully, but one cannot believe something quickly or carefully. What is more, actions and activities can be done intentionally, but believing can’t. Sarah can put the cat on the mat in order to please her mother or because she wants to vacuum the floor. But she cannot believe that the cat is on the mat in order to please her mother or as a means to achieve some other end she has. These differences suggest that believing belongs in a different ontological category than actions and activities.2 One might question the metaphysical framework I have assumed. Matt Boyle (2011b), drawing on ideas he finds in Aristotle, has argued that we need to recognize a category of agency in addition to bodily actions and activities, and that believing is an instance of this kind of agency. One difficulty is that perception also seems to fit that Aristotelian category, or at least Aristotle thought so, but perception does not seem to be an exercise of agency.3 Pamela Hieronymi (2009) goes further than Boyle and questions whether there is such a sharp distinction between states and events. Their views have been influential, but more work is needed to develop the ontological frameworks they rely on. Any account of the nature of believing has to account for the fact that people usually know what they believe and why they believe it without having to rely on observation. Such self-knowledge is also present when a person does something intentionally, and some have suggested that an account of epistemic agency can explain this similarity between believing and acting.4 But debate continues about the origins of this self-knowledge in the epistemic case and about its scope. On some views, believing is essentially self-conscious, whereas others say that it arises through a sort of inference. And it is hard to deny that people can sometimes be in denial about things they believe. Moreover, self-consciousness seems to be present with other mental phenomena, such as perception and proprioception, where there seems to be little or no room for agency. So the links between self-knowledge and agency remain matters of ongoing debate and investigation.

3 Making up one’s mind and direct epistemic control Even if believing something is not itself an active phenomenon, a person is able to make up or change her mind, and this might involve a form of agency. A person makes up her mind when she makes a judgment or draws an inference. But just what is it to make a judgment or to draw an inference, and where could agency fit in to it? For the sake simplicity let’s focus on inference. Usually, when a person draws an inference she starts believing one thing on the basis of or after considering certain other things. Suppose Sarah lifts up the cat’s food bag and after weighing it in her hand concludes that the bag is nearly empty. Everyone can agree that lifting the bag and weighing it in her hand are actions or activities and so exercises of her agency. But what about when she concludes that the bag is nearly empty? Was that an exercise of agency too? Gilbert Ryle (1949) famously mocked the idea that an inference is a mental action.5 He is often accused of being a behaviorist, someone who believes that all mental phenomena are just bodily movements or dispositions. If behaviorism is true, then it is hard to see how inference could be a mental phenomenon at all, since it is plainly not a bodily movement or disposition. I think this accusation is unfair to Ryle, but it is true that he denied that inference is a mental action. If Sarah’s inference were an action, he said, then it would make sense to ask how long the inference took and whether she did it slowly or quickly. But these questions don’t seem to make much sense. So Ryle denied that inferences are mental actions. It is not 228

Epistemic agency

clear what positive proposal he had in mind instead. Some of what he says suggests that an inference is the onset of a state of believing, in something like the way arriving in Toronto is the onset of a state of being in Toronto. If so, then it is hard to see how an inference can be an exercise of agency. But Ryle seems to have assumed that every action is an event, and while this is a standard view about action, it has been challenged.6 Some hold that an action is a causing and that causings are not events. The metaphysics here are subtle and controversial, but what matters is that if actions are not events, then this might explain why the questions Ryle raised about inference make no sense. If inferences are not events, then they have no duration and cannot be fast or slow. Still, more work is needed to clarify the metaphysics of action before a fan of mental agency can be content with this view of inference. Nevertheless, the view that inference is an action faces other difficulties. As we saw, actions can usually be intentional, but it is hard to see how a person can draw an inference in order to please her mother or as a means to some other end. Moreover, if inference is an action, it would be a basic action, in the sense that it is not done by or in doing something else. Sarah moved the cat by lifting it off the mat. She did the first by doing the second. But she did not draw that inference by doing something else. An inference would be more like blinking or wiggling a finger, a basic action that does not require doing something else. The distinction between basic and non-basic action matters here because, in standard cases anyway, a person can perform a basic action at will. Sarah can blink at will, just like that. But it is hard to see how an inference can be done at will. One cannot simply infer at will that some proposition is true.7 (Nor can one believe at will.) An inference, it seems, needs to be done on the basis of other things or in light of other considerations. This would not entail that an inference is not an action, but it would entail that an inference is unlike familiar kinds of action, and this in turn might raise doubts about the alleged similarities that encourage the idea that believing involves epistemic agency. We should distinguish whether an action is done at will from whether it is done voluntarily. The distinctions here are subtle and may be mostly terminological. Roughly, though, whether an action is voluntary depends on whether the agent had alternatives, whether she could have done something else, and this seems to be independent of whether what she did is something she could do at will. Sarah cannot lift the cat off the rug at will, but she can do it voluntarily. This matters for the study of epistemic agency because some have recently argued that inference can be voluntary, even if it cannot be done at will. They have in mind cases where a person’s evidence for some proposition is strong but not conclusive. It would be reasonable for her to believe it, but also reasonable for her to remain agnostic. When Sarah raises the food bag and feels its weight, it would be reasonable for her to infer that it is nearly empty but also reasonable for her to remain undecided on the matter. This is a matter of some controversy, and more work is needed to clarify the sense of voluntariness at issue.8 But this is important since whether an action is voluntary is relevant to an assessment of the person who did it. If inference can also be voluntary, then this might help explain why we assess people on the basis of the inferences they make.

4 Epistemic activity and indirect epistemic control One might thus question whether either believing or inferring is an exercise of agency. But everyone must agree that investigating is an exercise of agency. When Sarah sets about collecting evidence on whether she needs to buy more cat food, she does a variety of things that concluded with her belief that the bag was nearly empty. Perhaps her believing that is 229

David Hunter

not an exercise of agency. And perhaps her inferring it from her evidence is not an exercise of agency, either. But the things she did in order to collect the evidence surely are actions or activities, done intentionally, with the purpose or aim of getting to the truth of the matter. In this sense, they constitute indirect epistemic activity. Importantly, the ethical features that originally prompted a search for epistemic agency are present in epistemic activities, such as investigation. It can be true of a person that she ought to investigate some matter. Sarah ought to check whether she needs more cat food, since the store is closing soon and the cats are hungry. And we can assess the quality of a person’s epistemic investigations. How thoroughly did Sarah check for food? Did she also look in the closet where an extra bag is sometimes kept? Indeed, we assess a person’s quality based on how she conducts her investigations. For, arguably anyway, a good person is also a good critical thinker. A good person doesn’t jump to conclusions, keeps an eye out for contrary evidence, and does not let her emotions sway or cloud her investigations.9 But none of this is very surprising or controversial, since everyone already agrees both that epistemic inquiry is an activity and that a person can be assessed based on her actions and activities. The original search for epistemic agency was prompted by the idea that there are ethical similarities between action and believing itself, not just between actions and the things we do in conducting investigations that lead to believing. There are things that a person ought to believe, not just things she ought to investigate. And what a person believes is relevant to our assessment of her, not just to an assessment of her competence as a critical thinker. It is not clear that we can make sense of these ethical features of belief wholly in terms of the ethical features of epistemic activities like inquiry.

5 Ethics of belief The search for a distinctive form of epistemic agency is prompted by the idea that believing is subject to the same sorts of ethical concepts as intentional action. That is, it is prompted by the idea that there is an ethics of belief that mirrors in relevant ways an ethics of action. It is worth ending by considering how similar these ethics are. Earlier we saw that some properties are sortal properties. Being a chair, being a feline, being a person are sortal properties. And some sortal properties are goodness fixing properties, in the sense that they set a standard for when a thing is a good example of that kind. Some chairs are better chairs than others. Some cats are healthier cats than others. Some people are better people than others. Some actions types are also goodness fixing in this sense. One tennis serve can be a better tennis serve than another. And, as we saw, one investigation can be a better investigation than another.10 But it is not clear that belief properties are goodness fixing properties. Is there such a thing as a better or worse case of believing that the cat is on the mat? Or is any case of it just as good, qua believing that the cat is on the mat, as any other? If there can’t be better or worse cases of believing, then belief properties are not goodness fixing. Arguably, the property of being an action is also not a goodness fixing property, since no case of action can be better qua action than any other. But what if the belief is true? Or what if it is rational? Wouldn’t that make it a better case of believing? Many have said that believing ‘aims’ at truth or rationality, and that this is ‘constitutive’ of believing. And some have then sought to ground an ethics of belief on this sort of fact.11 But so far as I can see, whether a person is right or reasonable in believing something

230

Epistemic agency

makes no difference to whether their believing it is a better case of believing it than another. Belief properties, it seems to me, are not goodness fixing properties. Still, even if belief properties are not goodness fixing properties, whether a person is good in one or another way does depend on what she believes. A good doctor knows, and so believes in, the latest medical information. A good student knows, and so believes in, the relevant rules on academic misconduct. Arguably, whether a person is a good person also depends on whether she knows (and so believes) the things that she ought to know (and so believe). Facts about a person’s beliefs can, in this way, be to her credit or discredit. And when a person fails to believe something she ought to believe, and this is through her own negligence, then she may be to blame for not believing it.12 Importantly, these similarities are present in other mental phenomena where there is little or no appearance of agency. As just noted, there are things a person ought to know, and what a person knows (or does not know) can be to her credit or discredit. Likewise, a person should sometimes be ashamed and it can be to her credit when she is. The same holds for hoping, admiring, respecting, and a wide variety of other mental attitudes and emotions. It seems unlikely that each of these also involves an exercise of a distinctive form of agency. But all of this is a rich subject for further investigation.

Related topics Mental agency; Rational agency; Agency and practical knowledge; Agency and normativity.

Notes 1 This distinction and its importance for the philosophy of mind have been stressed most recently in (Steward 1997), (Marcus 2009), and (Marcus 2014). Their work echoes views in (Ryle 1949) and (Kenny 1989). I discuss all of this in more detail in (Hunter 2018a). 2 For more discussion, see (Chrisman 2016) and (Setiya 2008). 3 It is another matter whether perception is an exercise of a cognitive power or capacity. For discussion, see (Hyman 1994), who also counts inference as an exercise of a cognitive power. 4 An influential discussion of agency and self-knowledge is (Anscombe 1957). (Hintikka 1962) is an important work on self-knowledge in formal epistemology. Recent work on self-knowledge and believing traces to (Burge 1998) and (Moran 2001). (Roessler & Eilan 2003), (O’Brien 2007), (Boyle 2011a), (Rödl 2007), and (Byrne 2018) contain excellent discussions. 5 For discussion see (White 1971). 6 For the standard view, see (Davidson 1984). For some of the challengers, see (Bach 1980), (Thomson 1977), and (Alvarez and Hyman 1998). (Broome 2013) and (Boghossian 2014) are influential recent discussions of inference. 7 A thorough history of views on believing and inferring at will is in (Pojman 1986). Contemporary discussion in (Hieronymi 2009) and (Setiya 2008) of believing at will is influenced by (Williams 1973) and (Bennett 1990). Excellent discussions can also be found in (Winters 1979) and (Govier 1976). 8 For more on voluntariness, see (Hyman 2015). For more on whether inference can be voluntary, see (Govier 1976) and, more recently, (Nickel 2010), (McHugh 2014), (Roeber 2019), and (Hunter 2022). For discussion of the epistemic matters underlying this, see (White 2005). 9 For discussion of this, see (Montmarquet 1993), (Zagzebski 1996), (Owens 2000), and, more recently, (Sosa 2009) and (Cassam 2019). 10 This paragraph develops ideas in (Thomson 2008). 11 Influential discussions are in (Williams 1973), (Feldman 2000), and (Shah 2006). For critiques, see (Steglich-Petersen 2007), (McCormick 2015), and (Hunter 2018b). 12 For recent treatments of epistemic blame and responsibility, see (Meylan 2015) and (Peels 2016).

231

David Hunter

Further reading Peels, R. 2016. Responsible Belief: A Theory in Ethics and Epistemology. Oxford: Oxford University Press. This recent book provides an excellent discussion of the various notions of responsibility and accountability. While it is focused primarily on believing, it offers valuable insights into the similarities between action and believing. Pojman, L. 1986. Religious Belief and the Will. London: Routledge, Kegan, and Paul. This book provides an excellent survey of the history of discussions by philosophers of doxastic freedom. Thomson, J. 2008. Normativity. Peru, IL: Open Court. This book is primarily about the nature of normativity, a branch of meta-ethics. But it offers an excellent introduction to the complex issues involving responsibility, blame, and goodness. It also includes excellent discussions of epistemic duties. Williams, B. 1973. Deciding to believe. Chap. 9 in Problems of the Self. Cambridge: Cambridge University Press, 136–151. This article has been enormously influential. It discusses the nature of believing, whether believing is essentially normative, the nature of inference, and whether a person can believe at will. Just about all contemporary discussions of epistemic agency trace back to this article.

References Alvarez, M. and Hyman, J. 1998. Agents and their acts. Philosophy, 73, 219–245. Anscombe, G.E.M. 1957. Intention. Second Edition. Cambridge, MA: Harvard University Press. Bach, K. 1980. Actions are not events. Mind, 89(353), 114–120. Bennett, J. 1990. Why is belief involuntary? Analysis, 50, 98–107. Boghossian, P. 2014. What is inference? Philosophical Studies, 169, 1–18. Boyle, M. 2011a. Transparent self-knowledge. Proceedings of the Aristotelian Society, supplementary volume LXXXV, 223–241. Boyle, M. 2011b. “Making up your mind” and the activity of reason. Philosopher’s Imprint, 11(17), 1–24. Broome, J. 2013. Rationality through Reasoning. Malden, MA: Wiley-Blackwell. Burge, T. 1998. Reason and the first person. In C. Wright, B. Smith & C. MacDonald (Eds.), Knowing Our Own Minds (pp. 243–271). Oxford: Oxford University Press. Byrne, A. 2018. Transparency and Self-Knowledge. Oxford: Oxford University Press. Cassam, Q. 2019. Vices of the Mind. Oxford: Oxford University Press. Chrisman, M. 2016. Epistemic normativity and cognitive agency. Noûs, 52(3), 504–529. Davidson, D. 1984. Essays on Actions and Events. New York: Oxford University Press. Feldman, R. 2000. The ethics of belief. Philosophy and Phenomenological Research, 60(3), 667–695. Govier, T. 1976. Belief, values, and the will. Dialogue, 15, 642–663. Hieronymi, P. 2009. Believing at will. Canadian Journal of Philosophy, supplementary volume 35, 135–187. Hintikka, J. 1962. Knowledge and Belief. Ithaca, NY: Cornell University Press. Hunter, D. 2018a. The metaphysics of responsible believing. Manuscrito, 41(4), 255–285. Hunter, D. 2018b. Directives for knowledge and belief. In D. Whiting et al. (Eds.), Normativity: Epistemic and Practical. Oxford: Oxford University Press: 68–89. Hunter, D. 2022. On Believing: being right in a world of possibilities. Oxford: Oxford University Press. Hyman, J. 1994. Vision and power. The Journal of Philosophy, 91(5), 236–252. Hyman, J. 2015. Action, Knowledge and Will. Oxford: Oxford University Press. Kenny, A. 1989. The Metaphysics of Mind. Oxford: Oxford University Press. Marcus, E. 2009. Why there are no token states. Journal of Philosophical Research, 34, 215–241. Marcus, E. 2014. Rational Causation. Cambridge, MA: Harvard University Press. McCormick, M. 2015. Believing against the Evidence. New York: Routledge. McHugh, C. 2014. Exercising doxastic freedom. Philosophy and Phenomenological Research, 88(1), 1–37. Meylan, A. 2015. The legitimacy of intellectual praise and blame. Journal of Philosophical Research, 40, 1–15. DOI:10.5840/jpr201511537. Montmarquet, J. 1993. Epistemic Virtues and Doxastic Responsibility. London: Rowman & Littlefield Publishers. Moran, R. 2001. Authority and Estrangement. Princeton, NJ: Princeton University Press.

232

Epistemic agency Nickel, P. 2010. Voluntary belief on a reasonable basis. Philosophy and Phenomenological Research, 81(2), 312–334. O’Brien, L. 2007. Self-knowing Agents. Oxford: Oxford University Press. Owens, D. 2000. Reason without Freedom. London: Routledge. Peels, R. 2016. Responsible Belief: A Theory in Ethics and Epistemology. Oxford: Oxford University Press. Pojman, L. 1986. Religious Belief and the Will. London: Routledge, Kegan, and Paul. Rödl, S. 2007. Self-consciousness. Cambridge, MA: Harvard University Press. Roeber, B. 2019. Evidence, judgment, and belief at will. Mind. DOI:10.1093/mind/fzy065. Roessler, J. and Eilan, E. (Eds.). 2003. Agency and Self-Awareness. Oxford: Oxford University Press. Ryle, G. 1949. The Concept of Mind. New York: Barnes & Noble. Setiya, K. 2008. Believing at will. Midwest Studies in Philosophy, 32, 36–52. Shah, N. 2006. A new argument for evidentialism. Philosophical Quarterly, 56 (225), 481–498. Sosa, E. 2009. A Virtue Epistemology Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. Steglich-Petersen, A. 2007. Against essential normativity of the mental. Philosophical Studies, 140, 263–283. Steward, H. 1997. The Ontology of Mind. Oxford: Oxford University Press. Thomson, J. 1977. Acts and Other Events. Ithaca, NY: Cornell University Press. Thomson, J. 2008. Normativity. Peru, IL: Open Court. White, A. 1971. Inference. Philosophical Quarterly, 21(85), 289–302. White, R. 2005. Epistemic permissiveness. Philosophical Perspectives, 19, 445–459. Williams, B. 1973. Deciding to believe. Chap. 9 in Problems of the Self. Cambridge: Cambridge University Press, 136–151. Winters, B. 1979. Believing at will. Journal of Philosophy, 76, 243–256. Zagzebski, L. 1996. Virtues of the Mind. Cambridge: Cambridge University Press.

233

21 AGENCY AND PRACTICAL KNOWLEDGE Kim Frost

1 Introduction Mature agents seem to first-personally know what they are doing and why, at least to some extent, when they act intentionally. I know, for example, that I’m typing in order to write about practical knowledge. I don’t seem to need to look and see to know this, as I might were I to investigate what you’re doing. Nor do I need to theorize, as I might were I to guess the cause of some spasm. I just know, apparently without needing to perceive or theorize to work it out. Call this practical knowledge. Practical knowledge is philosophically important for three reasons. First, practical knowledge seems partly constitutive of rational intentional agency. Nonrational animals act purposively, but they aren’t sophisticated enough to form complex, abstract, first-personal, linguistically expressible conceptions of what they are doing and why. By contrast, I know what I’m doing, and why, and I can tell you: I’m writing about practical knowledge, to fulfill a contractual obligation, and, ultimately, for the sake of education. Outside philosophical contexts, no one seriously doubts these claims to knowledge. What’s more, if I don’t know what I’m doing, it’s plausible that I’m not acting intentionally at all. If I were sleep-typing this, say, one could probably explain what I do by appeal to my psychological states, skills, circumstances, and unconscious mechanisms that bring these things together, perhaps in an expected way. But the explanandum wouldn’t be an intentional action, precisely because I fail to practically know what I do and why. (The problem is not just that I am asleep; see Setiya 2008: 389, fn 7.) If I were carefully stacking bottles and suffering from schizophrenia, and felt that someone is stacking bottles but didn’t know it was me, it would similarly be implausible that I am acting intentionally. Intentional action exhibits self-consciousness that is absent in unsophisticated animals, or even sophisticated animals that are sleep-typing or suffering from certain mental disorders, and the term ‘practical knowledge’ refers to a crucial aspect of that self-consciousness. Second, practical knowledge may refute empiricism, that is, the claim that all knowledge concerning particular matters of fact ultimately has its epistemic basis in a receptive occurrence called experience. Practical knowledge doesn’t seem epistemically based on experience. If there is practical knowledge, but it is spontaneous rather than receptive, then empiricism is false.

234

DOI: 10.4324/9780429202131-27

Agency and practical knowledge

Third, practical knowledge may embody moral knowledge. In knowing myself to pursue the noble end of education, say, I seem to know myself as a moral agent. Most contemporary literature on practical knowledge says little about this possibility. I only say a little at the end. Anscombe (1963) did the most to introduce practical knowledge into contemporary discussions (see also Hampshire 1959). I focus on her view and reactions to it. Practical knowledge is fiercely complex, so I focus on the first two reasons why it is important, and skip many rich debates about details. In this opinionated introduction, I also argue that some objections to Anscombe’s view are mistaken. Some terminology: by ‘self-consciousness’ I mean knowledge (awareness) of oneself as oneself. I sometimes use, for example, ‘know’ instead of ‘self-know’, for the sake of elegance.

2 Skeptical reactions Anscombe (1963: 50) says that practical knowledge is of what one does (intentionally, on an occasion). Anscombe (1963: 87) also says that it provides the descriptions under which what is done is intentional. This implies a necessary connection between intentional action and practical knowledge. If someone acts intentionally, then she practically knows what she does, else nothing would provide the descriptions under which what she does is intentional. Skeptical reactions to this claim are common because of a famous counterexample. Davidson (2001: 50) imagines someone making ten carbon copies intentionally. It’s a difficult task, so he needn’t believe, and may well doubt, that he is succeeding. Although he is intentionally making ten carbon copies, he doesn’t know what he does under the description: ‘I’m making ten carbon copies’. The counterexample can be resisted. Thompson (2011: 210) argues that ordinarily – when signing documents, say – one checks and keeps going until one does things right. In extraordinary cases there’s no opportunity for this, but then it’s unclear whether the act is intentional, because success is so lucky, akin to winning a lottery. So arguably the object of the agent’s doubt is not whether he is making ten carbon copies, broadly speaking, but whether he has succeeded already. A more radical response is to question the assumption that all knowledge requires belief or confidence. Anscombe (1963: 57) thinks that practical knowledge is distinct from contemplative knowledge, which depends on agreement with facts that obtain independently of being known contemplatively for its status as knowledge. Unlike contemplative knowledge, Anscombe (1963: 87) thinks that under certain conditions, practical knowledge is ‘the [formal] cause of what it understands’. Anscombe (1963: 56–57) also thinks that if purported practical knowledge and its object don’t agree, and this alone constitutes a mistake, the mistake is in the agent’s performance, and not in some judgment. Perhaps Davidson’s counterexample shows that contemplative knowledge is not required for acting intentionally. But it seems question begging to directly extend that conclusion to Anscombean practical knowledge, which may have different relations to mistakes, belief, and confidence (see below). Davidson himself thought that his counterexample had limited force. For Davidson (2001: 50), ‘[a]ction does require that what the agent does is intentional under some description, and this in turn requires … that what the agent does is known to him under some description.’ So the carbon copier knows what he does in some way: he knows he is pressing hard on the paper, say. Hence, we should carefully distinguish two different necessary connections: GENERAL: If I act intentionally, I know what I do (somehow). SPECIFIC: If I am A-ing intentionally, I know that ‘I am A-ing.’

235

Kim Frost

Davidson’s example only challenges Specific; indeed Davidson endorses General. The distinction is important because Davidson’s example is sometimes treated as a reason to ignore the purportedly necessary self-consciousness of intentional action. If General is true, that attitude is misguided. Although it doesn’t challenge General, a common reaction to Davidson’s example is to retreat from General and substitute a weaker link between intentional action and, for example, reliable (not necessary) knowledge; or necessary belief (not knowledge); or knowledge of what one merely intends (not what one does). Examples of such retreat are given below. What is gained by such retreat? In one sense, retreat from General is dialectically puzzling. To my knowledge, no one has produced an excellent counterexample to General. General is a very weak claim, because it’s so unspecific. General says nothing about how much knowledge is required, or how many kinds of knowledge there are, or whether contemplative knowledge is the only kind. Examples like Armstrong’s (1968) absentminded driver do nothing to upset General. If the driver has no knowledge of what she does – not even that expressible by ‘Hmm? Oh, I’m driving. Sorry, my mind wanders.’ – then plausibly no intentional action is in view. (We are back to sleep-driving.) Setiya’s (2008) agent, who is recovering from paralysis and tries to clench her fist, and does so but has little confidence that she will succeed, plausibly knows that she is testing her medical condition. Besides, the example only upsets General if such knowledge always requires belief or confidence. Setiya does not argue that it does; he assumes it. If it doesn’t, and the connection of self-awareness to action here is completely accidental, then again, plausibly, no intentional action is in view. (We are back to winning lotteries.) General is, I think, quite plausible. Where every relevant non-accidental union of self-awareness and action fails (e.g. sleep-typing; severe schizophrenia; outrageous luck), it seems like there is no intentional action. But if General is true, retreat from it only defers problems. One cannot easily explain necessary knowledge of what one does with materials that are defined as weaker than that; mere conjunction of those things with extra conditions in principle allows room for Gettier-style counterexamples to creep in (or their analogues, for non-propositional accounts of the relevant knowledge). Even if General overgeneralizes, one must still explain why agents usually know what they do when they act intentionally, what that knowledge is, and why it seems spontaneous. In another sense, retreat from General is not dialectically puzzling. If General is true, and the relevant knowledge spontaneous, then there is, as Rödl (2007: 121) puts it, ‘spontaneous knowledge of a material reality’. This can sound like a divine power, and not something possessed by mere humans. Davidson’s example may then serve as an indirect expression of proper modesty and a deep skepticism about the possibility of spontaneous (human) knowledge of a material reality. I suspect a heady mix of empiricist assumptions, ‘dualist hangovers’, and overly abstract approaches to propositional attitudes explains why Davidson’s example is often thought to demolish the necessary connection between intentional action and knowledge. Let’s consider empiricist views.

3 Empiricist views Anscombe (1963: 50) says that practical knowledge is non-observational (i.e., not epistemically based on observation) and non-inferential (i.e., not inferred from evidence). Anscombe (1963: 13) also introduces practical knowledge by comparing it to non-observational knowledge of 236

Agency and practical knowledge

the position of one’s limbs. This invites an objection from anyone who thinks that proprioception is the basis of knowledge of the position of one’s limbs. If this is observational knowledge, why couldn’t practical knowledge be similarly observational? Ginet (1990) says voluntary actions have an ‘actish phenomenal quality’. Why couldn’t something like that be the epistemic basis of practical knowledge? Practical knowledge would then be based on experience, and empiricism would not yet be refuted. Examples of empiricist views are O’Brien (2003) and O’Shaughnessy (2003). Such views face considerable difficulties. It doesn’t seem like experience is the epistemic basis of my knowledge of what I’m doing; pre-theoretically, the presumption is against an empiricist account. One dialectical difficulty then is to explain why the posited experience must ground the agent’s practical knowledge, without begging the question of whether empiricism is true. If it is an open question whether empiricism is true, we can imagine creatures who reasonably credit themselves with practical knowledge, but who don’t (need to) appeal to experience to ground their claims to know what they are doing and why. If they can know themselves that way, then why can’t we? Another difficulty concerns necessary connections. If General is true, intentional action can’t exist without any knowledge (however dim or partial) of that action. But as Setiya (2008: 399n38) notes, receptive occurrences, like perception and its analogues, are contingent in relation to their objects; the objects are there, and have the nature they do, independently of being experienced. Why should it be, then, that there is a necessary – not merely reliable – connection between acting intentionally and knowing what one does? There are many possible empiricist accounts, so I cannot canvass all the ways they might respond. In relation to Anscombe, the empiricist objection clearly misses its target. Anscombe (1963: 13) thinks one has observational knowledge when one (correctly) employs appearances as one’s ‘criterion’ for judgment. For example, I might judge that someone is sneaking about because the shadows look that way. By definition, there is room to meaningfully disagree about whether a criterion supports a contrasting judgment. So for Anscombe, to say that practical knowledge is non-observational is, roughly, to say that it is not inferred abductively from observational evidence. The empiricist account clearly lacks that kind of judgment. One doesn’t usually judge one’s arm is down on the basis of a criterion; similarly for what one does. Proprioception might be required for knowing where my arm is. It might help explain why I am so reliable in my judgments. It might be what I appeal to in a few outré cases, or what I guess I should appeal to, after being pressured to confabulate an epistemic basis. But it is not what I always appeal to when justifying my claim that my arm is down. Ordinarily, I don’t appeal to anything, and nobody thinks I must. The way the empiricist objection misses its target indicates that any account of practical knowledge must clearly define terms like ‘epistemic basis’. Anscombe thinks an epistemic basis is something which is in principle available to the agent’s self-conscious awareness (perhaps after reflection) and which she can and does deploy in what Sellars (1997) calls the ‘space of reasons’: the language game of giving, asking for, and criticizing reasons for what one does and what one thinks is true. So Anscombe’s view seems prima facie compatible with the claim that proprioception, or even a special ‘sense of agency’, is required for knowledge of the position of one’s limbs, or what one is doing and why. It’s just that one doesn’t usually (need to) appeal to these things in the space of reasons. That’s what she means by saying practical knowledge is non-inferential (and non-observational). Different definitions of terms like ‘epistemic basis’ correspond to different explanatory projects. To see why, let’s consider inferentialist views. 237

Kim Frost

4 Inferentialist views The leading contemporary inferentialist view of practical knowledge is Paul (2009) (which revises and refines a view in Grice 1971). Paul argues that practical knowledge is evidentially based on knowledge of intention. Paul thinks knowledge of intention is non-observational, and so she agrees with Anscombe that practical knowledge is also non-observational. But she also thinks she disagrees with Anscombe in thinking that practical knowledge is inferential, because it is inferred from evidence about intention. Paul doesn’t obviously disagree with Anscombe, because Paul has different definitions of inference and evidential basis. Paul (2009: 10) says inference is ‘non-conscious information processing [where] beliefs resulting from such a process are derived from prior states that (ideally) evidentially support those beliefs’. Paul’s ‘inference’ and ‘evidential support’ are conceived as largely external to what the self-conscious subject has, and does, in the space of reasons. Paul (2009: 10) acknowledges that some might balk at calling her information processing inference, because it requires no self-conscious awareness of the evidential relation between the terms of the so-called inference. She says this is a labeling dispute. But if so, Paul misrepresents her opposition to Anscombe. As with proprioception, it seems prima facie compatible with Anscombe’s view that such unconscious information processing and relations of evidential support, externally conceived, are required for practical knowledge: it’s just that one doesn’t (need to) appeal to such things in the space of reasons. So for all Paul has said, Paul’s practical knowledge is non-inferential, in Anscombe’s sense. Relatedly, it’s no objection to Paul’s view that practical knowledge doesn’t seem inferential. Self-conscious seemings are cheap when doing speculative psychology about unconscious information processing. This dialectical confusion points to a difference in explanatory projects. Paul’s main question is not ‘What is practical knowledge?’ but rather ‘How does practical knowledge work?’ The answer to this question aims at a relatively external ‘engineer’s understanding’ of the topic, typical of many projects in naturalized epistemology and philosophy of mind. Unless one can prove that this is the only good way to understand or specify the nature of knowledge (inference, evidence, etc.), we don’t have a clear disagreement on whether practical knowledge is inferential in Anscombe’s sense. Paul’s view faces considerable difficulties. Given that the relevant processes are defined in terms of operations on belief – not knowledge – she seems committed to a conjunctive account of practical knowledge: practical knowledge is belief conjoined with extra conditions. So this view may confront Gettier-style counterexamples when completing the (deferred) task of explaining what practical knowledge is. Paul also faces problems with necessary connections. As Setiya (2008: 394–395) notes, the kind of inference Paul invokes, like perception, is contingent in relation to its objects: the objects could exist without the inference. Paul (2009: 5–6) acknowledges this, and explicitly rejects Specific, but doesn’t explicitly consider General as such. If General is true, there is still an unexplained necessary connection between acting intentionally and knowledge. To the extent that the ‘engineer’s understanding’ only provides reliable contingent connections, explaining that necessary connection requires a different kind of account. Interestingly, Paul (2009) allows necessary connections between an object of knowledge and knowledge of that object, but only in a restricted domain. Paul (2009: 4) says her ‘basic strategy is to push any groundless component of our knowledge of our own intentional actions back into the head, while still capturing the sense in which that knowledge is non-observational’. In Paul (2009), the groundless component is knowledge of intention; in Paul (2012) and Paul (2015) the groundless component is only knowledge of conscious 238

Agency and practical knowledge

decision. Either way, I think the basic strategy betrays a ‘dualist hangover’. Many still feel the after-effects of substance dualism – even those who swear off drinking the stuff forevermore. Here the hangover is felt in the idea that there are different rules for spontaneous knowledge of material and (some) mental realities. Paul (2009: 1) drily lampoons spontaneous knowledge of a material reality as something that only belongs to God, if anything. But Paul accepts spontaneous knowledge of a mental reality like intention, or at least decision. On this view, we humans are divine, in our small way: we are little gods of a small corner of the latter-day physicalist analogue of the Cartesian theater. Such ‘different rules’ for spontaneous knowledge might make sense for substance dualists; it’s unclear how they make sense for physicalists. For those who believe in such ‘different rules’ for spontaneous knowledge, it’s natural to be mystified by apparently spontaneous practical knowledge. Indeed, Paul (2009: 22) says one of the great advantages of her view is that it demystifies practical knowledge, by clarifying how it is a species of contemplative knowledge, derived from evidence in the usual ways. But there are ways to demystify practical knowledge that don’t involve nursing a dualist hangover. Let’s consider doxastic views.

5 Doxastic views The leading contemporary doxastic view of practical knowledge is Setiya (2008) (which revises and refines a view in Velleman 1989). Setiya (2008: 395) argues that ‘to intend something is, in part, to [wholly or partially] believe that one is doing it or that one is going to’. Practical knowledge, when one has it, either is that belief, or at least necessarily involves that belief somehow. (Doxastic views are commonly called ‘cognitivist’; the term assumes cognition implies belief.) Setiya also argues that practical knowledge is non-inferential and non-observational. There is then a question of why such belief isn’t irrational wishful thinking, as it isn’t based on sufficient evidence. Setiya says that knowing how to perform the relevant act entitles one to the relevant belief, so the belief has a non-inferential epistemic basis. Setiya rejects General, but he accepts something like it: namely, a necessary connection between acting intentionally and (partial) belief. So his account does not face the problem about necessary connections which faces empiricist and inferentialist accounts. Contingent connections cannot explain a necessary connection. Setiya explains his necessary connection with a theory about the nature of intention: namely, that intention necessarily involves belief. Here we see how the truth of General, or something like it, dictates the form of an adequate account. Setiya accepts the spontaneity of practical knowledge, so he needn’t posit ‘different rules’ for spontaneous knowledge of mental and material realities. Even so, he still thinks practical knowledge needs demystification. His central question is: ‘Why is practical knowledge rational, given that it is belief without sufficient evidence?’ The urgency of the question implies that beliefs, generally, are subject to norms of theoretical rationality. Setiya’s posit of know how, as an epistemic basis, explains why some beliefs satisfy norms of theoretical rationality (excepting the norm of sufficient evidence), not why they are exempt from them. If theoretical knowledge is propositional knowledge subject to norms of theoretical rationality, then Setiya thinks practical knowledge is a species of theoretical knowledge. To the extent Setiya thinks intention could cause intentional action without corresponding knowledge, his practical knowledge is also a species of contemplative knowledge, because such knowledge is beholden to a prior fact which obtains independently of being known. 239

Kim Frost

One of Setiya’s (2008: 396) main reasons for thinking intention involves belief is that intention is assertorically expressible; one usually expresses intentions by saying what one is doing, or will do, and why. I think this betrays an overly abstract approach to propositional attitudes. To see why, let’s consider Anscombean views.

6 Anscombean views Anscombe (1963: 57) says modern philosophy is largely blind to practical knowledge, because it has an ‘incorrigibly contemplative’ conception of knowledge. Empiricist, inferentialist, and doxastic views tend this way: they tend to think practical knowledge must be a (peculiar) species of theoretical or contemplative knowledge. Call an Anscombean view of practical knowledge any view which denies that practical knowledge is a species of theoretical or contemplative knowledge. How could practical knowledge fail to be theoretical? Given my definition, it would fail if it were non-propositional. Campbell (2018a) argues that practical knowledge is grounded in a mental state of intention, whilst theoretical knowledge is grounded in a mental state of belief. Campbell (2018b) also argues that intentions are attitudes to ‘do-ables’ rather than propositions. Similarly, Frost (2018) argues that practical knowledge isn’t propositional knowledge, though it is linguistically expressible. Either way, practical knowledge isn’t a species of theoretical knowledge. If so, then overly abstract approaches to propositional attitudes may make a subtle mistake. Call the unity of something its real metaphysical structure. Campbell thinks practical knowledge has the unity of an attitude to a ‘do-able’. Frost (2018: 19) thinks practical knowledge has the unity of a practical syllogism. Both think that the unity of something needn’t be isomorphic to the unity of assertions that express that thing. The mistake is to move directly from assertion of a proposition held true to positing a corresponding belief asserted, as Setiya does. It’s true that expressing belief is a sufficient condition for assertion. But expressing non-propositional knowledge might be another. An overly abstract approach to propositional attitudes misses this alternative, because it defines belief as whatever (in the subject) is expressed by assertion. It’s a substantial claim, in need of defense, to say that this method of abstract definition is the only good guide to the unity of mental things. How could practical knowledge fail to be contemplative knowledge? Anscombe thinks practical knowledge is the formal cause of intentional action. This implies General, and that practical knowledge cannot be contemplative knowledge, because its object depends on being known. If practical knowledge is not of a prior fact, it can seem independent of its object. But if practical knowledge is completely independent of its object, then it could exist without its object. As Anscombe (1963: 82) says: ‘Someone might say that it was a funny sort of knowledge that was still knowledge even though what it was knowledge of was not the case!’ Anscombe’s response to this objection is obscure. Small (2012), Frey and Frey (2017), and Frost (2018) all argue, in different ways, that practical knowledge is not completely independent of its object: if one merely thinks one is writing, but one is not, then one cannot have practical knowledge of writing. Something similar applies to merely thinking one knows why one acts. The relative merits of these Anscombean views haven’t been clearly worked out, but they do seem internally consistent. If so, then Anscombean views are, in principle, viable. We aren’t forced to think of practical knowledge on ‘incorrigibly contemplative’ lines. Anscombe might still turn out wrong but Anscombean views deserve more careful consideration. Consider Davidson’s 240

Agency and practical knowledge

counterexample again. It depends on assuming that practical knowledge relates to belief and confidence the way that theoretical or contemplative knowledge does. Now imagine someone walking in and asking Davidson’s agent: ‘What are you doing?’ Even for agents who know that making ten carbon copies is difficult, the natural answer is not ‘I’m not sure.’ It is rather, ‘I’m making ten carbon copies (for the house sale).’ This is indeed what the agent is doing, and their practical self-awareness and what they are doing are non-accidentally unified. That would make sense if practical knowledge is Anscombean, and yet is assertorically expressible.

7 Reason and knowledge Most literature on practical knowledge focuses on knowledge of what one does. But practical knowledge is also supposed to be knowledge of why one acts. This raises several difficult questions about the relation between reason and knowledge. In concluding I’ll raise the questions without answering them, to indicate directions for further inquiry. Why should practical knowledge of what one does and why go together? Anscombe (1963: 57) says one cannot understand practical knowledge unless one understands practical reason. Anscombe (1963: 58–81) offers a specific, contentious theory of practical reason on which a practical syllogism’s conclusion is action, where the syllogism articulates a means-ends order present in action (if all goes well, and the order is executable, and executed). Anscombe (1963: 82) also suggests one only has practical knowledge to the extent one has settled on a sufficiently determinate means-ends order. If what one does (typically) embodies a meansends order, and practical knowledge is constitutive knowledge of a means-ends order, that would explain why practical knowledge of what one does and why go together. I cannot compare relative merits of alternative views here. It’s enough to note that alternatives will maintain that something else constitutes means-ends orders, so there is no a priori reason to expect that knowledge of what one does and why must go together. Such views must explain why it seems that knowledge of what one does and why spring from the same source. How far does practical knowledge of why one acts extend? Anscombe (1963: 72–76) thinks the first premise of practical reasoning mentions something the agent thinks somehow good, and reveals a way it’s considered good, made out in terms of ends like pleasure and power. So for Anscombe, practical knowledge extends to knowledge of one’s final ends. Practical knowledge may even extend to moral knowledge, if one must be conscious of the necessity of such first premises (see Rödl 2011 in ‘Further Reading’; cf. Engstrom 2009). Alternative views of practical reason are likely to restrict practical knowledge of why one acts. For example, if one thinks practical reason is a process that takes a desired action-specification as input and produces an intention to perform a means as output, then even if one thinks practical knowledge depends on practical reason, one probably won’t think practical knowledge must extend to final ends. Pleasure, after all, is not something I aim to do, the way I do aim to finish this paragraph. To what extent do practical knowledge and practical reason depend on theoretical epistemic standards? Many premises in practical reason represent mere facts about the world; they don’t directly represent what one does. For example, I know that typing is a way to write, but this doesn’t directly represent what I do. Anscombe (1963: 50) thinks that ‘worldly premises’ need only concern what can happen (not what must, or will probably, happen) if one pursues certain means, and that the agent’s corresponding beliefs need only be true in order to inform corresponding intentions through a correct exercise of practical reason. So on Anscombe’s view, the constraints of theoretical epistemic standards on practical reason 241

Kim Frost

and knowledge are pretty minimal. Other views – especially those following ‘incorrigibly contemplative’ lines – would likely balk at the idea that knowledge could rest on, say, an accidentally true superstition concerning a worldly premise. This leads to a final, deep, difficult question: what is the genus to which practical and theoretical (or contemplative) knowledge belong as species? No one has an explicit, excellent, generally accepted answer yet. That is one main reason why ‘incorrigibly contemplative’ views are so popular: theoretical knowledge is at least very familiar and well-examined. That said, I have argued that practical knowledge could well be Anscombean. If so, there is work to do in general epistemology to make it part of the family.

Related topics Agency and causation; Agency, Attention and consciousness; Agency, powers, and skills; Agency and self-knowledge; Agency and practical reasoning; Agency and morality.

Further reading Rödl, S. (2011). ‘Two Forms of Practical Knowledge’. In Anton Ford, Jennifer Hornsby, and Frederick Stoutland (eds.), Essays on Anscombe’s Intention. Cambridge, MA: Harvard University Press: 211–241. Explores the idea that Anscombe and Kant had similar views of practical knowledge (as knowledge of what one does, and why, and also moral knowledge). Roessler, J. (2010). ‘Agent’s Knowledge’. In Constantine Sandis and Timothy O’Connor (eds.), A Companion to the Philosophy of Action. Hoboken, NJ: Wiley-Blackwell: 236–243. An alternative summary of relevant issues concerning knowledge of what one does. Schwenkler, J. (2015). ‘Understanding “Practical Knowledge”. Philosophers’ Imprint 15. Traces Anscombe’s theory to its roots in Aquinas and Aristotle. Setiya, K. (2017). Practical Knowledge. Oxford: Oxford University Press. The most detailed recent comprehensive study of practical knowledge and its importance to action theory and ethics.

References Anscombe, G. E. M. (1963). Intention. 2nd edition. Oxford: Basil Blackwell. (Reprinted 2000 by Cambridge, MA: Harvard University Press). Armstrong, D. (1968). A Materialist Theory of Mind. London: Routledge. Campbell, L. (2018a). ‘An Epistemology for Practical Knowledge’. Canadian Journal of Philosophy 48(2): 159–177. Campbell, L. (2018b). ‘Propositionalism about Intention: Shifting the Burden of Proof ’. Canadian Journal of Philosophy, DOI: 10.1080/00455091.2018.1512819. Davidson, D. (2001). Essays on Actions and Events. Oxford: Oxford University Press. Engstrom, S. (2009). The Form of Practical Knowledge. Cambridge, MA: Harvard University Press. Frey, C. and Frey, J. (2017). ‘G.E.M. Anscombe and the Analogical Unity of Intention in Perception and Action’. Analytic Philosophy 58(3): 202–247. Frost, K. (2018). ‘A Metaphysics for Practical Knowledge’. Canadian Journal of Philosophy, DOI: 10.1080/00455091.2018.1516972. Ginet, C. (1990). On Action. Cambridge: Cambridge University Press. Grice, P. (1971). ‘Intention and Uncertainty’. The Proceedings of the British Academy 57: 263–279. Hampshire, S. (1959). Thought and Action. Notre Dame, IN: Notre Dame University Press. O’Brien, L. (2003). “On Knowing One’s Own Actions”. In Johannes Roessler and Naomi M. Eilan (eds.), Agency and Self-Awareness. Oxford: Clarendon Press. O’Shaughnessy, B. (2003). “The Epistemology of Physical Action”. In Johannes Roessler and Naomi M. Eilan (eds.), Agency and Self-Awareness. Oxford: Clarendon Press. Paul, S. (2009). ‘How We Know What We’re Doing’. Philosopher’s Imprint 9(4): 1–24.

242

Agency and practical knowledge Paul, S. (2012). ‘How We Know What We Intend’. Philosophical Studies 161(2): 327–346. Paul, S. (2015). ‘The Transparency of Intention’. Philosophical Studies 172(6): 1529–1548. Rödl, S. (2007). Self-Consciousness. Cambridge, MA: Harvard University Press. Sellars, W. (1997). Empiricism and the Philosophy of Mind. Cambridge, MA: Harvard University Press. Setiya, K. (2008). ‘Practical Knowledge’. Ethics 118(3): 388–409. Small, W. (2012). ‘Practical Knowledge and the Structure of Action’. In Abel Günter and James Conant (eds.), Rethinking Epistemology. Berlin: DeGruyter, 133–228. Thompson, M. (2011). ‘Anscombe’s Intention and Practical Knowledge’. In A. Ford, J. Hornsby and F. Stoutland (eds.), Essays on Anscombe’s Intention. Cambridge, MA: Harvard University Press. Velleman, J. D. (1989). Practical Reflection. Princeton, NJ: Princeton University Press.

243

22 AGENCY AND EVIDENCE Berislav Marušić and John Schwenkler1

It is a special and hard problem to understand how evidence figures into the reasoning of an agent. It is a problem for philosophers but also one we all encounter in our daily lives. Our aim in this chapter is to identify the problem and outline a possible solution to it.2

1 The problem In a nutshell, the problem that agents face is how to consider evidence in their decisionmaking without taking it as grounds for a prediction, that is, without turning their decision into an estimate of what they are going to do.3 Let us illustrate this with an example. Imagine that you are considering running a marathon with a limited number of starting places.4 The starting places are assigned through a lottery. Now suppose that consideration of your chances reveals the following: your chance of getting a starting place is pretty good—say 80%. Your chance of actually running the marathon, conditional on getting a starting place and the world cooperating with your plans, is also pretty good—say also 80%. If a well-informed bookie were offering bets on whether (a) you would get a starting place in the lottery, and whether (b) you would actually run the marathon, conditional on getting a starting place and the world cooperating, they would offer the very same odds for both. The problem is that you cannot consider the question of whether you will actually run the marathon in the same way as that of whether you will get a starting place through the lottery. As an agent, you must face the respective uncertainties in two very different ways. That is because whether you get a starting place in the race is not up to you, and so you can’t decide to get a starting place. You have to make your decision to run the marathon conditional on winning a starting place in the lottery—and, in general, conditional on the world’s cooperation. But whether you actually run the marathon is, we may suppose, entirely up to you. (Let us ignore the possibility of getting injured, or sick, or hit by a car or a meteorite. Though some of these things are to some extent up to you, they also involve ways in which the world must cooperate.) And if it is entirely up to you whether you actually run the marathon, then you cannot make your decision to run it conditional on—your running the marathon! You have to regard the uncertainty that arises from the possibility of deciding not to run the marathon in a different way than the uncertainty that arises from the possibility of not getting a 244

DOI: 10.4324/9780429202131-28

Agency and evidence

starting place, or of the world not cooperating with your choice. The fact that it is up to you whether you actually run the marathon makes a difference to how you should think of the evidence concerning whether you will actually do this. However, it would be a mistake to conclude that, since it is up to you to run the marathon, you needn’t consider evidence concerning the difficulty of following through. You cannot make a good decision if you ignore how difficult it is to actually run a marathon. You must consider the fact that running a marathon will require resolve and persistence. So, if you make a good decision, you will take the difficulty into account in considering which ends you set for yourself and which means you choose to achieve them. That is to say, you will consider whether this difficult project is really worth the effort. And, if you decide that it is, you will consider how best to pull it off: you will train hard in advance of the race, and you will be mindful of the need for resolve during the race. In this way, you will not ignore the evidence of difficulty. But you will also not use it as the basis for prediction. The problem is to say how you should consider this evidence, without using it as grounds for prediction of your future behaviour. In what way can you reason about the possibility that you will fail to do something that it is up to you to do?5

2 Answer I: decision theory A first answer might be suggested by decision theory. The rationality of a decision is a function of both preferences and probabilities. In making a decision, an agent has to take into account both how much they prefer the possible outcomes of their actions and how probable those outcomes are, conditional on performing those actions. However, we do not think that this is an adequate answer, for two reasons. First, this view does not really address the problem we have identified. Second, it leads to incoherence between an agent’s practical and theoretical conclusions.6 Let us illustrate this with an example. Suppose you dramatically prefer to run a marathon, on the grounds that it will secure you abiding admiration from those you care about the most. You then set out to consider how to do it. And suppose you have three ways of pursuing your goal. First, you could set out to train on your own. Second, you could join a training group at a local gym. Third, you could join a training group and hire a personal trainer at the same local gym. (Note: you can only hire a personal trainer if you join the training group!) Needless to say, joining the local gym is expensive, and joining the gym as well as hiring a personal trainer is extremely expensive. If you went by your preferences alone, you would pursue your goal of running the marathon by training by yourself, thus saving yourself the money for a gym membership and a personal trainer. However, since you have to take into account the difficulty of running a marathon, decision-theoretic calculation suggests that one of the other courses of action is the more rational one. Let us suppose that it is, in fact, joining a training group and foregoing a personal trainer, as the increase in costs for hiring the trainer does not justify the increase in odds of success. Finally, let us suppose that, in arriving at this conclusion, you have assigned probability 0.8 to the outcome of successfully running the marathon, conditional on joining a training group at the local gym. But now we can see the problem. In your decision-theoretic calculations, you are doing nothing other than making predictions about your future actions—you are assigning probabilities to outcomes that are up to you. And our challenge was to articulate how to consider the uncertainty that arises about matters that are up to the agent is different from the uncertainty that arises about matters that are not up to the agent. The decision-theoretic approach has not done that. 245

Berislav Marušić and John Schwenkler

Moreover, the decision-theoretic approach has also given rise to incoherence between your practical thought and your theoretical thought. On practical grounds, you have reached a conclusion that you might express by saying, “I will run the marathon by joining a training group.” But on purely theoretical grounds, you have reached a conclusion that you might express by saying, “There is an 80% chance that I will run the marathon by joining a training group.” 7 And that is odd. You are, in effect, committed to the conclusion that you might express by saying, “I will run the marathon, though there is a 20% chance that I won’t.” Your view of the future is incoherent.8

3 Answer II: practical reasoning Suppose that you deliberate about whether to run the marathon in a way we might represent as follows: (1) There’s a marathon being held in my home city six months from now. Running the marathon will secure me abiding admiration from those I care about the most. If I enter to run, there’s an 80% chance that my name will be drawn in the lottery. It will also cost about $150 to register for the race if my name is drawn. In order to complete the entire race, I’ll need to do a lot of training. I will be much better able to accomplish this if I join a training group rather than attempting to train on my own. However, joining the training group will be expensive, since it requires joining a gym. My household finances are in good enough shape to cover the $150 fee and the gym expenses, though it will be tight. So I’ll enter to run, pay the $150 fee if my name is drawn, join a training group to help me prepare, and then complete the race on the day it is run. Here, the “so” of your conclusion in the final sentence is an instance of what Jonathan, Dancy (2018, 15) helpfully calls the “practical ‘so’”: it is part of an act in which a decision to do something is drawn from a body of considerations that one takes to rationalize or support it. The decision is a conclusion of practical reasoning. The hard question is: How are the considerations listed as premises in (1) supposed to support your drawing this conclusion? The first thing we should note is that the expression of your conclusion in (1) is shorthand for a judgement with a lot more internal complexity. That is, the decision you reach when your reasoning concludes is not the decision that, if your name is drawn in the lottery, you will do the things that it describes no matter what—that you will do them even if, say, your spouse has a medical emergency and your checking account is overdrawn, or the training group meets during your child’s weekly soccer games, or on the day of the race you have a seriously injured ankle. These conditional aspects of your intention, which have been explored in detail in important work by Luca Ferrero (2009), are left implicit in (1) but might be stated explicitly if they became contextually relevant—if someone were to ask, say, whether you really are going to run the marathon if your injured ankle isn’t healed.9 This is what we earlier tried to capture by saying that the world must cooperate with your plans. The conclusion that you reach in (1) is not, however, even implicitly conditional on anything like your still wanting to run the marathon on the day it happens or to train for it during 246

Agency and evidence

the months leading up to it. If the conclusion of your reasoning were conditional on things like those, then your conclusion would be better expressed by saying, for example, that “I’ll think about running the marathon,” or “I’ll start training for the race and then see how I feel,” or something of that sort. But what accounts for this difference? Looking forward to your future, you can appreciate that there’s a non-zero chance both that (a) when the day of the marathon comes you’ll have sprained your ankle training and will decide not to run the race in order to avoid making the injury more serious; and that (b) when the day of the marathon comes you’ll strongly desire to sleep in and will decide not to run the race on those grounds. We may suppose that neither your injured ankle nor your desire to skip the race will prevent you from running in the way that, say, being kidnapped would do. Then why is the injury, but not the desire to bow out, something your decision can be conditional on? The answer is not simply that you are able to control your desires in a way that you cannot control your bodily health—for in fact the reasoning in (1) can be supplemented with further practical reasoning aimed at doing both of these things. For example, in anticipation of getting cold feet, you might reason as follows: (2) I’m likely to want to skip the race if I haven’t prepared mentally for it. So I’ll spend the week before the race watching films like Chariots of Fire and Prefontaine. Similarly, in anticipation of a possible injury, you might reason along these lines: (3) I’m likely to injure my ankle if I play pickup basketball. So I’ll also avoid playing pickup basketball until the marathon is past. In both cases, you reason practically in light of the conclusion you reached in (1). Indeed, it seems like a commitment to reasoning in ways like these is a rational requirement of being seriously committed to the conclusion of (1). In each case, however, it seems clear that your strategy for preventing the undesired outcome might fail to do the trick. We are back with the question asked earlier: Why is the potential injury, but not the potential desire to bow out, something that your decision to race the marathon can be conditional on? Here is a better answer: The difference between an injured ankle and a desire not to run is that only the former gives you any kind of reason not to run the race. If on the day of the race your ankle is injured, this gives you reason to sit the race out, because completing it would cause you severe pain, or risk making your injury more serious. By contrast, if on the day of the race you simply want to skip it, then this gives you no reason at all not to run, but only a reason to suck it up. Of course, if your desire not to run the race is itself well grounded, say because there’s been an emergency at home that needs your attention, then you have a reason not to run the race. But in that case the reason you have is not your desire itself, but rather the situation at home. Simply wanting not to do something you’ve decided to do gives you no reason at all to abandon your decision in favour of this desire. Let’s return to the question we asked originally. In virtue of what do considerations listed as premises in (1) put you in a position to draw the conclusion you do? That conclusion has the form of a statement about the future: you’d express it by telling someone, “I am going to enter the lottery to run the marathon, and train for the race and then run it if my name is drawn.” Yet how can you conclude this on the basis of the considerations that you reason from? 247

Berislav Marušić and John Schwenkler

To see how this is possible, let us compare the reasoning in (1) with reasoning that might proceed from the following considerations: (4) All the evidence indicates that Candidate X is going to enter the race. If she enters the race, she will raise millions of dollars in campaign funds. If she raises millions in campaign funds, she will have an 80% chance to win the primary. If she wins the primary, she will win the general election. Imagine that on the basis of the considerations in (4) you draw a conclusion that you might express by saying, “Candidate X is going to enter the race, raise millions of dollars in campaign funds, and win the general election if she wins the primary.” This conclusion, just like your conclusion in (1), is a description of future events. Yet whereas the conclusion in (4) will be reasonable only if it is supported by sufficient evidence that this is what Candidate X is going to do, it seems reasonable for you to conclude as you do in (1) even if the considerations you reason from, or any others you might be able to appeal to, do not provide sufficient evidence that you’ll act in the way described.10 That is because in reasoning as you do in (1), in contrast to (4), the conclusion that you draw is such as to depend on the reasoning in light of which you draw it. That is, when you reason as you do in (1), the question “Am I going to run the marathon?” is not a question that you treat as having an answer that it is the aim of your reasoning to discover. This is rather the way that you reason in (4), where Candidate X’s likely decisions, and her prospects in the election depending on what she decides, are matters that you treat as fixed, and that you aim in your reasoning to reflect.11 By contrast, when you reason about what you are going to do, the answer that you reach is supposed to be settled by the reasoning that leads you to it, not by some independent matters which that reasoning aims to reflect. Indeed, we can now see the error of predicting matters that are up to us. The error is to treat the subject matter as to be reflected in our reasoning rather than to be made so by it.12

4 Practical knowledge This final section will relate the position we have developed to a further thesis, frequently identified with the work of G.E.M. Anscombe, namely that practical reasoning about what to do gives an agent practical knowledge of what she is doing or is going to do. The version of this thesis that Anscombe herself explicitly endorses extends only to present action, insofar as this action is intentional: it says that practical reasoning is the ground of practical, non-observational knowledge of what one is presently in the process of doing (Anscombe 1963, 11–12 and passim). But Anscombe brings her account of practical reasoning to bear on the topic of future action as well, claiming that statements of the form, “I am going to φ,” where they express intentions rather than estimates of one’s future behaviour, are grounded in reasons that do not suggest “what is probable, or likely to happen,” but rather in reasons “suggesting what it would be good to make happen with a view to an objective” (Anscombe 1963/2000, 4). Other philosophers, most prominently David Velleman (1989/2007) and Kieran Setiya (2007, 2008), have treated belief about what one will do as a central element in their accounts of an agent’s practical knowledge. The account of practical reasoning that we have developed here suggests a way around a common objection to positions like Velleman’s and Setiya’s, namely that in positing an essential connection between intention and belief they demand that an agent make an epistemically unjustified “leap of faith” in forming the intention to do something.13 By contrast, 248

Agency and evidence

a central element in our account is that beliefs about one’s future actions are necessarily not subject to the evidential norm governing beliefs about other matters. To the extent that your future actions are up to you to decide, they are such as to depend on the reasoning in light of which you decide this, which means there is not a requirement that they be supported by adequate evidence. However, we also hold that an agent’s practical reasoning does not always give her practical knowledge that she will do what she decides.14 When, through practical reasoning, you conclude that you will run the marathon, you do not necessarily have practical knowledge that you will do so. That is because the fact that it is difficult to run the marathon may be evidence that defeats your practical knowledge, preventing you from knowing that you will act as you have decided. To see this, consider a very different kind of case, involving mathematical knowledge. Suppose you calculate something by longhand. Your calculation is accurate and you have made no mistake. It seems that you have a priori knowledge of the result of your calculation: your knowledge of the result does not rest on evidence or observation but is grounded in the calculation alone. However, if you acquire evidence that there is a good chance that you made a mistake—say, because you often make mistakes when you perform such calculations, or because people who are in many respects like you often make mistakes, or because a bookie will offer bets that reflect a significant chance that you have made an error—this will defeat your a priori knowledge, even if your calculation was perfect. Analogously, the fact that there is a 20% chance that you will fail to run the marathon serves as a defeater for any potential practical knowledge you might have had. Your situation would be analogous to a case in which you proved something mathematically but had empirical evidence that you made a mistake. It is instructive to consider why this must be so.15 If our practical reasoning always had to conclude in practical knowledge in order to be good reasoning, there would be a constraint on the strength or significance of our reasons: reasons could speak in favour of φ-ing only if they put us in a position to know that will φ if we decide to φ. The resulting view of the strength or significance of our reasons would be implausible, because we may have overwhelming reasons to do something that is exceedingly difficult. We may strongly want, immensely value, or even simply be obliged to do something, and this can make it practically rational to do it—but the fact that we desire or value or are obliged to do it need not put us in a position to know that we will do it, since the difficulty of doing it would be a defeater for our purported knowledge. The argument of the preceding section also shows that an agent is in a different position with respect to evidence about the difficulty of her intended actions than with respect to evidence concerning matters that are not up to her. For example, the fact that it will be difficult to run the marathon, due to the physical demands of the race as well as the challenge of keeping yourself sufficiently confident and motivated, is something you can account for in your practical reasoning by forming plans that will help you to overcome these difficulties. Yet plans like these cannot eliminate entirely the possibility of failing to do what you intend. In particular, they cannot eliminate the need for resolve in sticking to a decision in the face of temptations to abandon it. What is the right response to this limitation? It does not lie in saying, “I’ll do it … unless I don’t,” in any situation where the possibility of failure becomes salient—for as we have argued, this is to regard one’s intention as conditional on one’s future state of mind (cf. Anscombe 1963/2000, §52). The moral is rather that, in making up our mind, we sometimes must be prepared to be resolute in the face of difficulty—even when we know that no bookie will think in the way that we do. 249

Berislav Marušić and John Schwenkler

5 Conclusion Kant’s famous dictum is that we act under the idea of freedom. In the present paper, we do not mean to make headway on the question of what freedom is. However, we hope to have shown that freedom has an important epistemological significance. It licenses us to think of the future in a distinctive way. In particular, when something is up to us, then no matter how hard it is for us to do it, what will happen is something for us to decide—and deciding is a practical matter. We evade our responsibility if in those circumstances we seek to predict what will happen—and sometimes such evasion is couched in claims of theoretical rationality. As agents, we should forego such comfort.

Further topics Agency and practical knowledge; Agency and self-knowledge; Diachronic agency; Planning agency; Agency, time, and rationality; Agency and practical reasoning.

Notes 1 Both authors contributed equally to this work. 2 For further discussion, see Marušić (2012, 2013, 2015), Marušić and Schwenkler (2018), and Schwenkler (2019, Chs. 5–7). The present discussion synthesizes and builds upon that work. 3 This use of “estimate” is due to Anscombe (1963/2000, 2). The problem is sometimes discussed in decision theory in terms of the question of whether deliberation crowds out prediction. See Spohn (1977), Levi (1997, 2007), Rabinowicz (2002), Joyce (2002), Hájek (2016), Vavova (2016), Liu and Price (2019). We think that, with the exception of Vavova (2016), those discussions don’t quite get to the heart of our problem, because they fail to distinguish the problem as it arises for decision theory from the problem as it arises for the ordinary agent. Our topic here is the latter. 4 This example organizes the argument of Marušić (2015). 5 The question here bears on the extensive discussion of the case of Professor Procrastinate from Jackson and Pargetter (1986). See Goldman (1976), Portmore (2011, 2019), Ross (2012), and Jackson (2014). The view defended here has affinities with Thomason (1981) and Liu and Price (2019). 6 A further problem is that, on this view, there is no resting place for practical thinking, because the fact that one is making a decision (on certain grounds) affects the probability assignment of the outcomes. Since the outcomes are not independent of one’s decisions, it is not clear how the decision-theoretic approach applies. We won’t pursue this problem here. The structure of the problem is nicely brought out by Bok (1998), albeit not in probabilistic terms. 7 Assuming something like Lewis’s Principal Principle (e.g. Lewis 1994). 8 One might reply by defending non-cognitivism about practical reason, according to which the conclusions of practical reasoning are not beliefs (e.g. Bratman 1987). However, even on a non-cognitivist view, there are coherence requirements on one’s intentions and beliefs, so that one’s overall view will be irrational. See Marušić (2013, 2015, Ch. 3.1) and Marušić and Schwenkler (2018) for further discussion. 9 For further development of our view about this, see Marušić and Schwenkler (2018, §3.1). 10 You will, however, need to have adequate evidence for many of the premises you reason from— e.g. that the race is being held on such-and-such a date, that you have the money to pay the registration fee—but that is another issue. 11 There is, of course, a general problem about how to understand rational belief and knowledge of the future, if the future is genuinely open. This is an issue we set aside. 12 Velleman’s distinction between conclusions we accept “so as to reflect the truth” and conclusions we accept “so as to create the truth” informs our present discussion (Velleman 1996, 195, n.55). However, we do not adopt the entirety of Velleman’s view. For differences, see Marušić (2015, Ch. 6.2). 13 See esp. Langton (2003) for this objection in response to Velleman (1989/2007). 14 We take a similar position in regard to present action as well. For relevant arguments, see Schwenkler (2015, 17–25) and (2019, Ch. 6.4). 15 See Marušić (2015, esp. 115–116) and Schwenkler (2019, Ch. 6. 4) for further discussion.

250

Agency and evidence

Recommended reading Anscombe, G.E.M. (1963/2000). Intention, 2nd ed. Cambridge, MA: Harvard University Press, 2000. The seminal contemporary work on the nature of practical reason. Jackson, Frank and Robert Pargetter. (1986). ‘Oughts, Options, and Actualism’. The Philosophical Review 95, 233–255. Presents the influential example of Professor Procrastinate, which raises the problem of whether prediction informs decision. Marušić, Berislav. (2015). Evidence and Agency: Norms of Belief for Promising and Resolving. Oxford: Oxford University Press. Develops the view that is outlined in this chapter. Schwenkler, John. (2019). Anscombe’s Intention: A Guide. New York: Oxford University Press. Provides an interpretation of Anscombe with extensive discussion of the form and structure of practical reasoning. Velleman, David. (1996). ‘The Possibility of Practical Reason’. Chap. 7 in The Possibility of Practical Reason. Oxford: Oxford University Press, 2000. Shows how belief can be grounded in practical reasoning.

References Anscombe, G.E.M. (1963/2000). Intention, 2nd ed. Cambridge, MA: Harvard University Press, 2000. Bok, Hilary. (1998). Freedom and Responsibility. Princeton, NJ: Princeton University Press. Bratman, Michael. (1987). Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Dancy, Jonathan. (2018). Practical Shape: A Theory of Practical Reasoning. Oxford: Oxford University Press. Ferrero, Luca. (2009). “Conditional Intentions.” Noûs 43, 700–741. Goldman, Holly. (1976). “Dated Rightness and Moral Imperfection.” The Philosophical Review 85, 449–487. Hájek, Alan. (2016). “Deliberation Welcomes Prediction.” Episteme 13, 507–528. Jackson, Frank. (2014). “Procrastinate Revisited.” Pacific Philosophical Quarterly 95, 634–647. Jackson, Frank and Robert Pargetter. (1986). “Oughts, Options, and Actualism.” The Philosophical Review 95, 233–255. Joyce, James. (2002). “Levi on Causal Decision Theory and the Possibility of Predicting One’s Own Actions.” Philosophical Studies 110, 69–102. Langton, Rae. (2003). “Intention as Faith.” In Action and Agency, edited by Helen Steward and John Hyman. Cambridge: Cambridge University Press: 243–258. Levi, Isaac. (1997). The Covenant of Reason: Rationality and the Commitments of Thought. Cambridge: Cambridge University Press. Levi, Isaac. (2007). “Deliberation Does Crowd Out Prediction.” In Hommage à Wlodek: Philosophical Papers Dedicated to Wlodek Rabinowicz, edited by Toni Rønnow-Rasmussen, Björn Petersson, Jonas Josefsson and Dan Egonsson, Lund University. Lewis, David. (1994). “Humean Supervenience Debugged.” Mind 103: 473–490. Liu, Yang and Huw Price. (2019). “Heart of DARCness.” Australasian Journal of Philosophy 97, 136–150. Marušić, Berislav. (2012). “Belief and Difficult Action.” Philosophers’ Imprint 12, 1–30. Marušić, Berislav. (2013). “Promising against the Evidence.” Ethics 123, 292–317. Marušić, Berislav. (2015). Evidence and Agency: Norms of Belief for Promising and Resolving. Oxford: Oxford University Press. Marušić, Berislav and John Schwenkler. (2018). “Intending Is Believing: A Defense of Strong Cognitivism.” Analytic Philosophy 59.3: 1–32. Portmore, Douglas. (2011). Commonsense Consequentialism. New York: Oxford University Press. Portmore, Douglas. (2019). Opting for the Best: Oughts and Options. New York: Oxford University Press. Rabinowicz, Wlodek. (2002). “Does Practical Deliberation Crowd out Self-Prediction?” Erkenntnis 57, 91–122. Ross, Jacob. (2012). “Actualism, Possibilism, and Beyond.” In Oxford Studies in Normative Ethics 2, edited by Mark Timmons. Oxford: Oxford University Press.

251

Berislav Marušić and John Schwenkler Schwenkler, John. (2015). “Understanding ‘Practical Knowledge’.” Philosophers’ Imprint 15, 1–32. Schwenkler, John. (2019). Anscombe’s Intention: A Guide. New York: Oxford University Press. Setiya, Kieran. (2007). Reasons without Rationalism. Princeton, NJ: Princeton University Press. Setiya, Kieran. (2008). “Practical Knowledge.” Ethics 118, 388–409. Spohn, Wolfgang. (1977). “Where Luce and Krantz Do Really Generalize Savage’s Decision Model.” Erkenntnis 11, 113–134. Thomason, Richmond. (1981). “Deontic Logic and the Role of Freedom in Moral Deliberation.” New Studies in Deontic Logic: Norms, Actions and the Foundations of Ethics 152, 177–186. Vavova, Katia. (2016). “Deliberation and Prediction: It’s Complicated.” Episteme 13, 529–538. Velleman, David. (1989/2007). Practical Reflection, 2nd ed. Stanford: CSLI Publications. Velleman, David. (1996). “The Possibility of Practical Reason.” In The Possibility of Practical Reason. Oxford: Oxford University Press, 2000, 270–299.

252

23 AGENCY AND SELF-KNOWLEDGE Brie Gertler

1 Self-knowledge The term “self-knowledge” is standardly used by philosophers to refer to the kind of knowledge that each of us (sometimes) has of our own mental states: what it is that we think, feel, believe, or desire. This chapter concerns self-knowledge in that sense, with a focus on knowledge of what we believe to be true and knowledge of what we intend to do. In particular, we will examine the influential idea that beliefs and intentions are exercises of agency, and the implications of that idea for accounts of the individual’s knowledge of her own beliefs and intentions.

1.1 An initial contrast: agency and knowledge Think about some ordinary ways in which we exercise agency. I kick the can, and it bounces down the road. I call my dog, and he runs toward me. In these cases, my agency involves my affecting objects or phenomena. By contrast, ordinary cases of knowledge typically involve some object or phenomenon affecting me. I know that the can is bouncing down the road because its movement affects my visual experience—I see it bouncing. I know that my dog is running toward me because the crackling of dry leaves under his paws shapes my auditory experience—I hear him coming. So the direction of control or causation ordinarily operative in knowledge seems to be the reverse of that in which agency consists. Agency typically involves affecting some object or phenomenon, whereas knowledge typically involves being affected by some object or phenomenon.1 This general idea, that knowledge involves being affected by the object or phenomenon that is known, clearly applies to some knowledge of one’s own mental states. When the doctor asks me whether the pain I’m feeling is a dull ache or a sharp twinge, I reflect on the pain so that my response will be shaped by the quality of the pain: its dullness or sharpness. When trying to decide what flavor of ice cream to order, my preference—the fact that pistachio appeals to me more than vanilla, say—will shape my thought that I’d prefer pistachio, and will thereby help to explain why I order pistachio.

DOI: 10.4324/9780429202131-29253

Brie Gertler

1.2 “Active” thoughts and attitudes We have just seen that, on a standard model, knowledge consists in an awareness—for example, a thought or belief—that is shaped by what is known. According to a view I call agentialism, this model cannot exhaustively explain self-knowledge, knowledge of our own mental states. Agentialists claim that the standard model is plausible only as regards self-knowledge of passive states, like pain and flavor preferences.2 On their view, there are also active mental states, and an alternative model is needed to explain self-knowledge of these active states. Active mental states are those that involve exercises of agency. They form two broad classes. The first class is exemplified by episodes of goal-directed conscious thinking, such as imagining how Fred would look in a purple cardigan. It’s natural to say that this is something that you do, rather than something that happens to you or an experience you merely undergo. And some philosophers will maintain that our knowledge of what we’re doing is not achieved purely through observation: after all, simply observing the image of a man in a purple cardigan does not reveal whether you’re imagining Fred or his identical twin. We will not discuss these sorts of cases here. Our focus will be the second class of mental items regarded as active: committal attitudes such as beliefs and intentions. According to agentialists, we don’t simply find ourselves with the belief that liberal democratic ideals are on the decline, or with the intention to visit an ailing friend. In paradigmatic cases, believing and intending are things we do, for reasons. Suppose that, upon assessing current geopolitical trends and considering them against the backdrop of history, you conclude that liberal democratic ideals are waning in influence. Weighing the evidence and drawing a conclusion are something that you do, not merely an experience you undergo. You are similarly active when, reflecting on the obligations of friendship, or simply thinking about your ailing friend, you decide to visit him. These examples illustrate that beliefs and intentions often result from theoretical or practical reasoning. Moreover, existing attitudes are apt to change in the face of new evidence or practical considerations. In this sense, beliefs and intentions are often responsive to reasons. Agentialists take reasons-responsiveness to constitute a kind of agency. So they regard believing and intending, in a way that is responsive to reasons, as exercises of agency: that is, as a type of doing.3 And since we are generally responsible for what we do, we are responsible for our beliefs and intentions. We are therefore subject to normative assessment—credit or blame—according to how well our beliefs and intentions conform to our reasons. “Our rational beliefs and intentions are not mere mental attitudes, but active states of normative commitment” (Korsgaard 2009, 39). The agentialist maintains that, because reasons-responsive beliefs and intentions are exercises of agency, self-knowledge of those attitudes contrasts with self-knowledge of passive states like sensations. In particular, self-knowledge of beliefs and intentions does not fit the picture described earlier, on which knowledge involves being affected by the known object or phenomenon. Suppose that someone asks you whether you intend to visit your friend. If you answer this question by consulting your memory, your belief that you intend to do this is shaped by that intention itself, which you later remember. In effect, you are regarding your intention from the perspective of a passive observer, the perspective you occupy when observing an ache or tickle. You are not treating the intention to visit your friend as a commitment to be endorsed. Similarly, if you respond to the question “do you believe that liberal democratic ideals are on the decline?” by trying to remember your past statements on that topic, you’re treating this opinion as an inert state to be discovered, rather than as a conviction to uphold or disown, that is, a commitment for which you bear responsibility. In 254

Agency and self-knowledge

both of these cases, the method you use is available to others: someone else could check your calendar or remember your past comments about liberal democracy. Agentialists allow that self-knowledge of sensations and other passive mental states is a matter of observing (or otherwise tracking) those states. But, they contend, a subject cannot apprehend her own beliefs and intentions as “active states of normative commitment” merely by observing her cognitive life. For this reason, agentialists maintain that self-knowledge of beliefs and intentions differs significantly from self-knowledge of passive mental states.

1.3 First-person authority Accounts of self-knowledge generally aim to explain “first-person authority”, the special authority that each of us seems to possess as regards our own mental states. As evidence for this authority, consider how odd it is, in ordinary circumstances, to challenge someone’s report about what they believe or intend. If you say, “I believe liberal democratic ideals are on the decline” or “I intend to cook dinner tomorrow”, it would be odd for someone to respond, “No you don’t”. Although it’s generally agreed that we possess first-person authority, the nature and source of this authority are disputed. On one view, the oddness just described is fully explained by the idea that we have first-person epistemic authority about our own attitudes: that is, each of us is in an epistemically privileged position to identify our beliefs and intentions. This means that, as compared with others’ beliefs about my attitudes, my own self-attributions are more strongly justified, or open to fewer sources of error, or produced by processes that are more reliable (etc.). Claims of first-person epistemic authority are standardly expressed as claims about privileged access to one’s own mental states. Agentialists argue that construing first-person authority in purely epistemic terms neglects the more significant kind of authority we possess in believing or intending, namely, that these are acts of first-person agency. In saying, “I believe liberal democratic ideals are on the decline” or “I intend to visit my friend tomorrow”, I am not simply reporting on an internal state that I’m in a special position to know about, as I might report that I have a stomachache. Instead, I am expressing my commitment to the corresponding view or plan, and thereby taking responsibility for it. Agentialists call such expressions of commitment avowals. My avowal conveys that you can hold me accountable for the avowed attitude: you can reasonably ask me to justify this pessimism about liberal democracy or to explain why it is reasonable or appropriate for me to visit my friend tomorrow. The authority to avow an attitude belongs exclusively to the subject, because only the subject herself can take responsibility for her attitudes. In this sense, the agent is said to possess first-personal agential authority relative to her attitudes. Agential authority contrasts sharply with epistemic authority. Others can know that I am in pain, or that I intend to cook dinner tomorrow. In this sense, privileged epistemic access admits of degrees: it may consist in the subject’s having a way of detecting her own mental states that is more reliable than the ways others can detect them, or in one’s own selfascriptions being more strongly justified than others’ ascriptions of mental states to her (etc.).4 By contrast, the authority to take responsibility for a belief or intention is an all-ornothing matter, as beliefs and intentions can be avowed only by the person who believes or intends. Consider the following scenario. Nigella wants to cook eggplant parmigiana, but she loathes grocery shopping. So she tells her brother Nigel “You should cook eggplant parmigiana tonight!”, knowing that once he endorses this plan he’ll buy the necessary ingredients. Nigella stills plan to cook the dinner herself—Nigel is a terrible cook. 255

Brie Gertler

In this case, Nigella exercises agency in bringing it about that Nigel intends to cook tonight. Still, Nigel’s intention to cook is not an exercise of Nigella’s agency. Nigella cannot avow Nigel’s intention. Persuading Nigel to cook, and asserting “Nigel intends to cook tomorrow”, do not make Nigella responsible for justifying his intention. After all, Nigella thinks it would be unreasonable for Nigel to cook. To justify her action, Nigella need only cite possible benefits of Nigel’s intending to cook tonight. By contrast, Nigel can justify his intention only by specifying reasons (for him) to actually cook tonight—for example, possible benefits of his cooking. It is now clear why our agential authority, relative to our beliefs and intentions, strikes many philosophers as more remarkable than our privileged epistemic access to those attitudes. Agential authority is all or nothing, as only the subject can take and bear responsibility for her attitudes, whereas epistemic privilege is a matter of degree. Moreover, agential authority uniquely attaches to one’s own reasons-responsive attitudes and other exercises of agency, whereas epistemic privilege can extend to any phenomenon the subject is especially well-situated to appreciate: I have a degree of epistemic authority as regards my dog’s whereabouts so long as no one else is around to observe him. Now it may appear that agential authority, however remarkable, doesn’t bear on the question of self-knowledge, since self-knowledge is an epistemic phenomenon. But as we will see, leading versions of agentialism take the distinctive epistemic characteristics of self-knowledge of beliefs and intentions to derive from the subject’s agential authority as regards these attitudes.

2 Agential models of self-knowledge We’ve just seen the case against assimilating self-knowledge of beliefs and intentions to self-knowledge of sensations, flavor preferences, and other “passive” states. We now face the question of how to account for self-knowledge of beliefs and intentions. Granting that beliefs and intentions are sometimes acts of agency, what does this mean for self-knowledge of these attitudes? In other words, how is first-personal epistemic authority related to first-personal agential authority?

2.1 Rational agency as providing epistemic entitlement to self-knowledge The most prominent agentialist accounts (Burge 1996; Moran 2001; Boyle 2009) take our epistemic authority, relative to our attitudes, to derive from our agential authority— specifically, our authority as rational thinkers. These accounts are motivated by the view that beliefs and intentions are active insofar as they are responsive to reasons. Tyler Burge’s version of agentialism centers on the idea that, insofar as we are rational thinkers, we have certain obligations. We are obligated to (try to) avoid inconsistent beliefs, to conform our beliefs to our evidence, and to strive to satisfy other rational norms. Burge claims that we could not hope to satisfy these obligations unless we had the capacity for self-knowledge, since we must know what we believe in order to assess our beliefs for consistency with each other and for conformity with our evidence. In this way, he argues, our obligations as rational thinkers ensure that we are capable of self-knowledge. Burge takes this to mean that our rationality gives us a kind of “epistemic entitlement” to beliefs about our attitudes; it is because of this entitlement that such beliefs (when true) qualify as self-knowledge. 256

Agency and self-knowledge

On Burge’s account, our epistemic authority as regards our attitudes—our capacity for self-knowledge—derives from our agential authority, the responsibility we bear as rational thinkers. Burge is concerned to locate the basic source of our epistemic authority, but he does not provide an account of the specific method we use to achieve self-knowledge. Richard Moran shares Burge’s general outlook. Like Burge, he thinks that our capacity to rationally shape our attitudes constitutes a kind of agency, and that this agential authority over our attitudes provides for epistemic authority. Moran goes beyond Burge’s earlier discussion in advancing a view about how our agency provides a route to self-knowledge. Moran’s proposal is inspired by what he takes to be a distinctive characteristic of the agent’s relation to her own attitudes. What rationalizes your belief that p, or your intention to w, are reasons directly bearing on the attitude’s content. For example, your evidence bearing on democracy is on the decline rationalizes your belief to that effect; your reasons to make eggplant parmigiana rationalize your intention to do so. In contrast, others’ reasons for you to have a belief or intention are at best indirectly related to the attitude’s content. Nigella’s reason for persuading Nigel to (intend to) make eggplant parmigiana is that this intention will lead him to buy groceries. This rationale is entirely compatible with Nigella’s firm belief that there is no good reason for Nigel to make eggplant parmigiana. But Nigel’s intention is rationally justified, according to these philosophers, only if Nigel should make eggplant parmigiana is supported by reasons, at least from his perspective. Similarly, if I lead you to believe that democracies are on the decline, I can justify this act of persuasion without invoking any reason to think that democracy is on the decline: perhaps I think your pessimism would be somehow beneficial. But you can’t rationally justify believing that liberal democracy is on the decline except by providing evidence that liberal democracy is indeed on the decline. To summarize, reasons bearing directly on a belief or intention, rather than on the having of the attitude, concern the attitudes’ content (democracy is on the decline or to cook eggplant parmigiana) directly, and directly justify the attitude only for the subject herself, the person who believes or intends.5 These considerations lead Moran to argue that the signal characteristic of agential attitudes is that such attitudes are “transparent” to the subject (see also Boyle 2011). I can determine what I believe about p by considering evidence bearing on p itself, rather than evidence about whether I believe that p. I can determine whether I intend to w by considering reasons for or against w itself, rather than evidence about whether I intend to w. This is known as the transparency method: I grasp my attitudes by “looking through” them, as it were, directly to the facts or actions they concern. Moreover, on this view the transparency of my attitudes (to me) reflects what makes those attitudes truly my own, namely, my rational authority over them. Others may possess the power to control my attitudes through non-rational means (such as brainwashing). My attitudes genuinely belong to me because I exercise rational agency over them. Moran argues that our agential authority, rooted in the capacity to conform our attitudes to our reasons, explains our epistemic authority relative to our attitudes. On his account, we have the right to presume that we are rational thinkers: that is, that our beliefs and intentions are responsive to reasons. This “epistemic right”, a kind of epistemic entitlement, licenses the inference from “the evidence favors p” to “I believe that p” (Moran 2003). Your epistemic authority consists in the fact that you—and only you—can grasp your attitude by directly considering your reasons. This epistemic authority is constituted by your special agential position vis-à-vis your attitudes: your capacity to conform your attitudes to your reasons. 257

Brie Gertler

2.2 Self-knowledge as assured by our normative commitments Akeel Bilgrami also maintains that our responsibility for our attitudes ensures that we possess self-knowledge, but offers an alternative view about how self-knowledge is linked with responsibility. On his view, our rational agency guarantees self-knowledge in two ways. The first focuses on the fact that beliefs and intentions motivate our behavior. Suppose that Nigel buys tomatoes because he intends to make eggplant parmigiana and he believes that tomatoes are a necessary ingredient for that dish. Assuming that purchasing the tomatoes is a free action, Nigel is responsible for it; however, Bilgrami says, we cannot be responsible for an action unless we understand the attitudes that motivated it. This means that we possess self-knowledge of beliefs and intentions that rationally motivate our behavior. The second way that rational agency guarantees self-knowledge, on Bilgrami’s view, derives from the fact that beliefs and intentions are normative commitments. Having a belief commits me to behaving in ways that are rationalized by this belief: for example, my belief that it will rain today obligates me to carry an umbrella (assuming that I want to remain dry). Bilgrami maintains that, insofar as we have a commitment, we are disposed to try to live up to it, and to upbraid ourselves if we fail. This link to dispositions requires that we are aware of our commitments. He concludes, “We cannot therefore have commitments without believing that we have them” (Bilgrami 2006, 287). In this way, the normative dimension of beliefs and intentions ensures that we have self-knowledge of those attitudes. Whereas Bilgrami claims that commitments partly consist in dispositions to think or behave in certain ways, a related view focuses on events associated with commitments, namely, avowals. It’s worth noting that, on all of these agentialist views, avowals like “I intend to cook dinner” are themselves exercises of agency. To avow a belief or intention is to take responsibility for it. And to take responsibility is not simply to describe an independent aspect of reality: instead, it is to make it the case that I can reasonably be held accountable for the attitude. If I avow an intention, I can be criticized for failing to carry through on it, or even for having it—if the intention is not aligned with my reasons, or if it is malevolent or foolish, say. Because avowals play this role, they are “commissive, not descriptive” (McGeer 1996, 508). The idea that avowing a belief or intention is commissive forms the basis for Annalisa Coliva’s version of agentialism (Coliva 2012). She argues that, so long as a thinker is rational and possesses the relevant concepts, self-ascriptions of beliefs or intentions will create the attitudes ascribed, by committing the thinker to those attitudes. She extends this “constructivist” view to rationally held conative attitudes, such as desires. When understood in the way proposed, a judgement (or a sincere assertion) such as “I believe/desire/intend/wish/hope that P” is like a performative, namely like “I promise to buy you an ice-cream”…: it makes a certain thing happen, for it does create the first-order propositional attitude as a commitment.… [This] is possible precisely because judging “I believe/desire/intend/wish/hope that P” becomes just an alternative way of undertaking the same commitments one would make by judging that P (is worth pursuing or having)… (Coliva 2012: 235–236) For Coliva, judgments about one’s attitudes constitute forming (or committing to) the attitude. The resulting commitment partly consists in dispositions, as in Bilgrami’s view. But the act of agency just described is what makes it a genuine commitment for which the thinker is responsible. 258

Agency and self-knowledge

2.3 Doubts about rational agency Agentialists take our capacity for self-knowledge of beliefs and intentions to derive, in one way or another, from our rational capacities. Some commentators have objected that agentialist accounts rest on an overly idealized picture of human cognition. Psychological studies indicate that our attitudes are often shaped by non-rational factors such as emotions, moods, and arbitrary features of the context or situation in which we form the attitude. Worse yet, we often manufacture or selectively attend to evidence that confirms our beliefs. The capacity to conform our attitudes to our reasons is of questionable value if our “reasons” are shaped by emotions or biases. And the fact that our attitudes correspond to what we take to be our reasons—and are thus transparent to our reasons—is not a triumph of rationality if our sense of our reasons is shaped by non-rational factors. Suppose that Roy’s daughter asks him for a loan so that she can realize her dream of opening a restaurant. Roy carefully deliberates about this decision. He is aware that most restaurants close within the first year. But because of his affection for his daughter, and his desire that she succeed, that statistic does not weigh as heavily in his deliberations as it would if the request were coming from someone else. His hopes for his daughter incline him toward an optimistic view of the restaurant’s prospects even before he deliberates. Moreover, this optimistic outlook directs his attention away from counterevidence, such as his daughter’s lack of business experience. And it leads him to overvalue favorable evidence, such as her impressive cooking skills. In the end, Roy concludes that the restaurant would likely succeed. This belief is transparent: Roy can report on it simply by considering (what he regards as) the evidence. But the transparency of his belief cannot be attributed to his rational agency, since his belief shapes his sense of his reasons, rather than vice versa. The well-established fact that attitudes are often shaped by non-rational factors challenges the epistemic basis of rational agency accounts of self-knowledge. If non-rational factors lead us to overvalue some reasons and blind us to others, then our capacity for rational thought is seriously flawed and limited, and it is hard to see how this capacity could ground an epistemic right to assume that our attitudes are transparent. Even if our reasons result from confirmation bias, post-hoc rationalization, or other dubious sources, our attitudes may well be transparent to what appears to us to be our reasons. But that kind of transparency may derive from our psychological need to fit our reasons to our attitudes—or to selectively attend to reasons that justify our attitudes—rather than our rational capacity to conform our attitudes to our reasons. The influence of non-rational factors challenges the idea that our rational agency, and the attendant obligation to reason critically about our attitudes, could provide an epistemic entitlement that undergirds self-knowledge. Put another way: since in some cases the transparency of a belief or intention is due to the fact that the attitude shapes the thinker’s sense of her reasons, the epistemic authority associated with transparency is not (always) grounded in agential authority.6 Victoria McGeer presses this kind of objection (McGeer 1996, 2007). She argues that a more accurate depiction of our cognitive lives will rely on a more expansive understanding of first-person agency, beyond the agency involved in responsiveness to reasons. Given that we are not ideally rational cognizers, she says, fulfilling our responsibility for our own attitudes sometimes requires adopting a detached, third-person perspective on our cognitive lives. (Moran regards this as a kind of self-alienation.) From this perspective, we can better identify biases, impulses, and other non-rational factors affecting our sense of our reasons. Recognizing these non-rational factors is required if we are to curb their influence (McGeer 2007, 102). 259

Brie Gertler

McGeer’s argument is one strand in a larger skepticism about agentialist accounts of self-knowledge. Given our cognitive flaws and limitations, it’s hard to see how our capacity for rational thought could ground the kind of epistemic rights or entitlements envisioned by agentialists like Burge and Moran. Another strand of skepticism concerns the claim, shared by Bilgrami and Coliva, that there is a conceptual connection between having a commitment and believing or judging that one does. We are often ignorant about the motivations for our decisions and actions, and we sometimes invent—or “confabulate”—motivations that rationalize our choices or otherwise support our self-image (Nisbett and Wilson 1977). Bilgrami does not deny this, but claims only that we could not be responsible for our actions unless we know the attitudes motivating them. But that claim may imply that we are responsible for far fewer of our actions than we ordinarily assume. Similarly, we may be disposed to criticize ourselves for failing to live up to our commitments, if by “commitments” we mean beliefs and intentions that we’re aware of, or at least would readily endorse. But some of our beliefs and intentions operate outside of conscious awareness. And some of these implicit attitudes—including some racist beliefs and retributive intentions—are at odds with what we would avow. This perhaps means that they are not commitments in Bilgrami’s sense. The upshot is that Bilgrami’s account applies most naturally to conscious attitudes that we endorse. The thesis that we possess self-knowledge of conscious attitudes that we would readily endorse is comparatively uncontroversial. Similar questions can be raised about Coliva’s view. We can grant Coliva’s claim that a judgment or sincere assertion creates a commitment, and that in those cases the commitment is known to the subject. But this doesn’t explain self-knowledge of attitudes that are not created in this way. (This indicates a limitation of Coliva’s view, but not a decisive objection to it, as the claim in question is not intended as a comprehensive account of self-knowledge.)

3  Conclusion We typically act as if individuals exert rational agency over their beliefs and intentions: we hold ourselves responsible for our attitudes, and we think it appropriate to demand justifying reasons for beliefs and intentions. Some philosophers take these practices to reflect a deeper truth, that our beliefs and intentions express our rational agency, and maintain that our rational agency is key to self-knowledge of our beliefs and intentions. Others are skeptical of the extent and significance of our alleged rational agency relative to our attitudes.

Related topics Epistemic agency; Agency and evidence; Agency and practical knowledge; Intentional agency; Rational agency; Agency, reasons, and rationality.

Notes 1 The idea that knowledge involves one’s beliefs being shaped by the facts is most plausible for a posteriori knowledge. Many philosophers deny that necessary truths knowable a priori, such as analytic and mathematical truths, are part of the causal order. On that view, such truths cannot affect thinkers or their beliefs. We will not be concerned with a priori knowledge, since the objects of self-knowledge—e.g., that one is undergoing a certain experience or has certain beliefs or ­intentions—are knowable only a posteriori.

260

Agency and self-knowledge 2 Pains and flavor preferences are passive states, even if they arose because of past exercises of agency. If I deliberately kicked myself or cultivated an appreciation of pistachio ice cream, then I exercised agency in bringing about my pain or my flavor preference. But neither phenomenon—the experience of pain or the craving for pistachio ice cream—is itself an exercise of agency. 3 States like pains and flavor preferences are correspondingly passive because they are not sensitive to reasons. Reasoned argument to the effect that feeling a pain is unwise or unwarranted does not alleviate the pain. And although you could invite me to sample some ice cream in an effort to persuade me that vanilla tastes better than pistachio, you can’t convince me of this purely by enumerating reasons that I should prefer vanilla, or providing objective evidence that vanilla is gustatorily superior. 4 Some philosophers argue that there is one kind of first-person epistemic privilege that is an allor-nothing matter, namely, acquaintance. No one else can be acquainted with your mental states. Loosely speaking, a subject is acquainted with a mental state when her awareness of that state is directly linked to the state itself – where the notion of directness here is metaphysical and contrasts with the much weaker causal relations that are implicit in the standard model discussed at the outset of this entry. For more on the notion of acquaintance, see Hasan (2020). 5 Doesn’t my evidence that liberal democracy is on the decline justify my persuading you of this? Not unless there is some reason for me to maximize your true beliefs. 6 The claim that attitudes are transparent to reasons is in itself purely epistemic; it is not committed to the idea that attitudes are exercises of agency. Accounts of self-knowledge that center on transparency, and do not rely on claims about rational agency, have recently been advanced by Byrne (2018) and Fernandez (2013).

Further reading Cassam, Quassim. 2014. Self-Knowledge for Humans. Oxford: Oxford University Press. This book contains an extended examination and critique of agentialist accounts of self-knowledge. Coliva, Annalisa. 2016. The Varieties of Self-Knowledge. London: Palgrave Macmillan. This book evaluates a wide range of accounts of self-knowledge. It makes the case for a pluralist view, according to which self-knowledge takes a variety of forms.

References Bilgrami, Akeel. 2006. Self-Knowledge and Resentment. Cambridge, MA: Harvard University Press. Boyle, Matthew. 2009. Two Kinds of Self-Knowledge. Philosophy and Phenomenological Research 78: 133–164. Boyle, Matthew. 2011. Transparent Self-Knowledge. Proceedings of the Aristotelian Society, Sup Vol. 85: 223–241. Burge, Tyler. 1996. Our Entitlement to Self-Knowledge. Proceedings of the Aristotelian Society 96: 91–116. Byrne, Alex. 2018. Transparency and Self-Knowledge. Oxford: Oxford University Press. Coliva, Annalisa. 2012. One Variety of Self-Knowledge: Constitutivism as Constructivism. In The Self and Self-Knowledge, A. Coliva (ed.). Oxford: Oxford University Press, pp. 212–242. Fernandez, Jorge. 2013. Transparent Minds: A Study of Self-Knowledge. Oxford: Oxford University Press. Gertler, Brie. 2018. Self-Knowledge and Rational Agency: A Defense of Empiricism. Philosophy and Phenomenological Research 96: 91–109. Hasan, Ali and Richard Fumerton. 2020. Knowledge by Acquaintance vs. Description. The Stanford Encyclopedia of Philosophy (Spring 2020 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/ archives/spr2020/entries/knowledge-acquaindescrip/ Korsgaard, C.M. 2009. The Activity of Reason. Proceedings and Addresses of the American Philosophical Association 83: 23–43. McGeer, Victoria. 1996. Is “Self-Knowledge” an Empirical Problem? Renegotiating the Space of Philosophical Explanation. Journal of Philosophy 93: 483–515. McGeer, Victoria. 2007. The moral development of first-person authority. European Journal of Philosophy 16 (1): 81–108 Moran, Richard. 2001. Authority and Estrangement: An Essay on Self-Knowledge. Princeton, NJ: Princeton University Press. Moran, Richard. 2003. Responses to O’Brien and Shoemaker. European Journal of Philosophy 11: 402–419. Nisbett, Richard and Timothy Wilson. 1977. Telling More Than We Can Know: Verbal Reports on Mental Processes. Psychological Review 84: 231–259.

261

PART 6

Agency and moral psychology Introduction to Part 6 Luca Ferrero

Why does agency matter to us? A straightforward answer is that we value agency because it allows us to be efficacious in getting what we want: agency is the necessary and most general-purpose means to satisfy our desires, cares, and values. If this were all we cared about, what would really matter about our agency would only be its goal-directedness, possibly augmented by those additional features (like having a planning structure) that might increase its efficacy and effectiveness. Yet, this instrumentalist take appears to offer too narrow a view about what is valuable about our agency. For instance, this view is unconcerned about the origin of our goals, the nature of our capacity to acquire them, the various modes in which we might pursue them, and what having and pursuing those goals (and goals in general) would tell about what kind of beings we are, and why we should care about the kind of beings that we are. A richer understanding of the nature and role of agency in our lives shows that the agency we care most about is the one that supports our concerns for such things as freedom, autonomy, self-governance, and authenticity (on the self-regarding side), and accountability, responsibility, and morality (on the other-regarding side). At the heart of these features might lie not just our purposiveness but also and primary our capacity for being responsive to reasons and evaluations, possibly in a reflective and self-conscious mode. The chapters in this section explore some of the connections between our full-blooded agency and such notions as free will, autonomy, identification, responsibility, and some of the features of our psychology that support them, such as some of our emotions and our concern for agential and volitional unity. The chapters have been grouped under the label ‘moral psychology’ because this is the term of art currently used to embrace the various philosophical concerns about motivation, autonomy, responsibility, and related notions of self and identity. The term ‘moral psychology’ is somewhat misleading and confusing for those who might encounter it for the first time. First, moral psychology is not a branch of the science of psychology; it is not a matter of empirical investigation but of philosophical reflection (although one that can and should be empirically informed). Second, the topic of philosophical reflection is our ‘psychology’ in the broad sense of the distinctive configuration of properties, powers, and capacities that allow us to be (or at least to aspire to be) free, autonomous, responsible, and reason-responsive, and to have (or at least aspire to have) such things as (to use again some terms of art) an DOI: 10.4324/9780429202131-30263

 Luca Ferrero

‘agential standpoint’ or a ‘deep self.’ Many of these properties, powers, and capacities are in large part mental or psychological since they are the features of subjects with complex minds and a sense of self. But we should be wary of taking too an inward-looking approach to the philosophical investigation of these features. Ultimately, these are the properties, powers, and capacities of agents who are also embodied and social, possibly necessarily so. So the investigation of this constellation of properties, powers, and capacities is not just about (and in some cases not even primarily about) the internal configuration of what is sometimes called our ‘psychic economy.’ The ‘moral’ qualification in the term ‘moral psychology’ might indeed help counterbalance the excessive inward pull of this investigation, but it also risks overshooting. The concerns about motivation, autonomy, identification, self hood, freedom, and reason-responsiveness are not necessarily of a moral nature—at least at the first stage of philosophical investigation (whether we could go from full-blooded agency all the way to morality is still one of the open questions, as shown by the discussions in the chapters ‘Agency and normativity’ and ‘Agency and morality’ later in this volume.) So, what would be a better term to describe what this section is about if it is not really either about psychology or morality or their combination? In light of what I have just said, this section is about articulating the structure of full-blooded agency that bears directly on what we value about this very agency—a structure that is in large part, but not exclusively, a matter of our psychological make-up and organization as subjects who are mental, corporeal, and social. In a sense, this section is really about articulating the agential features of our agency. But ‘Agency and the agential configuration of the properties of mental, corporeal, and social subjects (with an eye to morality, maybe)’ is a mouthful and maybe not that informative either. So we will have to stick with the initial suggestion: ‘Agency and moral psychology.’

Agency, will, and freedom In ‘Agency, will, and freedom,’ Thomas Pink presents a long-standing debate about the nature and role of the will. There is a conflict between two competing models of the purposiveness of agency, of the relation between agency and content-bearing psychological attitudes, and ultimately of the nature of the powers of agency. According to one tradition (which comprises many Hellenistic philosophers, the medieval scholastics, and Kant) actions are performed through the exercise of a free will, which is a faculty with a specific kind of freedom. The will is the locus of motivations that, unlike the passions (including mere desires and emotions), are distinctively free and reason-involving. The will is the power of self-determined agency. The freedom of our agency as a whole depends on and is exercised as the freedom of the will. The opposing tradition (which stems from Hobbes) claims that the idea of a free will is a myth: the will is not required to explain either the nature of agency or its freedom. The idea of a free will exaggerates the metaphysical character of human agency and its distance from the rest of nature (especially from animal agency). According to this Hobbesian tradition, actions are not a distinctively practical kind of content-involving and internally goal-directed attitudes. Rather, actions are occurrences that bear no content and their goals are never internal to them. Actions are distinct from the attitudes that motivate them and that give them their purposiveness. The two conflicting models reflect a deep disagreement about the powers that give rise to agency. According to the Hobbesian tradition, the only power productive of human agency 264

Agency and moral psychology

is causal. Agency is the contentless effect of passive motivations, which cause actions and supply them with their goals. According to the other tradition, agency is a matter of distinctively practical attitudes, which are inherently goal-directed and move the agent in virtue of their contents. Agency is the product of a distinctive kind of power: the power of agents to determine actions for themselves, independently of their passive appetites and passions; a power of reason to move agents to act rationally, in response to the force of the justification that operates through the content internal to action. According to this tradition, as Pink summarizes it, “what moves an agent into action need not be a prior inclination. It can be the goodness of the action’s goal.”

Agency and responsibility In ‘Agency and responsibility,’ Pamela Hieronymi investigates how the idea of agency relates to the notion of responsibility and the practice of holding each other responsible. As she remarks in the opening of the chapter, ‘agency’ is a term of art, whereas ‘responsibility’ occurs in ordinary discourse and in a variety of different meanings. There are therefore many different ways in which agency might bear on responsibility. Hieronymi opens the chapter by offering a taxonomy of the senses of responsibility and later illustrates how agency might relate to them. The rest of the chapter is devoted to discussing what she takes to be the most interesting relationship between agency and what she calls ‘responsibility as mattering.’ This is the kind of responsibility at stake when we hold someone responsible because the quality of their will matters to us. This is what is as manifested by our responding to them with what Strawson calls ‘the reactive attitudes,’ such as gratitude, resentment, and indignation (these are opposed to the ‘objective attitudes’ which we adopt in response to things that do not manifest any quality of will or that we exempt from responsibility—to use Hieronymi’s example: “if a snowstorm forces a closure that absolves you of a distressing and burdensome task, you will feel relief. But if instead someone notices your distress and takes action to relieve you of the burden, you will feel gratitude.”) The ‘will’ whose quality we react to is comprised of the states of mind that manifest or embody the person’s take on the world. What we respond to with the reactive attitudes is the agency of the other person as this is embodied in the ways in which other people (ourselves included) figure into this person’s take on the world. What kind of agency (if any) is involved in the attitudes themselves? According to Hieronymi, philosophers have typically but mistakenly grouped reactive attitudes with sanctions and punishments. Even if the reactive attitudes might be used as sanctions, that is not their nature, she claims. Their agential status is different. Sanctions are negative consequences that are voluntarily created and imposed upon the wrongdoer for their violations. In contrast, reactive attitudes are natural human reactions that are not adopted at will (although communicating them might be a voluntary action). Even so, they are not simply involuntary reactions, like blinking. This is because these attitudes manifest or embody our takes on a variety of issues, such as “what is true, important, wonderful, admirable, offensive, contemptible,” none of which, however, depends on our will. A voluntary activity, by contrast, could only manifest our take on a specific question, that of what is all-things-considered worth doing. It does not follow, however, that reactive attitudes are only subject to explanation, and not to justification. We still enjoy a distinctive kind of agency with respect to the reactive attitudes: the agency we enjoy when we answer those questions (about what is true, important, worth doing, etc.) that shape our take on the world. 265

Luca Ferrero

Hieronymi closes by remarking that answerability, that is, being subject to a request to offer reasons for one’s attitudes and action, is also a kind of (often overlooked) responsibility, but one that is more fundamental than responsibility-as-mattering. Nonetheless, to be held answerable does not necessarily lead to being held responsible-as-mattering. There could be creatures who are capable of offering reasons in justification of their take on the world but such that they are not capable of prompting reactive attitudes (for instance, because they are not yet sufficiently mature or because they have been exempted from responsibility-as-mattering). Even so, the transition from answerability to responsibility-as-mattering need not be due to further facts about the agency of those involved, but rather to our sociability—to our ability to stand in complex forms of interpersonal relationships, where we matter to one another in distinctive ways.

Agency and identification In ‘Agency and identification,’ Agnieszka Jaworska discusses the philosophical views that account for the distinctive nature of full-blooded agency in terms of ‘identification.’ According to these views, some elements of the psychic economy of a person are more ‘deeply’ her own. These are the elements that represent her, that she identifies with, and that, as such, constitute her deeper self. This dimension of depth is lacking in simpler agents, and it is the basis of the distinctive responsibility and autonomy of persons. We are responsible for actions that originate from motives we identify with, and our autonomy is a matter of being governed by attitudes we identify with. According to Jaworska, there are two notions of identification at play in the philosophical literature: participant-identification and stance-identification. An agent identifies with an attitude or action as a participant when one sees the attitude or action as one’s own by being the subject of the attitude or the undertaker of the action. This is to be contrasted with the case of just perceiving or discovering that one has an attitude or is performing an action as a mere observer. According to the most prominent views of participant-identification, an agent identifies-as-participant with either attitudes that are acquired and revised or with actions that are performed out of one’s responsiveness to reasons and evaluations. By contrast, an agent stance-identifies with an attitude when that attitude is taken as a non-optional and inseparable element of the agential stance that one takes toward the world. (The distinction between the two kinds of identification is especially apparent in the case of weak-willed action: an agent participant-identifies with her weak-willed action—she is no mere observer of it—but she does not stance-identifies with it—the action is not essential to who she is.) Jaworska discusses at length the merits and demerits of different versions of stanceidentification. She contrasts the hierarchical theories (formulated in terms of wholehearted highest-order volitions, decisions, evaluations, or self-governing policies) with nonhierarchical theories (formulated either in terms of one’s carings or the integration of attitudes that define one’s character). She then argues that caring-based and integration-based views are able to avoid the criticisms that beset the hierarchical views, given that they do not require reflection upon one’s attitudes for identification. Jaworska concludes by speculating that stance-identification might be thought of as a kind of perfection or development of participant-identification.

Agency and autonomy What merely happens to an agent is not an action of hers. But among the actions that are attributable to the agent, it is not unusual to claim that some are ‘more’ of her own, and 266

Agency and moral psychology 

they are the actions that are ‘truly’ hers. In philosophical terms, these are the actions that are said to be genuinely self-governed or autonomous. In ‘Agency and autonomy,’ Andrea C. Westlund discusses various philosophical views about the conditions that make an action autonomous and why these conditions matter. A philosophical puzzle arises because autonomy seems to require more than the fact that some actions and choices are motivated by or expressive of attitudes internal to the agent’s own psychology. If this were all there is to autonomy, one would have to count as autonomous the actions produced by the like of addiction, compulsion, depression, brainwashing, indoctrination, and adaptive preference formation. The standard view of autonomy draws the distinction in structural terms: actions and choices are truly one’s own when we can ‘identify’ with them from the standpoint of the agent’s deliberative perspective, which is constituted by the agent’s self as a certain organization of various elements of the agent’s psychology. These actions and choices are to be contrasted with the heteronomous ones which stem from those components of our psychology we are ‘alienated’ from. These identificationist views focus on the ‘authenticity’ conditions of autonomy. They are attractive because they accommodate the intuition that to act autonomously is to be inwardly motivated. But Westlund argues that they ultimately fail because being authentic does not guarantee self-governance. In particular, identificationist views overlook cases where the problem is not alienation but excessive identification. Consider highly dogmatic agents (who appear to be supremely self-confident) or deeply deferential agents (who lack the disposition to engage critical challenges). What these agents lack is not a clearly defined self but a sense of responsibility for self. They make themselves impervious to the demand to be deliberately and reflectively answerable to reasons for their actions and attitudes. Identificationist views correctly point out that autonomy requires being motivated by attitudes that are internal to the agent’s justificatory perspective but they overlook the fact that autonomy also requires sensitivity to reasoning about the merits of that perspective itself. Westlund defends an alternative account of autonomy in terms of responsibility for self, centered on answerability. A key mark of autonomous self-governance is that one takes substantive responsibility for oneself: it is not up to anyone else but oneself to take care, in a reflective and critical manner, of the reasons one offers in the justification of one’s decisions and actions. Somewhat paradoxically, one is to hold oneself answerable in the face of external critical perspectives on one’s deliberative point of view. As Westlund writes: “the self-governing agent is not an isolated, self-enclosed center of activity, but rather, one reflective agent among many.” Westlund concludes by addressing the objection that autonomy might fail when one is too receptive to demands for justification. Her response is that responsibility for self requires some amount of confidence in one’s own judgment, even in the face of external challenges. Social and psychological challenges can undermine both receptivity and confidence, which shows that autonomous agents need to be adequately protected against such pressures. As she writes, “autonomous agency is precarious and vulnerable—which… makes it all the more important to concern ourselves with the conditions under which autonomous agency can develop and flourish.”

Agency and (the limits of) volitional conflict In ‘Agency and (the limits of volitional conflict),’ Sarah Buss argues that the metaphysics of agency allows us to experience only one kind of genuine internal volitional conflict—which is not to deny that there are other, possibly even more troubling, forms of psychic disunity that bear on our agency. 267

 Luca Ferrero

She begins with two uncontroversial observations: first, every agent is sufficiently unified to be a single agent, but, second, we all have many desires, values, and commitments that often support incompatible actions. It is thus unavoidable that we experience some psychic conflicts. Only one kind of conflict, however, is genuinely internal to our agency. To begin with, conflicts between desires and values are not by themselves volitional conflicts. Conflicts of goals and commitments, instead, are volitional, since they arise from constraints on the setting and the effective pursuing of ends. Volitional conflicts can often be resolved by revising or abandoning one’s commitments. We also face genuine normative dilemmas, where we are subject to more than one incompatible requirement at once, and it is impossible to revise our conflicting commitments because these are tied to our identity. In these dilemmas, we might find violating a commitment preferable to revising or abandoning it. These normative conflicts arise because of two different kinds of commitments. First, we have substantive commitments constitutive of our agential identity (our identity as end-setting agents). Second, we have an independent commitment to our identity as rational beings (as beings responsive to reasons and with a disposition to avoid incoherence). In cases of what Harry Frankfurt calls ‘volitional necessity,’ our all-things-considered normative judgments do not have the power to change our substantive commitments. Here we experience a genuine conflict internal to our agency: a conflict between requirements. As Buss argues, for an agent, her “substantive requirements can conflict with the specific ends manifest in her intentions because—and only because—they can conflict with what she takes herself to be required to do, all things considered.” She then argues that this is the only possible kind of genuine volitional conflict. First, one cannot regard any substantive requirement as overriding one’s all-things-considered normative verdicts. Second, one might not knowingly defy all-things-considered normative verdicts. What this implies is that, when we knowingly pursue an end while believing that we have overriding reasons to act otherwise, this is not a conflict within our agency. But the psychic disunity might lie elsewhere. For instance, we might avoid a conflict between our normative appraisals and our agency if we do not intend to be guided by these appraisals (which we can still apply to ourselves but only third-personally). This dissociation is not a form of volitional disunity. What we experience in this case is rather the suspicion that our will is weak because we might be operating on weak normative grounds, given our awareness that we might have foregone the opportunity to consider other potentially relevant considerations. But, at times, this dissociation is also necessary to avoid volitional paralysis (hence a sign of a strong, rather than a weak, will). We often face a trade-off between, on the one hand, forming an intention based on how things appear normatively to us at the moment and, on the other, reflecting on the actual justification of these appearances. Too careful and too frequent an examination of normative appearances might undermine our confidence in them and threaten our ability to form an intention and act on it. Hence, Buss concludes, “the more important challenge we face as agents is not the challenge of maintaining volitional coherence. It is, rather, the challenge of providing our will with an appropriate grounding without unduly interfering with our capacity to will.”

Agency and the emotions In ‘Agency and the emotions,’ Carla Bagnoli illustrates the complex relations between emotions and agency. She argues that emotions help with both intra- and interpersonal coordination of action and that, unlike passive or unreflective experiences, they can be recruited at 268

Agency and moral psychology 

the service of rational and moral deliberation. The most promising theories of the relations between emotions and agency are those that account for both the motivational and cognitive features of emotions. This is especially true of enactivist accounts, which conceive of emotions as temporal and action-oriented. Emotions play a variety of roles in the generation of action. They not only motivate but also help with the appraisal of the situation of action by offering criteria of salience and directing attention to the relevant features. The contribution of emotions, however, can also be more indirect. Emotions can prepare for action by shaping our imagination or by being activated in paradigmatic scenarios, which prepare us for specific courses of action. Emotions tend to be expressed and communicated in action. It is no surprise that across cultures there are norms that regulate the emotional expressive components of actions. These norms are central to many social interactions that play a role in the determination of the boundaries of social groups and communities. Susceptibility to emotions is often seen as hindrances to rationality and morality but Bagnoli argues that the opposite is often true: emotions help with both rationality and morality. Emotions help with the agent’s understanding and interpretation of her situation of choice and her self-understanding (which is not to deny that there are cases in which one’s deliberation suffer from emotional occlusion or obfuscation) Some emotions are directly tied to our status as agents, in particular ‘self-respect.’ Other emotions play a central role in our personal and moral relations, such as the reactive attitudes (e.g., blame and resentment) and the pair of love and hatred. Certain emotions (such as empathy and hope) and the phenomenon of emotional contagion help establish trust and shared agency. And it can be argued that some emotions (especially empathy) are constitutive of moral agency. Finally, emotions exhibit a temporal structure that differentiates them from whims and moods. Emotions might help us relate to different portions of our temporal existence (consider, for instance, regret for the past and anticipation of the future), and they unfold over time according to intelligible patterns of transformation and transmutation. The dynamics of these emotions should thus inform our understanding of the agency to which they contribute.

269

24 AGENCY, WILL, AND FREEDOM Thomas Pink

1 Agency and the will Agency includes not only the deliberate or intentional performance of actions, but deliberately refraining from them as well. I might act intentionally by moving in order to alert you to my presence. Or I might intentionally refrain from moving in order to prevent you from detecting me. In either case what distinguishes agency from mere changes in state or position or mere absences of such change seems to be purposiveness or goal-direction (aka goal-directedness) – the use of means to pursue ends. Even deliberate actions that have no further purpose are still described by us as done for a purpose. They are described as done for their own sake; their goal lies in their very performance. Since Aristotle (1985 at 1094a) many philosophers have seen purposiveness as essential to agency and as what distinguishes it from non-agency. The idea of purposiveness tells us what agency is – employing means to ends. And it supplies something key to its explanation. Why is she moving her arms like that? We discover that she is moving them like that in order to signal to us, and that purpose or goal explains what she does by leaving it, to some degree at least, intelligible. Purposiveness applies as much to the actions of the most primitive agents as to the actions of the most sophisticated. A shark is involved in agency akin to our own to the extent that it too is pursuing goals or ends, as when it turns this way and that in order to catch a fastmoving fish. But philosophers have often supposed that more sophisticated cases of agency, such as the deliberate actions and omissions of adult humans, have a special character that does not generalise to more primitive cases. Human agency involves the faculty or capacity of the will. Consider rationality or reasonableness. The actions of the lower animals, especially animals as unintelligent as sharks, seem to lie beyond reason. They are a-rational, or no more reasonable or unreasonable than is the swaying of a tree. But humans possess a capacity for rationality, and just as much as their beliefs, so too their actions can be reasonable or unreasonable. We can reason or deliberate and form mental states based on this reasoning. These mental states are attitudes, states with contents provided by objects of our thought that can come with justifications supporting the formation of the attitudes directed at them.

270

DOI: 10.4324/9780429202131-31

Agency, will, and freedom

In deliberating, we entertain these objects of our thought and can form the attitudes toward them that those objects justify. We respond to reason theoretically when we form beliefs directed at objects of thought as true – beliefs which are supported by the likely truth of those objects. We can also respond to reason practically as agents when we entertain and deliberate about potential goals of action – goals that are supported by their likely goodness or desirability. The attitudes that the goodness of these potential goals supports are motivations. They are decisions and intentions – attitudes whereby we become motivated to pursue ends and means on the basis of their goodness. Decisions and intentions are sensitive to practical deliberation about the good as beliefs are sensitive to theoretical deliberation about the true. They are motivations that supply our agency with its goaldirection, and which do so by being responsive to the agency-justifying goodness of their objects – of what is decided on and intended. The will has generally been identified as this capacity for reason-responsive motivation in the form of decision-making and intention-formation. Possession of the will is limited to those agents who are rational. Non-rational agents such as sharks, if such there are, may be capable of more primitive motivations, such as brute desires or passions or more primitive emotions. But these do not supply on their own any capacity to respond to reason as it governs action. Mere desires or passions are to be found in humans too, but in humans they are generally viewed as occurring outside the will because mere desires are often viewed as less sensitive to reason; in any case, their rationality is less crucial to the rationality of our agency than is the rationality of our decisions and intentions. Our desires incline us to act. But it is what we decide and intend that settles what we eventually do, and so it is on the rationality of our decisions and intentions that the rationality of our agency depends. The will has also been used to characterise another distinctive aspect of human agency. This is its supposedly free or self-determined nature. The agency of the more primitive lower animals such as sharks seems to be a simple product of their motivations – of brute passion or desire. But human agency need not be driven by desire alone. Adult humans can determine for themselves which actions they perform or refrain from. The further role of the will is to provide the motivation involved in agency that is self-determined. Self-determination seems to be a form of power. A power is a distinctive capacity – a capacity to produce or prevent outcomes. One familiar form of power is ordinary causation. A brick hits a window, and thereby exercises a power: the capacity to produce a dramatic change, so that the window breaks. We have here a case of power exercised as the force of ordinary causation. Another case of power is the self-determination attributed to human agents, the capacity to produce or prevent their actions for themselves. This capacity to determine for ourselves how we act is often termed as freedom and understood as a power over alternatives – a multi-way power to do A or to refrain and perhaps to do some further thing B instead. This power over how we act constitutes a form of control and is reported when we say that it is ‘up to us’ how we act. This power has often been located in the will. We can control our actions at all only because we can decide how we shall act, and we can determine for ourselves how we decide. Freedom of action in general depends on a freedom specifically of the will. So the will is a locus of motivations that are not only distinctively responsive to reason, but which are also peculiarly under our control as the passions and mere desires and emotions that we feel are not. The freedom of our agency as a whole has been thought to depend on and be exercised as a freedom specifically of the will.

271

Thomas Pink

2 Scepticism The will is a supposed locus of motivations that are both peculiarly free and also peculiarly reason-involving. What is the origin of such a conception? Its history has long been a matter of debate. The medieval scholastics assumed that Aristotle already had a conception of the will. But they mostly used translations of his work that employed a Latin terminology of the will, of voluntas and volo, to read into Aristotle a commitment to a faculty of will not obviously evident in his more terminologically varied Greek. Aristotle did seem to think of certain kinds of motivation as distinctively reason-responsive, notably the capacity for prohairesis, for decision or choice based on deliberation. But he did not obviously defend a full-blown theory of self-determination as requiring freedom of motivation at the point of the will. Some writers claim that a theory of the will arises only after Aristotle, in Stoicism and in Christian thinkers such as Augustine. There is an extensive literature about this history (see Kahn 1988), as well as controversy about whether this is a history of the will’s discovery or rather of its invention. For many philosophers now suppose that there is no such thing as the will as traditionally conceived, as a capacity for motivation that is free or self-determined. One of the first modern sceptics was Thomas Hobbes, who thought that the very idea of a freedom of the will was an invention of Catholic scholasticism (see Hobbes 1999). There have been many other sceptics since, ranging from Nietzsche to Gilbert Ryle (1949) and Bernard Williams (1995). The sceptics all agree that belief in the will is based on and fuels exaggeration of the division between humanity and the rest of nature, especially between human agency and the agency of the lower animals. The sceptics even claim that belief in a free will is viciously regressive. Everyone admits that human and animal actions alike occur as uncontroversial cases of what Hobbes termed the voluntary – the doing of things, such as moving one’s body and its limbs, on the basis of pro attitudes to so doing, such as desires or decisions and intentions to move one’s body. But in the case of human agency the theory of the will seems to characterise such familiar actions by appeal to yet further actions. The will, traditionally conceived, involves the formation of motivations, decisions, and intentions to act, that are supposedly directly controlled and determined by us, and are even more reliably within our control than the familiar actions thereby motivated. That surely leaves these motivations as yet more cases of free and self-determined action. Scholastic action theory was quite explicit about its commitment to an agency-generating agency of the mind. It regarded decisions of the will as the primary case of agency – what was termed the elicited agency of the will itself – leaving the voluntary actions decided upon, the actions of deliberately moving one’s limbs and the like, as mere secondary effects or products (commanded actions) of these elicited actions of the will. The commanded actions decided upon were controlled by us only through our immediate control of the elicited actions of the will itself (see Pink 2004). But then we are explaining the familiar agency of the voluntary by appeal to yet further, and rather more mysterious, action-generating actions within the mind. How can postulating yet more agency in some ghostly internal form shed any light on what agency is; and once begun, why should this account of action in terms of action-generating action ever stop? (Versions of this argument are in Hobbes 1999 and Ryle 1949.) Behind this debate about a freedom and agency of the will lie two competing models of agency’s goal-direction or purposiveness, and related to these models a controversy about the nature of power – the capacity to produce or prevent outcomes – and about the kinds of power that produce human actions. 272

Agency, will, and freedom

3  Goal-direction and intentionality A goal is an object of thought – an aspect of the mental, just like an object of belief, or desire, or emotion, or other attitudes. I might, for example, pursue a banking career as a means to getting rich. That I become rich is the goal of what I do, but a goal that may never be attained. It may never amount to anything more than an unrealised object of my thought. The goal-direction of actions and omissions is an aspect of intentionality – of our mental capacity to become directed, through psychological attitudes, at content-providing mental objects. We need to locate agency with its goal-direction within the wider phenomenon of intentionality – our capacity to hold attitudes of various kinds directed at content-providing objects. The idea of agency at the point of the will involves a very simple intuition. Action occurs as a mode of intentionality that is distinctively practical. There is a mode of intentionality that is cognitive, as when we are directed through belief at an object as true. Then there is a mode of intentionality that is desiderative, as when we are directed through desire at an object simply as good or desirable. But there is also a mode of intentionality that is distinctively practical, when we are directed at an object not simply as good or desirable, but as our goal. This mode of intentionality constitutes agency, and it is to be found in motivations of the will. Notice that goal-direction is conceived of in this theory as internal to agency. Agency arises as a distinctively practical content-bearing attitude, in decisions of the will, and its immediate goal-direction can be provided by its own content, just as the object of a belief or a desire is internal to it, as provided by its own content. The immediate goal of a decision to raise one’s arm is provided by its own content – that one raises one’s arm. Whatever further purposes it might serve, the decision to raise one’s arm is taken in order to do just this. There is, however, a quite different theory of goal-direction and its motivation. This simplifies the theory of intentionality by denying that intentionality ever takes a distinctively practical or agency-constitutive form. Agency is one thing, and attitudes are another. Actions occur not as a distinctively practical kind of content-involving attitude, but as occurrences, such as limb motions and the like, that are not content-bearing. The goal of an action is never internal to it, a content that it possesses as an attitude in its own right. The goal comes to the action from without, from attitudes whose contents provide the action with its goal by inclining the agent towards the performance of such an action. Perhaps I want to signal and believe that raising my hand would signal. These attitudes with their contents incline me to raise my hand by leaving that action an apparent means to something I desire. Their contents can thus supply my hand-raising with a goal. That view of goal-direction lies at the heart of Thomas Hobbes’s identification of agency with voluntariness – with doing things on the basis of pro attitudes towards doing them. Actions are performed because they are desired means to various desired ends and get their goal direction from the contents of the prior attitudes that so leave them objects of our desire. A classic modern exposition of this view is given in Davidson (1980).

4  Goal-direction and motivating power There is a further element to agency that is closely linked to its goal-direction and that is often treated as the same thing. This is our agency’s motivation – what moves an agent to perform an action or omission. Not only are actions and omissions purposive and directed at goals, but very often something moves or motivates the agent into so acting or refraining – something linked to that goal-direction. Motivation is more than goal-direction in itself. Goal-direction is simply the 273

Thomas Pink

direction of an action or omission at an object. Motivation seems to involve some sort of power that gets or inclines the agent to act or omit in a way directed at that goal. The power involved seems not to be that of self-determination or freedom. The power, after all, is not being exercised by the agent. Rather, the power is one to which the agent is subject, along with his action. The agent is being moved, so is subject to a power being exercised by something else so as to get him to act towards the goal. Beyond simple goal-direction then there is also motivation understood not simply as direction at a mental object, but the exercise by something of a power to get the agent to direct their action at that mental object. This power is not only distinct from the power of self-determination, but in some tension with it. An agent may not be able to determine for himself what he does if he is subjected to a power to move him that is too great – that compels him into action. What kind of power might be moving us to act, and what might be exercising it over us? One obvious possibility is that it is a power exercised over us by our own attitudes and that the power exercised is a form of ordinary causation. Decisions, intentions, desires, and other pro attitudes, guided by beliefs, can all be motivating causes of the actions they explain. Even if, unlike Hobbes, we do admit that an action’s immediate goal-direction can be internal to it as a content-bearing attitude in its own right, we do seem to allow for additional purpose to come to the action from without, from prior motivating attitudes. Moreover, we do seem to view these motivating attitudes as causes. Consider the way decisions to attain ends provide an initiating goal-direction or purpose that gives rise to further decisions to employ means. I decide to go on holiday to Spain and that decision then exercises a causal power to move me to decide to visit the travel agent in the next street. The decision caused may indeed have a goal internal to it, from its own content – that I visit the travel agent. But it is also taken as a means to ensuring that I go on holiday to Spain. And that is because I have been moved to take it by an earlier decision to go on holiday to Spain and my belief that visiting the travel agent is a means thereto. My ultimate purpose in deciding to visit the travel agent comes then from its being part of a wider plan for going on holiday to Spain. The will’s role in means-end planning embeds decisions to adopt means with their immediate purpose (to secure those means) within prior projects of attaining ends. Central to this function is the operation of causal power. After all, as Michael Bratman (see Bratman 1987) and others have persuasively argued, the point of taking decisions about what ends we shall pursue is to enable us to coordinate the rest of what we do with those intended ends. If we are to adopt the right means to ends, it had better be settled which ends we shall in fact be pursuing, lest (for example) we start paying deposits on hotels in one country when it is still open and unsettled that it really will be to that country that we are going. But that requires that our initial decisions to pursue ends must exercise a force or power to remove the chance of our acting otherwise – at least for as long as there is no change in the information on which those decision were originally based. Those initial decisions to attain ends then exercise their power to guide subsequent decisions to adopt means, just as they provide those subordinate decisions with their ultimate purpose. We also seem to view desires and emotions outside the will itself as similarly motivating us towards action through the exercise of causal force. Sometimes this force lies behind or reinforces the operation of the will. I decide to go on holiday to Spain because from the very start that has been something I very much wanted to do. Sometimes the force of certain desires conflicts with that of our will, as when I struggle to stick to a difficult decision, such as to revise for an exam, that conflicts with other things that I want. This is a struggle because of what seems to be the resistance offered to my decision by the opposing causal force of my desire.

274

Agency, will, and freedom

On Hobbes’s view, goal-direction only ever comes to agency from outside – from the contents of motivating attitudes distinct from the agency they explain. Given our tendency to conceive of motivation by prior attitudes as the exercise of causal power, it is unsurprising that Hobbes and his modern successor Davidson view these motivating attitudes as causes. Agency, according to Hobbes, always occurs as an effect of psychological causes – of motivating attitudes that, because they are attitudes, are not actions themselves. Action-explanatory motivations are always passions – appetites that happen to the agent, rather than decisions that are his own free or self-determined doing. Goal-direction presupposes motivation, and agency is by its nature an effect of non-agency. Contrast the view that agency occurs as a distinctively practical attitude. On this view, there is no necessity that a deliberate action must be an effect of some motivation that is passive. A deliberate action need not be motivated by any prior attitude. The agent is indeed acting to attain a goal. But nothing need have motivated him so to act. An agent could just decide, say, to continue a walk rather than go home, without any desire having motivated him so to decide. The decision has its own content that provides it with an immediate goal – that the walk continue – which the agent simply adopts by taking the decision. Some desire might have been present to motivate the decision. But the model of agency and its goal-direction does not demand this. An action gets an immediate goal-direction from its own content, so that this goal-direction does not come from motivation provided by prior attitudes. On this view opposed to Hobbes, agency is not by its very nature an effect of non-agency – of motivations that are passive. On the theory of agency as a practical attitude, actions may indeed be performed through the influence of motivating passions. But this is not essential to their character as actions. Actions can also arise from the will independently of passion. The medieval scholastics speculated that there may even be agents above the level of humanity who lack passive motivations entirely. There can, in God or the angels, be such a thing as an agency of pure will. Hobbes’s view of agency was an important component of his scepticism about selfdetermination. Hobbes flatly denied that freedom existed as a genuine power of selfdetermination. The only power in nature, and so the only power the operation of which could ever give rise to actions and omissions, was ordinary causation. This was a power exercised not freely by agents themselves, but quite mechanically by motivations within them – by the passive appetites that guided agents in their deliberation, which thereby determined and settled how they would act and which exclusively provided the goals at which their agency was directed. This theoretical tension between the Hobbesian model of agency and the idea that agents determine for themselves what they do persists to our day. One solution is that of modern compatibilism about free agency. This resolves the tension through a form of denial. This compatibilism simply identifies the agent’s power of self-determination with what Hobbes took to replace it – the motivating power of passive attitudes. We are acting freely to the extent that we act as we desire. But this proposal faces an obvious and immediate objection. We are identifying a power which the agent is supposed to exercise himself with a power to which (by contrast) he is obviously subject; but these must surely be quite different forms of power. Another solution is to retain the Hobbesian idea that agency presupposes motivation by prior attitudes, but to deny that motivation implies any form of freedom-threatening exercise of causal power by those motivations. The modern literature contains ingenious theories of how passive attitudes such as desires can motivate actions without causing them (see, for example, Sehon 2005). Aside from any technical problems afflicting such proposals,

275

Thomas Pink

they face a deeper difficulty. This is the extent to which common sense psychology does seem to understand the relationship of prior attitudes, whether of decision or mere desire, to the subsequent actions that they motivate as involving the exercise of a power to move, and to understand this power in causal terms. Consider again the deeply causal nature of our understanding of how decisions settle in advance how we shall act and do so by generating further decisions about means.

5 The power of reason There is a form of motivation quite different from causation by prior attitudes. This is motivation by reason itself. Reason does not simply provide a standard by which we are appraised as more or less reasonable. It provides a mode of direction to which as beings capable of reason we suppose ourselves to be responsive. Reason can move us, a conception of reason that includes the practical, as Christine Korsgaard has insisted: ‘Thus it seems to be a requirement on practical reasons, that they be capable of motivating us’ (Korsgaard 1997, 317). The theory of action as a distinctively practical attitude allows not only for goal-direction apart from the agent’s subjection to any motivating power but also for motivating power to take a specifically rational form – not as a causal power of attitudes, but as a power of reason motivating the action through its own content. How does reason move a rational agent to decide to pursue a given goal? It does so through the justifiability of attaining it, which amounts to its goodness as a goal. The goal is a mental object, an object of our thought, whose goodness is a normative property that attracts us to adopt it as our goal by deciding to attain it. And so we arrive at the theory of rationality and the motivating power of reason defended within scholasticism, and with especial ingenuity and sophistication by Francisco Suarez in his Metaphysical Disputations of 1597 (for discussion, see Pink 2018). When we are moved by reason, what moves us to form attitudes are the objects of our thought, and they do so by exercising a kind of normative power over us – a force of reason based on the degree to which those objects do indeed justify the attitudes they produce. A potential object of belief is evident, or likely to be true – and this normative property of evidence or being theoretically justified moves us to believe it. A potential object of decision and intention is both good and likely to be attained through deciding on it – and this property of being justified practically as a goal moves us to decide on it as our goal. The power of reason to move us is explained as a normative power very different from ordinary causation. The power is not being exercised by an actual occurrence but, in the case of a possible goal, by an object of our thought that may never be realised. And the power is the power of a value, truth in the theoretical sphere, goodness in the practical sphere. Just as Hobbes denied the possibility of a power of self-determination, so too he denied the possibility of any such normative power. There was no force of goodness that potential goals could exercise over us as objects of our thought by attracting us to decide on them as our goals. Motivation could only come from prior attitudes exercising ordinary causation. When we use terms such as ‘good’ we are not reporting a force of reason exercised by mental objects that could produce and explain our motivations. Rather than explaining our motivations, reference to goodness simply presupposes them. Talk of goodness is merely expressive of motivations that we already have, and which are effects not of anything normative but of other quite natural causes: But whatsoever is the object of any mans Appetite or Desire, that is it, which he for his part calleth Good; and the object of his Hate, and Aversion, Evill; and of his Contempt, Vile and 276

Agency, will, and freedom

Inconsiderable. For these words of Good, Evill, and Contemptible, are ever used with relation to the person that useth them: There being nothing simply and absolutely so; nor any common Rule of Good and Evill, to be taken from the nature of the objects themselves… Hobbes (2012, 80–81)

6 The place of the will in action theory We now see how controversy about the reality of a distinctive faculty of will is ultimately about the powers that give rise to agency. On Hobbes’s view, motivations are always needed to explain actions as their causes, thereby supplying those actions with the objects that, as contentless non-attitudes, they would otherwise entirely lack. The nature of agency is to be the contentless effect of motivations that are passive. There is no place in Hobbes’s action theory for a faculty of will. There is no room for motivations that count as actions simply in virtue of their own content and their mode of direction at that content. Such active and inherently goal-directed actions are anyway not needed as potential products of some power other than a causal power of passions. There is no such other action-productive power. The view that agency occurs as a distinctively practical attitude, by contrast, leaves room for the forms of action-productive power that Hobbes rejected – a power of agents to determine actions for themselves, and a power of reason to move agents to act rationally. If an action can possess its goal internally, rather than through its production by a content-bearing passion, there can be actions that agents determine independently of their passive appetites, and so quite genuinely ‘for themselves.’ And agents can respond to reason in the exercise of their power of self-determination in a way similarly independent of passive appetites. What moves an agent into action need not be a prior inclination. It can be the goodness of the action’s goal.

Related topics Agency and causation; Mental agency; Animal agency; Rational agency; Agency and autonomy; Agency, reasons, and rationality; Agency and responsibility.

Further reading Hobbes, T. (1999) ‘Of liberty and necessity’ in ed. V. Chappell, Hobbes and Bramhall on Liberty and Necessity (Cambridge: Cambridge University Press), pp. 15–42. Hobbes’s classic and highly influential attack on the idea of a freedom of the will. Kahn, C. (1988) ‘Discovering the will from Aristotle to Augustine’ in ed. J.M. Dillon and A.A. Long, The Question of ‘Eclecticism’ (Berkeley: University of California Press), pp. 234–259. A famous historical account of the development in late antiquity of our conception of the will. Pink, T. (2017) Self-Determination (Oxford: Oxford University Press). This develops the themes of this article in much more detail.

References Aristotle (1985) Nicomachean Ethics, trans. T. Irwin (Indianapolis: Hackett). Bratman, M. (1987) Intention, Plans and Practical Reason (Cambridge, MA: Harvard University Press). Davidson, D. (1980) ‘Actions, reasons and causes’ Chap 1 in Essays on Actions and Events (Oxford: Oxford University Press): 3–19. Hobbes, T. (1999) ‘Of liberty and necessity’ in ed. V. Chappell, Hobbes and Bramhall on Liberty and Necessity (Cambridge: Cambridge University Press): 15–42.

277

Thomas Pink Hobbes, T. (2012) Leviathan, ed. N. Malcolm (Oxford: Oxford University Press): volume 2. Kahn, C. (1988) ‘Discovering the will from Aristotle to Augustine’ in ed. J.M. Dillon and A.A. Long, The Question of ‘Eclecticism’ (Berkeley: University of California Press): 234–259. Korsgaard, C. (1997) ‘Skepticism about practical reason’ Chap. 11 in Creating the Kingdom of Ends (Cambridge: Cambridge University Press): 311–334. Pink, T. (2004) ‘Suarez, Hobbes and the scholastic tradition in action theory’ in ed. Thomas Pink and M.W.F. Stone, The Will and Human Action from Antiquity to the Present Day (London: Routledge): 127–153. Pink, T. (2017) Self-Determination (Oxford: Oxford University Press). Pink, T. (2018) ‘Agents, actions and their objects in Suarez and Hobbes’ in ed. Constantine Sandis, Philosophy of Action from Suarez to Anscombe (London: Routledge): 1–28. Ryle, G. (1949) The Concept of Mind (London: Hutchinson). Sehon, S (2005) Teleological Realism: Mind, Agency and Explanation (Cambridge MA: MIT Press). Williams, B. (1995) ‘Nietzsche’s minimalist moral psychology’ Chap.6 in Making Sense of Humanity (Cambridge: Cambridge University Press): 65–76.

278

25 AGENCY AND RESPONSIBILITY Pamela Hieronymi

“Agency” is a term of art. Its meaning and use might be discovered by reading and studying this volume. In contrast, “responsibility” is an everyday term, one that regularly appears in kitchen-table conversations and on the pages of newspapers. The present chapter is meant to relate them.

1 Responsibility Let us start by examining the relatively wide variety of everyday uses of “responsibility” or “responsible.” In one, the word could be replaced, without loss of meaning, by “cause:” “The extreme heat was responsible for the engine failure” is equivalent to “the extreme heat caused the engine failure.” In another, the word could be replaced, instead, by words such as “duty,” “obligation,” or the more prosaic “job.” “Feeding the dog is my responsibility” means “Feeding the dog is my job” (or duty, or obligation). (In contrast, causing the failure was not the job of the extreme heat.) Relatedly, the thought expressed by, for example, “Maria is very responsible” could instead be expressed with, “Maria is conscientious; she takes her obligations seriously.” Someone who is irresponsible cannot be trusted with a job. In a third and more complicated sort of use, “responsible” could be replaced with “at fault for” or “to be thanked for,” as in “Who is responsible for the delay?” or “Who is responsible for this lovely reception?” Here, it seems we mean something like, “Whose actions or omissions foreseeably caused this (bad or good) outcome?” As is often the case with ordinary notions, these uses overlap and combine. Sometimes, in saying someone is responsible for something, we mean that they have an obligation, or duty, or job, because they are at fault for some bad outcome—as in (certain uses of ), “Who is responsible for this mess?” and “The mess is John’s responsibility.” John made the mess, and now it is John’s to clean up. (We do not use the word in this way in cases in which someone is to be thanked or rewarded, presumably because, in such case, it is we, not they, who have the obligation—of gratitude or of recompense.)

DOI: 10.4324/9780429202131-32279

Pamela Hieronymi

Sometimes, though, rather than impose upon John a duty or obligation of clean up, we impose upon John some fine, penalty, or punishment. In such cases, it seems to be because John’s actions or omissions foreseeably caused some bad outcome that some burden can now be aptly imposed upon John—he can be fined for the mess because he is responsible for it. If we are in the business of imposing a penalty or burden on John because he is at fault for the mess, we might justify our doing so by saying, “he is a responsible adult.” In this case, we are not saying that John is conscientious or trustworthy (like Maria). Rather, we are saying that he is such that he can be aptly fined, penalized, or punished for his errors. He is such that he can be, as we say, held responsible. (In fact, it is sometimes suggested that to fail to burden John for his malpractice would be to fail to treat him as a responsible adult—to treat him with disrespect.1) Holding responsible is something we do to one another. One way in which we do it is by making clear that an action or omission creates some obligation—by making clear our expectations of clean up, for instance. Another way, as just considered, is by imposing penalties, sanctions, or punishments. However, there is an additional, more complex or subtle way in which we hold one another responsible, one which has been the subject of voluminous philosophical reflection over the last six decades: we stand ready to respond to others with what Peter Strawson called “reactive attitudes.”2 Strawson introduced the term “reactive attitudes” in his landmark paper, “Freedom and Resentment.” These are attitudes or emotions, like resentment, gratitude, and indignation, which we form in response to what Strawson called the “quality of will” that we perceive in another. They contrast with an “objective attitude,” which we adopt in response to things that do not manifest a will of any quality. To illustrate: while you might be frustrated, even angry, if you find that a stray nail has punctured your tire and left you stranded, you will feel differently if you find that someone has slashed your tire on purpose. You will resent the slashing. (If you resent the nail in the same way, you may recognize this is a mistake, a kind of animating of the physical world.) Likewise, if a snowstorm forces a closure that absolves you of a distressing and burdensome task, you will feel relief. But if instead someone notices your distress and takes action to relieve you of the burden, you will feel gratitude. (You would feel grateful for the snowstorm if you thought it orchestrated by a divine being with you in mind.) Resentment and gratitude are reactive attitudes—they are, as Strawson puts it, “natural human reactions to the good or ill will or indifference of others towards us, as displayed in their attitudes and actions.” When contrasted with these, anger and relief are more objective attitudes. When we respond to the quality of another person’s will with reactive attitudes, we are, in an important sense, holding that person responsible, treating them as a responsible adult. If instead we fail to react, or react with more objective attitudes, we are, in an important sense, exempting that person from responsibility. Strawson noticed three kinds of cases of exemption, three kinds of cases in which, when someone shows us malice or disregard, we respond in a less reactive, more objective, way. First, we might do so if we learn the person was in extreme or unusual circumstances. We might then simply ignore the ill will, treating it, so to speak, as the kind of thing that happens when people are under extreme stress. Second, we do not react in the usual way when interacting with people who are in some way incapacitated for ordinary interpersonal relationships. Strawson’s examples are immaturity and pathology. Finally, Strawson noted that we sometimes adopt an objective attitude for more pragmatic reasons: for therapeutic purposes, or out of scientific curiosity, or to avoid what he called “the strains of involvement.” 280

Agency and responsibility

The last should be familiar. Sometimes, in our ordinary relationships, we encounter an especially difficult or trying person. We sometimes then opt out, so to speak, of ordinary interpersonal engagement: We treat that person as an “issue” or “challenge,” someone to be simply handled or managed. In recent philosophical discussion, responding to someone with negative reactive attitudes has been called “blaming,” and being the appropriate target of such attitudes has been called being “blameworthy.” These labels are unfortunate: the natural use of the English words “blame” and “blameworthy” do not pick out what Strawson had in mind, but rather what I called above being “culpable” or “at fault”—foreseeably causing some bad outcome for which you now rightly incur either obligations or vulnerability to some sanction or punishment.3 Perhaps because of the unfortunate label, philosophers have typically grouped reactive attitudes with sanctions and punishments, as though they were a burdensome treatment imposed upon the wrongdoer by the one reacting.4 This is a mistake. It overlooks a crucial difference in the agency at work. It is tempting, even natural, to group the reactive attitudes with sanctions and penalties in part because they are (unfortunately, to my mind) often used as sanctions or penalties. Frequently enough, people wield their reactive attitudes, through their voluntary actions and communications, aiming to burden the wrongdoer—intending or hoping to make them feel badly for what they have done. Such “guilt-tripping” is not only possible but prevalent. Yet it does not show that the reactive attitudes, themselves, should be understood as sanctions or penalties: they should not. To see this, note, first, that a sanction, penalty, or punishment is not simply a negative consequence that predictably follows from some action: a hangover is not a sanction. Neither is the pile of work, accumulated thorough procrastination, that now must be done by tomorrow. Sanctions, penalties, and punishments are, instead, negative consequences that have been created and attached, by somebody (or some body), to certain violations. In response to such violations, sanctions, penalties, and punishments are imposed through voluntary action. In contrast, reactive attitudes are not voluntary. Though, of course, one might communicate or express them via a voluntary action, the attitudes are not, themselves, adopted at will. The reactive attitudes are, as Strawson noted, “natural human reactions” to our perception of the quality of another’s will. Being the target of resentment is thus, in one way, more like having a hangover than being fined or sent to your room: it is not a burden voluntarily imposed in response to your misbehavior, but rather a natural consequence of your disregard, manifested in another. However, while being the target of someone’s resentment is more like having a hangover than being fined, it is also importantly unlike having a hangover. These natural human reactions, though not voluntary actions, are also not simply involuntary reactions, like flinching or blinking. To treat the reactive attitudes simply like a hangover or a flinch would be to misunderstand the agency of the one who reacts—it would be to put the reactor in bad faith. We will return to this.5 To summarize, the word “responsible” or “responsibility” is used in a variety of ways. Sometimes it simply notes causal relations. Sometimes it indicates obligations or duties, or conscientiousness with respect to one’s obligations or duties. Sometimes it indicates an obligation or duty that was created by some action or omission, as when one makes a mess and now must clean it up. Sometimes, to be “a responsible adult” is not just to be liable for cleaning up one’s messes, but also to be open to penalties, sanctions, and punishments for one’s actions or omissions. And, finally, we are, in an important sense, holding a person 281

Pamela Hieronymi

responsible when we respond to the quality of their will with what Strawson called the reactive attitudes. To be responsible, in this last sense, is to be such that the quality of your will matters to others in the usual way. For lack of a better label, I will call it “responsibility as mattering.”6

2 Agency Turning, now, to the term of art: “agency.” We can already see a variety of different ways in which an investigation of human agency might interact with questions about responsibility and a variety of different things one might have in mind if one wanted to focus, narrowly, on “responsible agency.” Starting with the first use of “responsible,” an investigation into human agency might have something to say about which of the indefinitely many events causally downstream of my action count either as part of my action or as salient causal consequences of my action, such that I am responsible for them in the causal sense. To take a well-worn example, if I am moving my arm in order to operate the pump, in order to replenish the water supply, in order to poison the inhabitants of the house, in hopes of hastening the end of the war, then it seems that, in moving my arm, I am therein pumping and therein both replenishing and poisoning the water supply.7 If the inhabitants are poisoned, I will have poisoned them. However, poisoning the inhabitants seems merely a consequence—an intended consequence—of my action of poisoning the water supply, that is, it does not seem that moving my arm just is poisoning the inhabitants, in the way it just is pumping the water and poisoning the supply (even less that it just is hastening of the end of the war). An investigation into the metaphysics of actions and omissions may help us to draw lines between action, intended consequences, hoped-for outcomes, and other downstream causal consequences. (Of course, it is possible that conclusions about responsibility will instead help to determine the boundaries of human action.) Moving to the second idea, that of being “at fault,” we might focus investigations of agency on the ability to foreseeably cause outcomes one should understand to be good or bad. This will require some ability both to foresee and to evaluate that which one will cause. Certain sorts of outcomes (such as the lovely reception) will, in addition, require an ability to form and execute complex plans.8 Such agency goes beyond what might be thought of as the most basic sort, for example, that displayed in simple organisms pursuing biological needs.9 We saw, next, that sometimes those at fault for a mess incur an obligation to clean it up, and sometimes they are (perhaps in addition) liable to sanction, penalty, or punishment. An investigation of human agency might yield insights relevant to the ethical question of whether a person at fault for a bad outcome can aptly or justly be burdened with an obligation, penalty, or punishment. To illustrate, the mess may be the child’s fault—they foreseeably caused it—and yet the resulting mess may be more than the child can be asked to clean: they lack the required capacities. In this case, facts about the child’s capacities as an agent serve as input into ethical reasoning about where obligations lie. For a second illustration, it seems unfair to sanction or penalize a person for a violation if they lacked an adequate opportunity to avoid it. Thus, again, investigation into our capacities for agency—for guiding ourselves in such a way as to avoid violations—might serve as input into ethical reasoning about when sanctions, penalties, or punishments are fair. One version of the traditional “problem of free will and moral responsibility” can be seen as a particularly fraught result of this last sort of investigation: On a natural line of thought, 282

Agency and responsibility

a moral sanction or punishment is fairly imposed or truly deserved only if the wrongdoer “could have done otherwise.” However, it seems to some that investigation into our agency reveals that we never have the ability to have done otherwise. Such thinkers conclude that we are never rightly deserving of moral sanction or punishment, and thus never morally responsible.10 While such skeptics believe that facts about human agency undermine moral responsibility, others believe that features of morality will guard against the skeptical conclusion. Many of these non-skeptics agree with the skeptic that facts about agency and facts about moral requirement must fit together so as to ensure we are capable of avoiding moral w rongdoing— in a slogan, they agree that “ought implies can.” However, these non-skeptics then part company with the skeptic in at least two different ways. Some reason that, since we evidently are morally responsible, we must have the relevant ability (using “ought implies can,” they draw conclusions about agency from facts about moral requirement).11 Others believe that morality will stretch or shrink, so to speak, to accommodate the facts about agency, whatever they turn out to be.12 While I am sympathetic to the thought that morality can accommodate the facts, whatever they turn out to be, I doubt that agency and moral requirement must fit together in the way suggested by “ought implies can.” Humans are born into the world in need of moral development and, tragically, such development often goes awry. People thus often arrive at adulthood selfish, petty, insensitive, ruthless, and so forth—people arrive at adulthood without the capacities required to satisfy moral demands. Yet those demands do not, for that reason, yield or shrink to fit. It seems common that moral demands rightly apply to people who lack the capacities required for their satisfaction. We are, we might say, subject to original sin.13 If we are, as I suspect, subject to original sin—if we are not always able to satisfy moral demands—this should inform our understanding of sanction and punishment. If, as the skeptic maintains, our ideas about sanction and punishment somehow presume that we can always avoid moral wrongdoing, then we may have to reinterpret them to avoid that false presumption.14 It seems to me that this could be done—the fact of original sin need not show that sanctions or penalties are always unfair.15 To my mind, the final sort of responsibility, “responsibility as mattering,” provides the most interesting material for investigations of agency. Reflection upon it prompts us to ask, not “what kind of agency is required to incur obligations or be fairly penalized?” but rather, “to what kind of agency do we aptly respond with reactive attitudes?” And, “what kind of agency (if any) is involved in those attitudes, themselves?” And, finally, “why do we not respond with these attitudes to children or those subject to certain kinds of pathology or those in extreme or unusual circumstances—is this because of a limitation in their agency, or does the explanation lie elsewhere?” Here is my own answer to the first and second of these questions, and a hunch about the third. The reactive attitudes are reactions to, as Strawson put it, the quality of another’s will. That “will,” as I would understand it, is comprised of those states of mind that manifest or embody one’s take on the world—one’s take on what is true, worthwhile, to be done, threatening, delightful, awe-inspiring, and so forth. A person’s will prominently includes their beliefs and intentions, but also their trusts and distrusts, admirations and contempts, cares and concerns, and so forth.16 The reactive attitudes react to one particular quality of a will: the way in which other people figure into a person’s take on the world— whether, for example, the interests of others are worth heeding, whether their needs are worth protecting, whether they are treated as of equal importance, and so forth. Thus, 283

Pamela Hieronymi

when one person is held responsible by another person, in the responsibility-as-mattering sense, one set of attitudes, one person’s take on the world and their place in it, generates a reaction in a second set of attitudes (which often prompts a further reactive change in the first, etc.). We next ask about the agency we enjoy with respect to these reactive attitudes—and the answer is not straightforward. On the one hand, we are not simply passive with respect to them. Resenting or admiring is not like having a headache, seeing spots, or catching a cold. In manifesting or embodying our take on things, these attitudes are, in some way, up to us. On the other hand, this range of attitudes is not voluntary—we cannot simply adopt them at will, in the way we can (if able-bodied and well-resourced) raise our right hand, dance a jig, or cook a meal at will. Importantly, non-voluntariness is essential to these attitudes: any state of mind that manifests or embodies your take on what is true, important, wonderful, admirable, offensive, contemptible, and so forth, cannot be voluntary.17 If an activity is voluntary, if it is done at will, it reveals your take on what is all-things-considered worth doing.18 But, for this reason, nothing voluntary could manifest your take on the distinct questions of what is true, important, wonderful, admirable, offensive, contemptible, and so forth.19 Notice, too, that these attitudes are aspects of ourselves for which we can be asked to provide a justification, rather than merely an explanation. We can be asked to defend them with reasons that show the content of our beliefs to be true, the object of our cares to be important, that which we prefer to be desirable, and so forth. They thus contrast with our headaches, the spots we see, or the colds we catch. We can explain these, but not justify them—we do not adopt or maintain a headache or a cold for reasons of our own.20 I have elsewhere suggested that we understand the agency we enjoy with respect to these attitudes as the agency we enjoy when we answer questions for ourselves—that our take on the world and our place in it can be well understood as our answers to questions about whether this is true, or whether that was offensive or malicious, or whether this is important, or all things considered worth doing, and so forth.21 Notice, we have just brought into view another kind of status, one that might be a form of responsibility thus far overlooked: the status of being answerable, in the sense of being subject to a request for one’s own reasons. Being answerable is, indeed, an important status, and one might call it a form of responsibility (though it is not, I think, a natural use of the everyday word).22 Importantly, though, answerability is more fundamental than, and should not be confused with, responsibility as mattering. A creature, even a person, can be answerable without being responsible in the responsibility-as-mattering way, as shown by Strawson’s categories of exemption. Not every instance of the relevant form of agency carries the relevant sort of significance. It seems to me likely that a wide range of creatures, actual and merely possible, are capable of answering questions, in the relevant sense—they plausibly have a take on the world of a sort for which they have their own reasons, and so they are, at least in principle, answerable. But not all of these will be capable of giving offense or of aptly prompting resentment, indignation, trust, gratitude, and so forth. I suspect that what takes us from answerability to responsibility-as-mattering is not any special enhancement of agency per se, but rather our ability to stand in complex forms of interpersonal relationships. As adults, we typically are able to recognize each other’s expectations and demands, to form expectations of our own, and to stand in relations in which we not only grant legitimacy to each other’s expectations but also know, collectively, that we each do so. In other words, I suspect that what takes us from answerability to responsibility is 284

Agency and responsibility

not further facts about our agency, per se (not further facts about our ability to do things, affect things, or control things) but rather important facts about our sociability—about our capacity to take one another into account, to matter to one another in distinctive ways. 23 These last thoughts, though, are idiosyncratic. In the earlier ones, I hope to have provided a useful taxonomy of, first, the variety of things that people might have in mind when thinking about responsibility, and, second, the corresponding ways in which an investigation of human agency may either illuminate or be illuminated by such thought.24

Related topics Agency and autonomy; Agency and free will; Agency and causation; Agency and identification; Agency and morality; Agency and the law

Notes 1 See, e.g., Morris (1968). 2 Strawson (1962). 3 Consider a lecture course going badly. The students will blame the instructor, thinking the lectures are boring, too difficult, or too flatly delivered. The instructor will blame the students, thinking they come to class inadequately prepared, they do not pay enough attention, or they are insufficiently curious. Each is thinking the other at fault for a bad outcome, and each is thinking it is the other’s job to fix the situation. Quality of will, of the sort Strawson had in mind, is not centrally at issue; yet this seems a paradigm case of blame. 4 See, e.g., Watson (1996) and Wallace (1996). 5 The ideas, and some of the text, of the last several paragraphs appear in both (Hieronymi 2019, 2021). 6 Scanlon (2015) has recently introduced the term “moral reaction responsibility.” He explicitly means the term to parallel burdens or obligations. Because I do not think reactive attitudes should be thought of in parallel with burdens and obligations, I have chosen another term. 7 The example belongs to Anscombe (1957). 8 For planning agency, see Bratman (1987). 9 See, e.g., Burge (2009). One might identify an even more basic form of agency in, e.g., chemical agents. 10 See, e.g., Pereboom (2001) and Strawson (1994). 11 See, e.g., Chisholm (1964), Kant (1997). 12 The strategy is pursued in very different ways by Wolf (1980), Wallace (1996), and Strawson (1962). I provide a sympathetic reading of Strawson’s strategy in Hieronymi (2020). 13 I am appropriating the idea of original sin. It is, properly, the Christian doctrine according to which each person is born into the world tarnished with the original sin of Adam and Eve, therefore unable to achieve righteousness on their own, and thus in need of God’s grace. I am appropriating it as a culturally available opposition to the popular idea that “ought implies can” (an idea that may also find origins in Christian doctrine). 14 On the interpretation of punishment, see Nietzsche (1998), second treatise, aphorisms 12–14, and Clark and Dudrick (2012). I am particularly indebted to conversation with Mark C. Johnson. See Johnson (2018). 15 For a justification of criminal punishment compatible with original sin, see Scanlon (1998, Chapter 6). See also Hieronymi (2021). 16 Cf. Harry Frankfurt, “if we consider a person’s will is that by which he moves himself, then what he cares about is far more germane to the character of his will than the decisions or choices he makes” Frankfurt (1988, 84). 17 One might put the point in a way that sounds paradoxical: non-voluntariness is an essential feature of the will. I think this is true. While ordinary action is voluntary, willing (intending) is not. (See, again, Frankfurt 1988 and Hieronymi 2006.)

285

Pamela Hieronymi 18 You may, of course, do something that you do not believe to be all-things-considered worth doing. But what you believe to be worth doing and what you take to be worth doing need not march in step. If you do something intentionally, then you have, therein, taken it to be, all-thingsconsidered, worth doing, even if you believe otherwise. See Hieronymi (2009). 19 One might think one could, e.g., be offended at will. Note, though, there are reasons that you might take to show being offended is worth doing—e.g., it would improve your bargaining position—that you do not take to show that anything was offensive. If being offended could be done at will, one could adopt it for such reasons. But a state of mind that could be adopted for such reasons would not be one that manifests or embodies offense. (None of this is to deny the possibility of motivated offense—anymore than denying that belief is voluntary rules out wishful thinking or self-deception. See Hieronymi 2006.) 20 G. E. M. Anscombe brought to our attention the fact that, if you are intentionally φ-ing, a certain why-question is, as she put it, “given application:” you can be rightly asked, “Why you are φ-ing?” The same is true of entire range of attitudes that manifest or embody your take on the world (Anscombe 1957). 21 See Hieronymi (2005), (2014). 22 The nearby “accountable,” in the sense of “obliged to provide an account,” may provide a bridge. 23 Hieronymi (2014). 24 Thanks to Mark C. Johnson for helpful comments and conversation. This publication was supported by a joint grant from the John Templeton Foundation and the Fetzer Institute. The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the John Templeton Foundation or the Fetzer Institute.

Recommended reading Clarke, Randolph, Michael McKenna, and Angela M. Smith, eds. 2015. The Nature of Moral Responsibility. New York: Oxford University Press. Coates, Justin, and Neal Tognazzini, eds. 2013. Blame: Its Nature and Norms. New York: Oxford University Press. ——— The above two anthologies compile recent work on “blame” and moral responsibility. Frankfurt, Harry. 1988. “The Importance of What We Care About.” Chap. 7 in The Importance of What We Care About, Cambridge: Cambridge University Press: 80–94. ——— This short paper challenges the idea that responsible human agency is ultimately grounded in our capacity for choice or decision. Watson, Gary, ed. 2003. Free Will. Second ed. Oxford: Oxford University Press. ——— A standard anthology of work on free will and moral responsibility. Hieronymi, Pamela. 2020. Freedom, Resentment, and the Metaphysics of Morals. Princeton: Princeton University Press. ——— This volume includes a short, opinionated primer on the problem of free will and moral responsibility, a reprinting of P. F. Strawson’s seminal “Freedom and Resentment,” and a close, lineby-line reading of that difficult article.

References Anscombe, G. E. M. 1957. Intention. Oxford: Blackwell Publishing Co. Bratman, Michael E. 1987. Intention, Plans, and Practical Reason. Cambridge: Cambridge University Press. Burge, Tyler. 2009. “Primitive Agency and Natural Norms.” Philosophy and Phenomenological Research 79 (2): 251–278. Chisholm, Roderick M. 1964. Human Freedom and the Self, Lindley Lecture 1964. Lawrence: Dept. of Philosophy, University of Kansas. Clark, Maudemarie, and David Dudrick. 2012. The Soul of Nietzsche’s Beyond Good and Evil. New York: Cambridge University Press. Frankfurt, Harry. 1988. “The Importance of What We Care About.” In The Importance of What We Care About, 80–94. Cambridge: Cambridge University Press. Hieronymi, Pamela. 2005. “The Wrong Kind of Reason.” The Journal of Philosophy 102 (9): 1–21.

286

Agency and responsibility Hieronymi, Pamela. 2006. “Controlling Attitudes.” Pacific Philosophical Quarterly 87 (1): 45–74. Hieronymi, Pamela. 2009. “The Will as Reason.” Philosophical Perspectives 23: 201–220. Hieronymi, Pamela. 2014. “Reflection and Responsibility.” Philosophy and Public Affairs 42 (1): 3–41. Hieronymi, Pamela. 2019. “I’ll Bet You Think this Blame Is About You.” In Oxford Studies in Agency and Responsibility: Essays on Themes from the Work of Gary Watson, edited by Justin Coates and Neal Tognazzini, 60–87. Oxford: Oxford University Press. Hieronymi, Pamela. 2020. Freedom, Resentment, and the Metaphysics of Morals. Princeton: Princeton University Press. Hieronymi, Pamela. 2021. “Fairness, Sanction, and Condemnation.” Oxford Studies in Agency and Responsibility, edited by David Shoemaker, 229–58. Oxford: Oxford University Press. Johnson, Mark C. 2018. “What Bestiality of Thought: A Nietzschean Critique of Guilt, Punishment, and the Economics of Suffering.” Ph.D., Philosophy, University of California, Riverside. Kant, Immanuel. 1997. Critique of Practical Reason. Translated by Mary Gregor. Edited by Karl A meriks, Cambridge Texts in the History of Philosophy. Cambridge: Cambridge University Press. Original edition, 1788. Morris, Herbert. 1968. “Persons and Punishment.” Monist 52: 475–501. Nietzsche, Friedrich Wilhelm. 1998. On the Genealogy of Morality, translated and edited by Maudemarie Clarke and Alan J. Swensen. Indianapolis, IN: Hackett Pub. Co. Pereboom, Derk. 2001. Living without Free Will. Cambridge: Cambridge University Press. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2015. “Forms and Conditions of Responsibility.” In The Nature of Moral Responsibility, edited by Randolph Clarke, Michael McKenna and Angela M. Smith. New York: Oxford: Oxford University Press. Strawson, Galen. 1994. “The Impossibility of Moral Responsibility.” Philosophical Studies 75: 5–24. Strawson, Peter F. 1962. “Freedom and Resentment.” Proceedings of the British Academy xlviii: 1–25. Wallace, R. Jay. 1996. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24 (2): 227–248. Wolf, Susan. 1980. “Asymmetrical Freedom.” Journal of Philosophy 77: 151–166.

287

26 AGENCY AND IDENTIFICATION Agnieszka Jaworska

The concept of identification has been developed in efforts to describe the characteristically human form agency: the agency of persons. Simpler beings, such as animals or infants, are capable of agency – their behavior is guided by purposes, they enact their desires, and these count as actions, or at least activity. But human agency is not only more sophisticated but appears to be of a different category altogether. What accounts for the distinctive nature of human agency? One line of thought is that human agency involves ownership of (some of ) one’s attitudes: these attitudes don’t merely happen to be present in the person’s psychic economy but truly represent the person. A person is not merely a collection of the attitudes that happen to coexist in one mind. Rather, her mind is organized further to make some of those attitudes fully hers. The person can thus be said to be identified with some of her attitudes and not identified with, and perhaps even alienated from, others. Identification marks a major boundary within the self: the attitudes the agent is identified with constitute a deeper self. By contrast, agency of non-persons (also called wantons) is flat, lacking this dimension of depth. The concept of identification has been employed to elucidate two key further concepts that characterize human agency: responsibility and autonomy. Regarding responsibility, the guiding thought is that we can understand why holding agents responsible makes sense despite causal determinism if we posit that one is responsible for actions that ultimately stem from motives that one is identified with. Regarding autonomy, various authors construe autonomy – that is, self-government – as government by attitudes with which one is identified. Identification is often discussed in the foregoing terms: attitudes being fully owned, genuinely authored, not alien, and so forth. However, all these descriptions can refer equally well to two significantly different, albeit related, concepts of identification: identification in the participant sense and identification as agential stance. Once we explicitly distinguish the two, sorting out various accounts of identification as elucidating one but not the other of these two concepts will help clarify what is at stake in two different debates.

288

DOI: 10.4324/9780429202131-33

Agency and identification

1 Two concepts of identification One is identified with an attitude qua participant when one fully owns the attitude in the sense of being unequivocally its subject. Here identification stands in contrast to merely discovering or observing oneself as having or manifesting an attitude. When one is identified qua participant, one cannot be bewildered or surprised by the attitude, nor can one treat it as a piece of data to contend with or ignore. If the attitude manifests in action, the agent does not simply find herself performing the action but is rather undertaking it. But identification comes in a stronger form when the person fully owns an attitude in the sense that (at least at that time) the attitude represents what the person is fundamentally about: the attitude is inseparable from and essential to the way she interprets the world and the range of possibilities for action. It is a part of her subjective perspective that she cannot set aside or suspend, a non-optional element of the stance she takes toward the world.1 The difference between these two senses of identification is displayed in the phenomenon of weakness of the will (broadly construed). A weak-willed agent who gives in to temptation (a mere desire for something enticing in the moment) identifies with the action in the participant sense: she is not bewildered by the action, she sees herself as undertaking it rather than just observing herself act. But she does not identify in the agential stance sense: this action is not essential to who she is. Insofar as she is simply tempted, this action is, in her subjective perspective, optional and, indeed, to be avoided. She can certainly assess her possibilities for action without the tempting desire. Fischer (2012) discusses the distinction between responsibility-identification and autonomy-identification. This distinction is drawn not by specifying what identification is, but rather in terms of the relevance of identification with a desire in delineating further phenomena – responsibility for the resulting action and the resulting action being autonomous, respectively. Nonetheless, Fischer’s distinction may well match the distinction sketched above. Responsibility-identification is likely a weaker sense of identification that simply casts the agent as an active doer rather than an observer of motivational forces operating within her. And, insofar as autonomy is defined as governance by one’s true self, autonomy-identification is likely a stronger kind of identification that delineates the true self as the non-optional elements of the psyche that permeate how one’s genuine possibilities for action are framed. Another indication of a match is a parallel in how weak-willed action showcases both distinctions. We saw above that the agent is participant-identified but not stance-identified with her weak-willed action. Similarly, as Fischer emphasizes, one would be held responsible for weak-willed actions even though weak-willed actions are paradigm examples of non-autonomous actions, actions that betray one’s true self. So the agent must be responsibility-identified but not autonomy-identified with the motivations behind her weak-willed actions. In proposing accounts of identification, many authors fail to distinguish between participant- and stance-identification, sometimes equivocating between them, or at best leaving it unclear which sense they address. Several authors seem to explicitly take themselves to offer an account of participant-identification – for example, they contrast identification with attitudes “merely visiting” – and yet their accounts are clearly accounts of stance-identification because it is possible to weak-willedly defy the sort of identification they delineate. Indeed, openness to this possibility is a litmus test for when an account cannot pertain to participant-identification.

289

Agnieszka Jaworska

In what follows, I begin by assessing several rival accounts of identification, interpreting them as accounts of agential stance. Once the details are out, I will defend this interpretation. Then I turn to two clear-cut accounts of participant-identification.

2 Stance-identification 2.1 Motivational hierarchy (Frankfurt) versus evaluation (Watson) Harry Frankfurt (1971) launched the contemporary debates about identification by positing that an agent identifies with a desire when she has a further (second-order) desire concerning that desire: she wants that desire to be the one that effectively moves her to action (“second-order volition”). This account introduced hierarchies of desires as an essential element of agency of persons. Unlike wantons, persons can take their own attitudes (including desires) as objects of further attitudes (such as second-order volitions). In so doing, persons privilege some of their desires as the ones that truly define them. They own those privileged desires in a deeper way. In wanting to act on those desires, they define what kind of agents they want to be, carving out their own stance within the vast sea of their psychological makeup. Gary Watson (1975) famously objected that wanting some desire to be the one that moves you to action is, on its own, just another desire, on the same footing as all other desires, including the one that it is about; the second-order desire has no special clout to define the agent or represent her deeper or more authentic stance. The distinctive feature of this desire is its second-order content (it is about another desire), but this feature alone does not prevent the agent from taking a stand against her second-order volition. To block this possibility, the Frankfurtian hierarchical approach could be redeployed, invoking a desire of yet higher order: suppose the agent wants her second-order volition to be effective (a third-order volition). But this only pushes the problem a step back. The third-order volition is also no more than a mere desire, and the possibility that the agent might take a stand against it looms again. Further iterations of the hierarchical approach would only lead to an infinite regress. Frankfurt has conceded that a hierarchy of desires cannot, by itself, account for identification, and has made attempts to supplement his hierarchical approach to avoid the regress. Frankfurt (1987) added a decision to commit to the highest-order volition to the set of conditions sufficient for identification. But this invites the objection that an agent can also, in principle, take a stand against her decisions (see Velleman 1992 and Bratman 1996). Frankfurt (1992) abandoned the idea that identification can be secured by attitudes of a particular type and instead proposed that identification consists in a specific larger configuration of a hierarchical psychic system, a configuration he calls satisfaction. Satisfaction with a desire is achieved as long as there is some hierarchical structure of desires that supports the first-order desire in question, the agent has reflectively considered whether to challenge the desire at the highest level of this hierarchy, and yet this highest-level desire stands unopposed by any attitude at the same or higher level in the hierarchy. Satisfaction is not a particular attitude but partly a matter of absence of conflict. As such, it is not something that the agent can stand against. Nonetheless, this proposal is vulnerable to Bratman’s (1996) worry that satisfaction can obtain even when the process of reflection is incomplete or unsettled, perhaps due to an emotional block or lack of energy to continue. Intuitively, in such cases, we would take the question of whether the desire represents what the agent stands for as also unsettled, contrary to the satisfaction account. Additionally, while the satisfaction proposal overcomes the regress, it persists in privileging desires according to their position in the hierarchy, without addressing Watson’s original “mere desire” objection that desires about desires lack special authority. 290

Agency and identification

Watson (1975) offers an alternative to the hierarchical approach. For him, the key to being identified with a motivation is its grounding in an evaluative judgment concerning how one should live one’s life. Identification is, fundamentally, a matter of first-order evaluation of life possibilities in terms of worth, while evaluation – if any – of one’s motivation (introducing hierarchy) is only derivative and inessential to identification. However, Watson (1987) raises reservations about his original proposal, pointing to “perverse cases” where a person can wholeheartedly embrace and stand behind motivations and the resulting choices, while acknowledging that they are not the best from the standpoint of value or worth. Consider someone fully “into” spectator sports or the latest fashions, enthusiastically devoting time and energy to them, against her judgment that these pursuits are shallow and unworthy of such devotion.

2.2 Self-governing policies (Bratman) versus caring (Seidman) In light of criticisms of both Frankfurt and Watson, Bratman (2000) proposed endorsement by self-governing policies as a vehicle of identification with a desire. Policies are general intentions meant to be discharged in recurring contexts. Self-governing policies are about which desires are candidates for realization, and, as such, sources of reasons in recurring motivationally effective deliberation. These policies specify how much weight or priority (if any) to give to the satisfaction of a given desire when a relevant occasion for deliberation arises. One identifies with a desire so long as it is taken up and supported by such self-governing policies (provided that opposing policies or policy-like attitudes don’t cancel each other out). Bratman sees the key advantage of his view over the aforementioned rivals in the explanation he can provide for why self-governing policies have the authority to represent the agent’s authentic stance. This authority is due to these policies’ role in helping to constitute the person as a temporally extended agent. Self-governing policies function to facilitate cross-temporal coordination and provide psychological continuity and connections that stitch together into one entity what might otherwise be a mere sequence of time-sliced selves. These policies are embedded in larger webs of plans, policies, and intentions that connect time-sliced selves to one another by crisscrossing reciprocal references. And since policies, like all intentions, answer to the norm of stability, their stability supports psychological continuities, not only of the policies themselves but also of the desires to which they refer, the patterns of deliberation that ensue, and the corresponding actions. A further advantage of Bratman’s view is that it removes the puzzle of Watson’s “perverse cases.” Identifying with a pursuit against one’s better judgment becomes completely straightforward: despite her negative evaluation of, let’s say, being an avid sports spectator, the agent has chosen to set this judgment aside and formed a policy to take seriously her desire to follow spectator sports and to treat this desire as at least a contender for satisfaction whenever opportunities arise. Self-governing policies, being a species of intention, are the sort of attitudes that one forms; it does not seem that one can find out or discover that one has a self-governing policy. But, on occasion, we do seem to find out or discover what we in fact stand for. Seidman (2009) sees this as a serious shortcoming of any policy-based view of identification. Can this problem be avoided without relinquishing the progress made by Bratman? Seidman’s own view, echoing Shoemaker (2003) and more explicitly Jaworska (2007), is that the emotion-based attitude of caring suffices for identification. These authors’ conceptions of caring differ in the detail, but all work from the core idea that caring involves a comprehensive emotional attunement and vulnerability to how (well or badly) the object of care 291

Agnieszka Jaworska

is doing, manifest in a wide range of specific emotions in concrete circumstances, such as joy at the object’s successes and fear of anticipated harms to the object. Caring is thus not a matter of choice or decision, and it is possible to discover that one cares about something. Correspondingly, discovering one’s own agential stance, if it is grounded in caring, is possible, unlike on Bratman’s view. According to its defenders, the caring-based view retains the key advantages of Bratman’s approach. First, Watson’s “perverse cases” are no puzzle. Identifying with a pursuit against one’s better judgment is simply a matter of caring about it and continuing to be emotionally invested in it in a manner impervious to a negative evaluation of the pursuit. Second, the authority of caring to represent the agent’s authentic stance can be explained using Bratman’s own idea that such authority stems from the attitude’s role in constituting temporally extended agency ( Jaworska 2007). Caring, somewhat like self-governing policies, is a rich source of crisscrossing referential connections across time: different emotions constitutive of caring manifested at different times refer to the same object; these emotions are tied to one another via rational norms requiring coexistence; caring inspires the formation of corresponding stable plans, intensions and policies, setting up a rich web of referential connections of the kind that Bratman describes. Also, caring tends to be stable, facilitating psychological continuity of an emotional profile and of the corresponding plans, intentions, decisions, and actions. Notably, carings’ function in support of temporally extended agency is corroborated empirically, since agents impaired in caring, such as patients with ventromedial prefrontal cortex damage, also tend to lose the coherence and coordination of their presentation across time. The caring-based view thus meets three criteria that eluded earlier accounts of stanceidentification (compare Seidman 2009): it allows for (1) discovery of one’s identification, as well as for (2) identification against one’s better judgment; and (3) it can account for the authority of the agential stance it proposes. Let’s turn to another view that also appears promising insofar as it satisfies these criteria.

2.3 Wholism versus special attitude-types (Arpaly and Schroeder) On Arpaly and Schroeder’s (1999) proposal, an attitude being part of one’s Real Self (their term for authentic stance) is a matter of its integration into one’s entire psychological makeup; Real Self is the self that is expressed in one’s more settled character rather than in the one-off or easily malleable attitudes.2 Regardless of further details, it seems clear that what is central to our psychological makeup is something that can be discovered, since we are not always attuned to how our psychology is structured. So this view allows for the possibility of discovering what one’s authentic stance is. This view also retains the two key advantages of Bratman’s approach. First, it demystifies Watson’s “perverse cases:” one can identify with a pursuit against one’s better judgment if the weight of one’s settled psychological makeup pulls against what one happens to judge to be good. This can occur in the so-called inverse akrasia where one holds on to ingrained misguided evaluative judgments against how one feels in all relevant situations. Second, the authority of one’s character to represent one’s authentic agential stance can seem obvious: we take a stance as agents in those attitudes that fit within our overall psychology with the least conflict and that most powerfully and stably control our behavior. (Additionally, Bratman’s requirements for authority are also satisfied, since such attitudes are the building blocks of psychological continuity across time.) 292

Agency and identification

So far, the caring-based view and the integration-based view have emerged as the two alternatives best able to avoid the criticisms of earlier accounts of stance-identification. Notably, unlike Bratman’s view, these views do not try to rescue the hierarchical model: neither requires reflection upon one’s own attitudes for identification. In caring about something, I am focused on the object of my care, not on my own psychology. And my attitude can be well-integrated and consistent with the bulk of my psychological makeup without me needing to reflect upon this fact. In this respect, these two views align with Watson’s account. And it is no coincidence that non-hierarchical views easily allow for one to be mistaken about and discover what one is identified with: such a discovery requires reflection upon one’s psychology, which is an additional step, not part of identification itself. How do caring-based views and integration-based views measure up against each other? While some of their disagreements may rest on conflicting intuitions, something more fundamental is at stake. Integration-based views, unlike all others discussed here, fail to serve one of the main purposes for which the concept of identification was introduced: they cannot ground a distinction between agency of persons and simpler forms of agency. Unlike second-order desires, values, self-governing policies, and carings, integration and consistency within one’s psychology do not mark a special new form of agency unavailable to beings lacking relevant abilities.3 Thus, if Arpaly and Schroeder’s view of the Real Self is deemed superior on independent grounds, this would suggest either that the search for a categorical difference between agency of persons and other, “lower” forms of agency is misguided or at least that the concept of stance-identification is ill-suited for the task.

2.4 Alternative interpretations fall short Having sketched the details of various competing accounts of identification, we are now able to appreciate why, despite some claims of their proponents to the contrary, it is best to interpret these views as describing stance-identification rather than participant-identification. For it is possible to succumb to temptation and weak-willedly betray, respectively, one’s wholehearted highest-order volitions, decisions, evaluations, self-governing policies, what one cares about, and well-integrated attitudes defining of one’s character. Such weak-willed action must still be action that the agent is identified with in the participant sense, but the proposed theories posit lack of identification in these (respective) cases. Some authors might be best interpreted not as conflating two senses of identification, but as accounting for participant-identification in terms of stance-identification. Velleman (1992) interprets Frankfurt this way: “Frankfurt… arrives at the conclusion that if a causal account of action is to include the agent’s contribution to his behavior, it must include the agent’s identifying himself with the operative motive” (133). (Perhaps this interpretation also applies to Watson 1975.) But it should now be clear that such a project cannot get off the ground: there are key instances – namely, weakness of the will – that involve participant-identification but not stance-identification, so any attempt to reduce the former to the latter is ill-fated. We now turn to accounts that allow identification with weak-willed action and, in other respects, fit participant-identification.

3 Participant-identification 3.1 The role of reason-responsiveness (Velleman versus Smith) Velleman’s (1992) account explicitly concerns participant-identification, although he reserves the term “identification” for a different concept. Velleman’s stated goal is to explain the 293

Agnieszka Jaworska

type of action where the agent is an active contributor who does the action in the full sense rather than being a mere arena for the casual operation of mental states (desires, beliefs, etc.). According to Velleman, in order to delineate the agent’s contribution to the production of the action, we must specify the mental state(s)/attitude(s) that play the functional role of the agent, so that the causal contribution of these mental state(s)/attitude(s) amounts to the agent doing the action in the full sense. Velleman (1992) proposes that the relevant mental state/ attitude is the desire to act in accordance with reasons, which (given his understanding of the constitutive aim of action and taking reasons as considerations conducive to this aim) is equivalent to the desire to do what makes sense to oneself. Subsequently, he softened his view to claim that the relevant mental state is not a desire (which would explicitly represent its aim), but rather a sub-personal mechanism or mental architecture so organized as to aim at one making sense to themselves (Velleman 2000). Velleman takes the aim of making sense to oneself as equivalent to the aim of knowing what one is doing: one knows what one is doing only if one grasps one’s action under a proper description, and any such description of an action will implicitly refer to its motives under the circumstances and so will contain an explanation of the action that would make sense of it (p. 27). The mechanism aiming at knowing what one is doing is at work, for example, when action is inhibited until one settles on what one is going to do, or in subsequent pressure to do what one has settled on doing. A mechanism organized around the implicit aim of knowing what one is doing responds to considerations conducive to this aim, and is reason-responsive in this sense. Thus, when this mechanism regulates action, blocking some motivations and reinforcing others, this amounts to basing action on reasons. (By analogy, a belief system organized around the aim of forming true representations of the world is also reason responsive when it responds to considerations conducive to its aim.) In sum, Velleman’s considered view is that an agent is a full participant undertaking an action (i.e., participant-identifies with the action) so long as a rational mechanism characteristic of persons exerts causal influence in regulating what would otherwise be a battleground of desires issuing in animal-like actions. This sub-agential mechanism is sensitive to all considerations that would make action more intelligible to its agent, settling on actions with the best rationale and knowingly channeling the motivational forces toward only those actions one has settled on (which allows the agent to know what she is doing as she is doing it). This amounts to regulating action via practical reason. Velleman refers to the type of full-blooded agency he strives to delineate as autonomous agency. This appears to undercut my earlier suggestion that autonomy-identification aligns with stance-identification, and not with participant-identification, which is Velleman’s focus. But this misalignment stems merely from an ambiguity of the term ‘autonomy,’ one that Velleman (2000: 6, n. 11) himself acknowledges. Sometimes, as in Velleman’s primary usage, autonomy is simply synonymous with what I have called participant-identification. Other authors, including Fischer (2012), invoke a more robust sense of autonomy, where autonomy, or living up to one’s true self, is a further achievement of a characteristically human agent. An agent who fails to attain autonomy does not thereby relinquish her characteristically human form of agency. A paradigmatic example of such a failure is weak-willed action. Velleman (2000: 28, n. 34) explicitly undertakes to show that his account of full-blooded agency covers even weak-willed agents, which further affirms that he does not have in mind this more robust sense of autonomy. Note that Velleman’s approach to participant-identification can be detached from his more specific views about the constitutive aim of action and the corresponding view of which considerations count as reasons for action. The core of Velleman’s approach is that an agent is participant-identified with an action when a psychological mechanism responsive to reasons 294

Agency and identification

for action exerts causal influence in regulating the motivational forces that issue in the action. And this can be combined with any number of views about what would count as reasons for action. Angela Smith (2004) criticizes rival views of participant-identification for being overly voluntarist. For her, they are problematic because they cast choice as the distinctive contribution the agent makes to the chain of events and thereby separate too starkly the agent from her attitudes: an attitude can only be properly “authored” by the agent if it is somehow chosen by her. Similar criticism seems to apply to Velleman insofar as he frames the agent’s distinctive contribution to generating an action as one over and above the causal contribution of her attitudes (desires, emotions, beliefs, etc.) and thus appears to imply that the agent cannot make this distinctive contribution to having these attitudes in the first place. Even if Velleman did not intend this – counterintuitive – implication, nonetheless, given his focus on action, he does not attempt to explain an agent’s participant-identification with an attitude, that is, for example, an agent’s distinctive participation in having a desire, feeling an emotion, and so forth. Smith, by contrast, posits a unified view of the agent’s distinctive contribution both in generating actions and in holding attitudes (desires, emotions, etc.): the agent’s contribution is the contribution of her evaluative capacities, and thus the agent is participant-identified with both actions and attitudes to the extent that both are responsive to evaluation. Note that, given the parallel between participant-identification and responsibility-identification, this facilitates the claim that we can be held responsible not only for our actions but also for our attitudes. Further, Smith’s view allows for participant-identification with weak-willed actions, and even with the motives underlying them, since both would normally engage some evaluative assessment – for example, evaluating the tempting objective as good because it would be pleasant. Velleman’s and Smith’s proposals may look similar since they seem to agree that the distinctive agential contribution is that of the capacity to respond to reasons. But the key stark difference is that, for Smith, reasons and evaluation seem to track an independent normative order, so that, for example, reasons for desire come under the guise of the good (Smith 2005: 270). For Velleman, by contrast, reasons for X (action, attitude, etc.) cannot be properly understood until we identify the constitutive aim of X. Neither actions nor attitudes aim at the good, and some attitudes, such as desires, may not have a robust constitutive aim (beyond aiming at the attainable), and therefore lack robust standards of correctness that could generate substantive reasons. It is nonetheless significant that the two leading accounts of participant-identification see identification as a matter of engagement of a reason-responsive mechanism. If this is correct, then, perhaps surprisingly, the competing accounts of participant-identification turn on the competing accounts of the nature and scope of (practical?) reasons.

4 Closing remarks The debate about the nature of identification gains in conceptual clarity by separating the discussion into two strands, stance-identification versus participant-identification. But if identification marks agency of a distinct category, the relationship between these two forms of identification must be addressed, as it would be implausible to claim that persons are capable of two different special forms of agency: participant-agency and stance-agency. Here I offer some very brief speculative remarks. It’s perhaps best to think of stance-identification as a kind of perfection or development of participant-identification. If this is the right approach, it introduces additional constraints on plausible accounts of each type of identification: an account 295

Agnieszka Jaworska

of participant-identification must show how a perfection of participant-identification can lead to a plausible account of stance-identification and an account of stance-identification  must show how it can be developed from a plausible account of participant-identification. Here are some sketches of how this might go. Velleman’s account of participant-identification as engagement of the mechanism that aims at making sense of oneself can perhaps be extended to an account of agential stance as an explicit articulation of actions and attitudes that make sense to the agent. And a caring-based account of stance-identification can perhaps be combined with a Velleman-style account of participant-identification based on the observation that cares are collections of attitudes that make sense as a package, so that in enacting a caring, an agent can make better sense of herself than in enacting a mere desire.4

Related topics Agency and autonomy; Agency and practical reasoning; Agency and responsibility; Agency and self-knowledge; Agency and the emotions; Rational agency; The aim of agency.

Notes 1 Jaworska (2017) invoked the same distinction under less descriptive names (bare-bones versus robust identification). 2 Arpaly and Schroeder focus on the concept of agential stance or Real Self used in accounts of moral responsibility, praise, and blame. Nonetheless, the concept is independent of these applications, and accounts of it can be assessed on independent grounds. 3 Relatedly, integration, unlike these other concepts, is straightforwardly a matter of degree. 4 I thank Luca Ferrero for editorial improvements.

Further reading Bratman, M. 2000. Defends a view of identification grounded in self-governing policies. Frankfurt, H. 1971. Introduces the seminal “hierarchy of desire” view of identification. Seidman, J. 2009. A comprehensive defense of a caring-based account of identification. Velleman, J.D. 1992. Defends a reductionist view according to which the role of the agent is played by the desire to act for reasons. Watson, G. 1975. Introduces influential criticisms of Frankfurt’s hierarchical approach and an alternative view tying identification to one’s evaluative judgments about how to live one’s life.

References Arpaly, N. and Schroeder, T. 1999. “Praise, Blame and the Whole Self.” Philosophical Studies 93(2): 161–188. Bratman, M. 1996. “Identification, Decision, and Treating as a Reason.” Philosophical Topics 24(2): 1–18. Bratman, M. 2000. “Reflection, Planning, and Temporally Extended Agency.” The Philosophical Review 109(1): 35–61. Fischer, J.M. 2012. “Mission Creep: Responsibility and Autonomy.” Philosophical Issues 22(1): 165–184. Frankfurt, H. 1971. “Freedom of the Will and the Concept of a Person.” Reprinted in Frankfurt 1988: 11–25.

296

Agency and identification Frankfurt, H. 1987. “Identification and Wholeheartedness.” Reprinted in Frankfurt 1988: 159–176. Frankfurt, H. 1988. The Importance of What We Care About. Cambridge: Cambridge University Press. Frankfurt, H. 1992. “The Faintest Passion.” Reprinted in Frankfurt 1999: 95–107. Frankfurt, H. 1999. Necessity, Volition, and Love. Cambridge: Cambridge University Press. Jaworska, A. 2007. “Caring and Internality.” Philosophy and Phenomenological Research 74(3): 529–568. Jaworska, A. 2017. “Identificationist Views.” In The Routledge Companion to Free Will, ed. K. Timpe, M. Griffith, and N. Levy. New York: Routledge: 15–26. Seidman, J. 2009. “Valuing and Caring.” Theoria 75: 272–303. Shoemaker, D. 2003. “Caring, Identification, and Agency.” Ethics 114: 88–118. Smith, A. 2004. “Conflicting Attitudes, Moral Agency, and Conceptions of the Self.” Philosophical Topics 32: 331–352. Smith, A. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115: 236–271. Velleman, J.D. 1992. “What Happens When Someone Acts?” Reprinted in Velleman 2000: 123–143. Velleman, J.D. 2000. The Possibility of Practical Reason. Oxford: Oxford University Press. Watson, G. 1975. “Free Agency.” Reprinted in Watson 2004: 13–32. Watson, G. 1987. “Free Action and Free Will.” Reprinted in Watson 2004: 161–196. Watson, G. 2004. Agency and Answerability: Selected Essays. Oxford: Oxford University Press.

297

27 AGENCY AND AUTONOMY Andrea C. Westlund

What is autonomy, and how is it related to agency? Theorists of autonomy seek to explain what makes an action self-governed. What is it that makes an action truly the agent’s own? Philosophers of action seek to identify conditions that distinguish what the agent does from what merely happens in or to her. These might sound like versions of the same question: isn’t whatever the agent does precisely the agent’s “own” action? If so, the term “autonomous action” would be redundant: anything that counts as an action must be initiated by the agent herself. “Actions” that are forced or accidental are certainly not autonomous, but then they are not properly speaking actions at all. So, what is the point of calling an action autonomous? According to a standard view, the key to autonomy is the distinction between psychological elements that are internal to the agent’s self and those that aren’t. Autonomous choices and actions are those motivated by or expressive of attitudes with which the agent identifies, while heteronomous actions are those from which the agent is somehow alienated. Such accounts are intuitively powerful. The term “self-governance” suggests that governance by a distinct and determinate self is the target phenomenon. I will argue, however, that such accounts are plausible as accounts of authenticity but not of autonomy. It can be misleading to think of self-governance as, literally, a matter of governance by one’s self. Autonomous agency also requires taking responsibility for oneself, as I will elucidate in this chapter. I begin by distinguishing between different cases where we are inclined to think autonomy is lacking. I then focus on cases that are most puzzling for theorists of autonomy, and give a brief explanation of the “identificationist” approach. After explaining how this approach falls short, I introduce the concept of responsibility for self and articulate an account of autonomy centered on answerability. Finally, I address the objection that some failures of autonomy may result from being too receptive to justificatory challenge. I argue that responsibility for self is a normative competence requiring not just receptivity to other points of view, but also a degree of trust or confidence in one’s own judgment. Both receptivity and confidence can be undermined by social and psychological pressures, a fact that highlights the importance of identifying social structures and educational practices that might help protect autonomous agents against such pressures.

298

DOI: 10.4324/9780429202131-34

Agency and autonomy

1 Actions that intuitively fall short of autonomy fall into several categories: First, there are those that depend on manipulation, deception, or coercion – forms of interference that do not exactly undermine the agent’s autonomy but opportunistically make use of it in the service of another’s ends. One who manipulates, deceives, or coerces another changes features of that agent’s choice situation in ways that will predictably bring about desired results, given how she could reasonably be expected to deliberate under the revised conditions. These are paradigmatically immoral ways of treating autonomous agents, but not ones that are especially useful in illuminating the nature of autonomous agency. They violate autonomy without, in any deeper sense, undermining it. In a second and third class of cases, autonomy does seem to be undermined in a deeper sense. In both classes, the agents’ actions are influenced by features of their own psychology – by mental states that are (in some trivial sense) their own, but in some other, harder to specify, sense, also not their own. Addiction, compulsion, depression, and the like belong to the second class; brainwashing, indoctrination, and adaptive preference formation belong to the third (as well as, perhaps, some cases of socialization into oppressive norms and practices). Roughly speaking, examples in the second involve issues of individual mental health or psychological functioning, while those in the third are social or interpersonal in nature and involve the direct or indirect influence of other agents on the content of one’s mental states. In both cases, the problem lies not in the immediate choice situation but in the way in which the agent develops, holds, or relates to her desires, values, or other attitudes. Unsurprisingly, theorists of autonomy have focused overwhelmingly on the latter two sorts of cases, for they pose a distinctive puzzle. Agents who become addicted, say, or indoctrinated, may initiate actions without any (immediate) interference from other agents.1 They are moved by elements of their own psychological economy. Yet, intuitively, these actions don’t seem to be fully the agents’ own. This intuition cannot be explained by showing that these actions have been causally influenced, at some point, by forces outside of the agent’s control. For all actions share this feature. The challenge for (compatibilist) theories of autonomy is to articulate a principled distinction between the problem cases and the benign. In recent years, the dominant approach to solving this problem has been a “structuralist” one, which focuses on the internal structure of the agent’s will rather than on its causal history. The key move is to articulate the relationship between elements of the psychic economy that knits them into a distinct, determinate “self ” – a set of desires, values, or action-guiding commitments that taken together amount to the agent’s deliberative perspective, with reference to which we identify particular choices or actions as truly her own. These accounts start from the observation that there are certain elements of one’s motivational economy with respect to which one might feel like a passive bystander. Typical examples of such “alienation” are addiction and compulsion. The central conceptual innovation arising from reflection on such cases is the idea of identification. One cannot be alienated from a desire or attitude with which one is identified, the thought goes; to identify with an element of one’s motivational economy just is to render it internal to the self. So what is identification? According to the best-known identificationist account, by Harry Frankfurt, one identifies with a desire when one has a higher-order desire, with which one is satisfied, in favor of being moved by that desire (see Frankfurt 1999a, 105). In earlier work, Frankfurt highlights the condition of wholeheartedness, or the absence of ambivalence about one’s volitional state, as an important form of health of the will. Later, he argues that the “continuity and

299

Andrea C. Westlund

coherence [of ] our volitional lives” (Frankfurt 1999c, 162) depends on our capacity to care. To care about something is to be committed to one’s first-order desire, where being committed entails being disposed to “support and sustain” (Frankfurt 1999c, 160) the desire even though one might choose, all things considered, not to act on it. Ultimately, Frankfurt offers a lovebased view of selfhood and autonomy: the self ’s essential nature is a subclass of cares – those one cannot help but (willingly) embrace as action-guiding – and to be autonomous is to be governed by the commands of love, or, equivalently, of the self (Frankfurt 1999b, 131 and 137). The idea that care is what renders certain motives internal has been developed, most notably, by David Shoemaker (2003), Agnieszka Jaworska (2007, 2009), and Jeffrey Seidman (2009). Michael Bratman (2003, 2007), instead, has taken elements of Frankfurt’s earlier satisfaction-based account focusing, not on care but on policies and plans in autonomous decision-making and action. On the planning account, policies governing what one treats as a justifying reason count as self-governing policies because they support and organize agency across time; on the care-based accounts, the cares or concerns not only unify agency across time, but also generate a rich, subjective point of view or “take” on the world (Seidman 2009, 296). Motives that are appropriately related either to relevant policies or concerns have a special claim to speak for the agent since they are always expressive of “who we really are.”

2 Identificationist views focus on the “authenticity” conditions of autonomy – conditions which, when met, help categorize a subclass of attitudes as truly the agent’s own, and with reference to which we can identify certain choices and actions, likewise, as hers in some non-trivial sense. It is unlikely that anyone thinks that an account of authenticity is a full account of autonomy. Most authenticity-focused theorists of autonomy at least gesture toward various “competence” conditions that must also be met – a set of cognitive, emotional, rational, and executive capabilities that must be in place to make and follow through on decisions. The views described above, for example, require that the agent be capable of varying degrees and types of reflectiveness, and have a range of volitional, emotional, and deliberative capacities. But, on these accounts, meeting such conditions is more a prerequisite for autonomy than a defining characteristic thereof.2 Since they give pride of place to authenticity conditions, I will refer to identificationist views as “authenticity-centered” views. Authenticity-centered views are attractive in part because they accommodate one very important intuition: that to act autonomously is to be, in some sense, inwardly motivated – to act, not solely in response to external pressure, constraints, or incentives, but in light of desires or interests that one embraces as one’s own. Appealing to Gerald Dworkin (1988), Amy Mullin (2014) suggests that the value of autonomy is linked to the value of meaningfulness in life: when one “embraces the goals associated with … her activities” (2014, 414), one acts in personally meaningful ways. That we value such meaningfulness arguably lends the boundary between self and not-self some of its importance. Authenticity-centered accounts speak to the interest we have in being ourselves. Unfortunately, sometimes just being ourselves constitutes a failure of self-governance. Or so I will argue. While authenticity-based accounts might succeed in picking out attitudes from which we cannot experience ourselves as alienated, they overlook important cases in which the problem seems to be not alienation but too strong an identification. Consider two ends of a spectrum. On one end, we find highly dogmatic agents, who appear to be supremely self-confident because they are impervious to doubt about their action-guiding commitments. They are either indifferent to critical perspectives on their commitments or 300

Agency and autonomy

aggressively defensive thereof. They have no disposition to engage or consider critical challenges in good faith. On the other, we find deeply deferential agents who don’t experience critical challenges as calling for a response from them at all, but instead defer all such challenges to another source of authority. Deeply deferential agents lack a disposition to engage or consider critical challenges. They may cite reasons for their choices or actions, but in doing so, they merely express their deference to a point of view distinct from their own. Their attachment to those reasons is, paradoxically, itself dogmatic. Dogmatic and deferential agents are both gripped by certain commitments in a way that makes it wrong to describe them as self-governing, even though they are not passive bystanders to the functioning of those commitments. What they lack is not a clearly defined self (for they may wholeheartedly identify with their commitments) but a sense of responsibility for self. They treat their own deliberative perspective as though it fell beyond the purview of their own governance. They render themselves invulnerable to the give and take of reasons that might otherwise lead them to “reshape” themselves by revising their practical commitments, in a deliberate, reflective manner. Although this problem cannot be articulated on a purely identificationist account, I think we can work our way from such an account to an understanding of what has gone wrong. Identificationist views appropriately treat our being reflective agents, capable of engaging in normative deliberation, as having important implications for self-governance: we cannot be self-governing unless our motivating attitudes are appropriately related to deliberation concerning what we have reason to do. Self-governance, for creatures like us, involves governance by a “justificatory point of view” (Bratman 2003, 169, n. 39). Authenticity-centered views specify what it takes to be part of the agent’s justificatory point of view, allowing us to pick out considerations that will appear, to the agent, to have the force of reasons. The problem described above, however, should make us consider the agent’s attitude toward that justificatory point of view itself. Being moved to act by an attitude that falls outside one’s justificatory point of view is certainly one way for motivation to fail to appropriately relate to practical deliberation. But being moved by a commitment internal to that point of view, yet not susceptible to normative reflection, is another: such motivation is indicative of an ongoing failure of responsiveness to considerations that bear on the policies or concerns that make up one’s justificatory point of view. Imperviousness to reflection on the commitments with which one identifies is incompatible with full self-governance. The attitudes by which such an agent is motivated are related to deliberation concerning what she has reason to do – they are internal to her justificatory perspective. But because they are not sensitive to reasoning about the merits of that perspective itself, their relation to normative deliberation is, at best, flawed. One might think a simple fix is in the offing. Perhaps all that is required, in addition to identification, is that one’s action-guiding commitments be formed or retained in a manner suitably informed by a process of critical reflection.3 On a historical interpretation of this requirement, one’s commitments must actually withstand a process of critical reflection in order to be fully one’s own. But this standard seems at once too strong and too weak. It is simply implausible that we have submitted more than a relatively small subset of our commitments (often the ones by which we have been most vexed) to sustained critical reflection at any point. Nor is it clear just how extensive such reflection would have to be, in order to declare the commitments it supports in good standing. Moreover, the historical standard also seems to imply that once one has reflected (enough) on one’s commitments, one may rest easy and simply treat the question of their authority as closed. But this is not what it is like to be an agent concerned for the justification of her choices and actions. Well-functioning 301

Andrea C. Westlund

reflective agents are prepared to consider new evidence and respond to new considerations as they arise, including considerations that bear on the concerns or policies most central to their agential point of view. The fact that one has once endorsed a commitment on the basis of some amount of reflection does not get one off the hook when reasons to rethink that commitment present themselves. One might instead adopt a hypothetical interpretation of the critical reflection condition, according to which autonomy requires that the agent’s commitments would withstand critical reflection, were she to undertake it. This standard accommodates the intuition that some unexamined commitments may be autonomously held. Oddly, however, it makes selfgovernance contingent, not on the agent’s actual relationship to her commitments but on the truth of counterfactuals concerning how she would regard them on reflection, even if she has no disposition whatsoever to engage in such reflection. Since self-governance is a type of self-relation, one would expect it to place some demands on one’s actual orientation toward one’s self-constituting commitments. What we need is an orientation toward one’s self-constituting commitments that itself carries the hallmarks of governance. A key mark of governance, which seems relevant to the shortcomings of both dogmatic and deferential agents, is that one who governs bears a certain kind of responsibility for the governed. By “responsibility,” I mean not the more commonly discussed state of being appropriately held morally responsible (treated as praiseworthy or blameworthy) for one’s actions, for which autonomy is often thought to be required. What I have in mind is, instead, closer to what T. M. Scanlon refers to as “substantive responsibility” – the sense of responsibility we invoke when we say to someone, “That’s your responsibility!” meaning that it is not up to anyone else but you to take care of the matter in question (Scanlon 1998, 248). Substantive responsibility concerns that for which one is appropriately expected to take or accept responsibility, as opposed to relinquishing or deferring responsibility to another. Colloquially speaking, it concerns the question of where the buck stops. One who is substantively responsible for something is answerable for how it fares, particularly insofar as its fate is sensitive to her decision-making and action. One in a position of governance is expected to take responsibility, in this sense, for that which falls under her jurisdiction, and accept answerability for the fruits of her governing activities. She does not actually count as governing unless she lives up to this expectation (at least to some minimum degree) and takes responsibility where appropriate. Self-governance, I propose, shares these features: one who is self-governing is expected to take responsibility for herself in the substantive sense, and is not in fact self-governing except insofar as she does so. If we identify the self with a particular justificatory perspective, then responsibility for self must be understood to extend back from the agent’s choices and actions to the action-guiding commitments that make up that perspective. The self-governing agent takes responsibility for the reasons she may cite, or that she has in view, in justifying her decisions and actions. Early identificationist views, including those of Dworkin and Frankfurt, suggest some such connection between self-governance and responsibility for self (Dworkin 1988, 32; Frankfurt 1988, 170); indeed, Frankfurt insightfully links both autonomy and responsibility for self to the human tendency to “care about who we are” (Frankfurt 1988, 163). I agree that autonomous agents manifest a form of self-care in taking responsibility for themselves.4 Cashing out this idea in terms of identification, however, does not do it justice. The attitudes involved in identifying with something are not intuitively the same as those involved in taking (substantive) responsibility for it. Simply experiencing something as non-alien, or as internal to the self, is not yet to treat it as something for which one is answerable. Perhaps 302

Agency and autonomy

we often do treat ourselves as answerable for that with which we identify, but answerability is not simply a part of identification. On my view, it is not identification (or any other condition of authenticity), but a distinct attitude of responsibility for self, that lies at the heart of self-governance: to be self-responsible is to hold oneself substantively responsible – and thus to treat oneself as ultimately answerable – for the action-guiding commitments that make one who one is. Such responsibility is not to be confused with causal responsibility for one’s character; as already noted, one’s beliefs, desires, and other attitudes are not under one’s direct, voluntary control. They are, however, characteristically judgment-sensitive, which means they are the sort of thing for which one can coherently take responsibility. Coming to recognize that one’s action-guiding commitments may be based on faulty beliefs or mistaken value judgments will, absent countervailing pressures, tend to lead one to step back, reassess, and (if warranted) revise them. Being disposed to engage in such exercises, on appropriate occasions, is at the core of responsibility for self. You do not become responsible for the fact that some desire is yours when you identify with it; rather, you take responsibility for continuing to have it when you hold yourself answerable for it in the face of external critical perspectives. Somewhat paradoxically, then, self-governance requires that the agent be open to certain kinds of influence from the outside – namely, the rational influence of external, critical perspectives on one’s deliberative point of view. One must be prepared to accord such perspectives normative weight and consider them accordingly, rather than dismissing by default any perspective that challenges attitudes with which one identifies. This openness is an aspect of the deep relationality of autonomy. The self-governing agent is not an isolated, self-enclosed center of activity, but rather one reflective agent among many, who are jointly subject to shared norms of justificatory dialogue. What constitutes appropriate responsiveness to other perspectives, under these norms, is difficult to specify in anything but a case-by-case manner, taking into consideration the specific merits of challenges raised. Clearly, however, responsibility for self requires a degree of humility about one’s own practical judgments. The self-responsible agent remains open to the possibility that she is mistaken in her commitments, and to the possibility of transformative encounters with perspectives external to her own. Some particularly trenchant critiques of answerability-centered views raise worries about this openness. Intuitively speaking, an excess of humility about one’s judgments may lead one to abandon them too readily in the face of persistent questioning or skeptical doubt. Kyla Ebels-Duggan (2015), for example, observes that students who think they are “entitled only to those views that they can fully defend against skeptics” (Ebels-Duggan 2015, 83), or that “they ought to abandon their commitments whenever they lack an answer to an objection to it” (Ebels-Duggan 2015, 83), will often abandon their views prematurely instead of holding fast to them when they should. Mirja Pérez de Calleja (2019) raises a similar worry, drawing on recent studies of youth radicalized by online exposure to ISIS propaganda. Pérez de Calleja observes that some “self-recruiting” victims of indoctrination are drawn into extreme points of view in ways that appear to prey precisely on their responsiveness to alternative points of view. Early in the process, these recruits are genuinely ambivalent and “take justificatory challenges from both jihadist and anti-jihadist perspectives as having normative standing” (Pérez de Calleja 2019, 205). But their transformation, however full of justificatory dialogue it might be, looks anything but autonomous to family, friends, therapists, or, indeed, to ex-converts themselves. I agree with Ebels-Duggan and Perez de Calleja in their assessment of these cases: both the students and the converts appear to suffer a lapse in autonomy, in spite of their apparent 303

Andrea C. Westlund

openness to external, critical perspectives. These cases show that openness to influence by external perspectives can go too far, and that, as Ebels-Duggan explicitly points out, humility about one’s action guiding commitments must be tempered by a degree of tenacity or confidence in one’s justificatory point of view (Ebels-Duggan 2015, 83).5 Fortunately, a dialogical conception of responsibility for self can accommodate these points. Being selfresponsible does require being alive to the possibility that some critic might give one reasons to reconsider even one’s most dearly held convictions. It does not, however, entail being prepared to abandon one’s views if one cannot satisfy all comers, including even (impossible to satisfy) skeptics and conspiracy theorists. Taking responsibility for oneself (as a way of caring who one is) includes caring whether or not one gives up one’s convictions for good reason: one simply doesn’t care who one is, in the relevant sense, if one is unable to muster up appropriate resistance to the prospect of abandoning one’s views to a particularly quick-footed, persistent, or overbearing critic.6 The vulnerable agents discussed by Ebels-Duggan and Perez de Calleja err not in holding themselves answerable to external, critical perspectives, but in failing to maintain proper limits to their answerability. Holding oneself answerable to external, critical perspectives involves the exercise of a relatively complex normative competence, which includes not just readiness to reflect on one’s own commitments but also readiness to assess the credibility of challenges with which one is faced. One’s readiness to entertain real doubt about one’s own commitments is, moreover, appropriately sensitive to, and regulated by, feedback regarding the credibility of particular challenges. Intense self-doubt or anxiety might interfere with one or the other, or both, components of this competence.7 Those who succumb to indoctrination tactics, for example, seem all too willing to entertain doubts about their own commitments, in ways that are not effectively tempered by signs of incredibility in their interlocutors. They seem to miss “red flags” about the perspectives into which they are drawn and find themselves in the impossible position of attempting to answer to an interlocutor who is simply not playing the same game. Such agents allow themselves to be placed entirely on the defensive, while their confidence in their own judgment is steadily eroded. This loss of confidence renders them even less able to sustain the normal reciprocity of justificatory dialogue (putting them ever more deeply on the defensive), and leaves them vulnerable to the temptation of greater certainty that their interlocutors seem to offer. Maintaining an attitude of responsibility for oneself, in such circumstances, may come close to being psychologically impossible. In Pérez de Calleja’s indoctrination cases, for example, converts’ ultimate transformation appears to be more a matter of psychological capitulation than an ordinary, reasons-responsive change of mind. I suspect it would be impossible to pinpoint a moment at which autonomy is lost, in such cases, in what appears to be an extended (even if rapidly unfolding) process of erosion. But it is important to note that, on my interpretation of this process, what is eroded is an aspect of the agent’s normative competence to judge and answer for herself, independent of her indoctrinators’ point of view. Excessive anxiety and self-doubt seem to interfere with and derail the normal functioning of that disposition – they turn the agent’s answerability against itself, so to speak, by undercutting the resilience and resistance to dogmatic challenge that are normally entailed by caring who one is. Further empirical research on the role of emotions in indoctrination and other similar experiences may either support or demand modification to my interpretation of these cases. I hope, here, at least to have outlined a promising approach. I also hope to have clarified that responsibility for self requires not just pure receptivity to external, critical perspectives (as if more receptivity were always better), but rather appropriate receptivity thereto. Appropriate 304

Agency and autonomy

receptivity does require humility about the merits of one’s own commitments, but it also requires implicit trust in one’s own judgment, in responding to challenges, and the confidence to exercise and bear the burdens of such judgment in social context. Given emerging connections between autonomy and various emotions (fear and anxiety on the one hand, and self-care and confidence on the other), further research might also shed light on the social conditions under which healthy dispositions toward answerability are developed and sustained. Feminist theorists of autonomy have already convincingly argued that many of the capacities involved in self-governance are deeply social and relational. That the disposition to answer for oneself may be subject to derailment under hostile social conditions should come as no surprise. Consideration of processes of indoctrination (and other autonomy-undermining forms of interaction) reminds us of the extent to which autonomous agency is precarious and vulnerable – which, in turns, makes it all the more important to concern ourselves with the conditions under which autonomous agency can develop and flourish. What exactly those conditions are, and how to achieve them, is, of course, a much larger question than I can hope to take on here.

Related entries Agency and free will; Agency and responsibility; Agency and morality; Agency and identification; Planning agency.

Notes 1 How the process of indoctrination, as opposed to the condition of already being indoctrinated, affects autonomy is something that I consider later in this chapter (see pp. 15–20 below). 2 Meyers (1987) and Bhandary (2020), by contrast, defend competence-focused accounts, on which autonomy is defined in terms of the possession of various capacities, and authenticity appears to play less of a central role. 3 I consider the alternatives discussed in the following two paragraphs at greater length in Westlund (2003). 4 See Westlund (2014). 5 Mullin (2007) makes a similar point about the importance to autonomy of having confidence in one’s convictions. 6 I elaborate this point in greater detail in Westlund (2014, 190–195). 7 Pérez de Calleja (2019), for example, points to the role played by indoctrinators’ cultivation of fear and anxiety in their targets.

Recommended reading David, Shoemaker (2003) “Caring, Identification, and Agency” develops a care-based version of an identificationist view. (See also Jaworska 2007 and Seidman 2009). Harry Frankfurt (1988) “Identification and Wholeheartedness” offers the basis for an influential and paradigmatically identificationist account of autonomy. Kyla Ebels-Duggan (2015) “Autonomy as an Intellectual Virtue” offers a critique of dialogical conceptions of autonomy, and suggests that autonomy requires not just humility but also an appropriate degree of confidence in the truth of one’s views. Michael, Bratman (2003) “Autonomy and Hierarchy” offers a planning theory of autonomous agency, and makes the point that self-governance involves governance by a justificatory point of view. Mirja Pérez de Calleja (2019) “Autonomy and Indoctrination: Why We Need an Emotional Condition for Autonomous Reasoning and Reflective Endorsement” also offers a critique of answerabilitybased views, and points out that emotions such as anxiety and fear appear to play a significant role in undermining the autonomy of apparently self-answerable agents.

305

Andrea C. Westlund

References Bhandary, Asha. 2020. Freedom to Care: Liberalism, Dependency, Care, and Culture. New York: Routledge. Bratman, Michael. 2003. “Autonomy and Hierarchy,” Social Philosophy and Policy 20:156–176. Bratman, Michael. 2007. “Planning Agency, Autonomous Agency,” Chap. 10 in Structures of Agency. New York: Oxford University Press, pp. 195–221. Dworkin, Gerald. 1988. The Theory and Practice of Autonomy. Cambridge: Cambridge University Press. Ebels-Duggan, Kyla. 2015. “Autonomy as Intellectual Virtue,” in The Aims of Higher Education: Problems of Morality and Justice. Eds. Harry Brighouse and Michael MacPherson. Chicago: University of Chicago Press, pp. 74–90. Frankfurt, Harry. 1988. “Identification and Wholeheartedness,” Chap. 5 in The Importance of What We Care About. Cambridge: Cambridge University Press, pp. 159–176. Frankfurt, Harry. 1999a. “The Faintest Passion,” Chap. 8 in Necessity, Volition, and Love. Cambridge: Cambridge University Press, pp. 95–107. Frankfurt, Harry. 1999b. “Autonomy, Necessity, and Love,” Chap. 11 in Necessity, Volition, and Love. Cambridge: Cambridge University Press, pp. 129–141. Frankfurt, Harry. 1999c. “On Caring,” Chap. 14 in Necessity, Volition, and Love. Cambridge: Cambridge University Press, pp. 155–180. Jaworska, Agnieszka. 2007. “Caring and Internality,” Philosophy and Phenomenological Research 74(3): 529–568. Jaworska, Agnieszka. 2009. “Caring, Minimal Autonomy, and the Limits of Liberalism,” in Naturalized Bioethics: Toward Responsible Knowing and Practice. Eds. Hilde Lindemann, Marion Verkerk, and Margaret Urban Walker. Cambridge: Cambridge University Press, pp. 80–105. Meyers, Diana Tietjens. 1987. “Autonomy Competency,” Ch. 2.4 in Self, Society, and Personal Choice. New York: Columbia University Press, pp. 76–97. Mullin, Amy. 2007. “Children, Autonomy, and Care,” Journal of Social Philosophy 38(4): 536–553. Mullin, Amy. 2014. “Children, Paternalism, and the Development of Autonomy,” Ethical Theory and Moral Practice 17: 413–426. Pérez de Calleja, Mirja. 2019. “Autonomy and Indoctrination: Why We Need an Emotional Condition for Autonomous Reasoning and Reflective Endorsement,” Social Philosophy and Policy 36(1): 192–210. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Seidman, Jeffrey. 2009. “Valuing and Caring,” Theoria 75: 272–303. Shoemaker, David. 2003. “Caring, Identification, and Agency,” Ethics 114(1): 88–118. Westlund, Andrea C. 2003. “Selflessness and responsibility for self: Is deference compatible with autonomy?” Philosophical Review 112 (4): 483–523. Westlund, Andrea C. 2014. “Autonomy and Self-Care,” in Autonomy, Oppression, and Gender. Eds. Andrea Veltman and Mark Piper. New York: Oxford University Press: 180–198.

306

28 AGENCY AND (THE LIMITS OF) VOLITIONAL CONFLICT Sarah Buss

Here is a not-so-trivial point which is trivially true: every (single) agent is unified enough to be a single agent. Here is an at-least-as obvious empirical point: we would be greatly surprised to discover a human being who had only one desire; this would be a symptom, if not a sufficient constituent, of pathology. To be human is to have many different desires, not all of which can be satisfied at any given time. It is to have many different values, only some of which can be expressed on any given occasion. Human beings also have many, mostly implicit, long-term commitments. And these commitments often support incompatible actions. I want to consider the relationship between these two uncontroversial observations. I want to focus, in particular, on (i) the constraints that being a single agent imposes on intra-agential conflict, as well as (ii) the sorts of intra-agential conflict made possible and effectively guaranteed by the heterogeneity of our commitments and goals. I will first discuss one sort of case in which an agent experiences a conflict between her ends. I will then explain why no other form of experiential volitional conflict is metaphysically possible. What am I ruling out in saying that there are significant metaphysical constraints on the possibility of experiencing volitional conflict? In addressing this question, I will shed light on the nature of agency, and especially on what is involved in setting and pursuing ends. There is, I will argue, only one way for a person to experience a synchronous conflict among her ends. Though human beings can fall short of psychic unity in many different ways, only one possible form of psychic disunity involves conflicting ends which the agent experiences as such. This is a fact about the metaphysics of agency: to understand what is involved in setting, and pursuing ends is to understand what distinguishes intra-personal divisions we necessarily experience as intra-agential conflicts from intra-personal divisions we necessarily experience as conflicts between ourselves and our own agency. * Not all psychic conflicts are internal to the will. Often, for example, we are forced to decide which of our values to privilege over others; in satisfying some desires, we are often forced to frustrate others. Though this can occasion feelings of frustration and regret, these feelings are directed at the world, and the way it constrains our possible goals. There is nothing

DOI: 10.4324/9780429202131-35307

Sarah Buss

incoherent about having a large number and range of desires; a person in this condition can have a perfectly unconflicted will. Conflicting goals or commitments are quite different. When someone fails to “live up to” one of her commitments, she fails to comply with constraints she imposes on herself. This can happen diachronically. Suppose, for example, that yesterday I formed the intention to go to work today. In so doing, I gave myself an end, thereby imposing a normative constraint on myself; I gave myself a reason to get out of bed that I did not have before.1 When I fail to “live up to” my commitment, there is an obvious sense in which I am defying and betraying myself. This inner conflict can give rise to self-criticism and guilt, as well as to regret at not having done what I had determined to do. But if I am simply defying my prior self, why should I—the person I now am—feel guilty? Haven’t I merely changed my mind about what it makes most sense for me to do? More importantly, even if I have good reason to criticize myself for changing my mind (because I lacked sufficient reason for doing so),2 my self-criticism is not itself an instance of synchronic inner conflict. If I now repudiate what I did earlier as a betrayal of what I had even earlier committed myself to doing, my stance toward this earlier action is not now in the least bit conflicted. To experience a conflict between one’s commitments, these commitments must be (and be regarded as being) contemporaneous. One must feel pulled in two different directions at once; one must experience the necessity of violating a constraint whose force one continues to endorse. Is such internal conflict really possible? Can someone really take herself to have more than one incompatible end at the same time? We discover this possibility in experiencing normative dilemmas—conflicts regarding what we are required to do here and now. A dramatic example involves someone who is required both to kill her child (because this is necessary to save the life of at least one of her children) and to refuse to kill her child (because every child has a right to life, and because she is this child’s mother). Under the stipulated circumstances neither requirement ceases to have its force as a requirement (from the agent’s own point of view), even if the agent believes (and rightly) that she has a more compelling reason to conform to one of them. Under such circumstances having a good reason not to comply with a requirement does not prevent an agent’s noncompliance from being a violation. Many philosophers insist that any such experience must be confused. There are, they assert, no situations in which someone is required to do two or more things when she can do only one of them. If, moreover, the reasons for doing A are more compelling than the reasons for doing B, then, under these circumstances, one is not required to do B. This philosophical position implies that we cannot experience the volitional conflict I have just described without misunderstanding our own ends, and even being confused about what it is to have an end. Given, however, that human beings set a wide range of goals for themselves, and given that they have a very limited ability to predict their future circumstances, there is nothing they can do to guarantee that their ends will never conflict. The possibility of such volitional conflict is the possibility of conflicting commitments; and the possibility of conflicting commitments is the possibility that the requirements we impose on ourselves will conflict. Of course, when we confront a volitional conflict, we can respond by revising our commitments. But, especially when these commitments are tied to our identity, we may find it extremely difficult, if not impossible, to do this. We may find it far more desirable to violate a requirement. We can gain insight into this possibility by considering the phenomenon Harry Frankfurt has dubbed “volitional necessity”: our commitments ground requirements that are sometimes 308

Agency and autonomy

at odds with our conclusions regarding what we have overriding reason to do, all things considered. To take one of Frankfurt’s examples: even though someone might believe he has overriding reason to solicit certain sensitive information about his fiancée from his servant, he might also have commitments (involving his status and what this means about how he should comport himself in relation to his “inferiors”) according to which this course of action is off limits. The commitments that are relevant to the phenomenon of volitional necessity are the long-term, mostly implicit, general commitments constitutive of a person’s identity: the commitment to being a respectable member of the upper class, the commitment to being a good enough mother, the commitment to treating all people with concern and respect. To violate such commitments is to betray an essential aspect of who one is or takes oneself to be.3 We discover these aspects of our identity when we discover that in paying the cost of foregoing a given action, we would be defying constraints we continue to impose on ourselves. As Frankfurt explains, it is not just that we are unwilling to oppose these constraints; our “unwillingness is itself something which [we are] unwilling to alter” (Frankfurt 1988, 87). In short, “the force of volitional necessity … coincides with—and is, indeed, partly constituted by—desires which are not merely [our] own, but with which [we] actively identify [ourselves]” (86). Experiences of volitional necessity reflect the fact that aspects of our agential/end-setting identity are independent of our identity as rational beings (independent, that is, of our disposition to avoid incoherence and to respond to reasons4). Insofar as we are rational beings, we do things for reasons and are self-conscious about this fact to the extent that we acknowledge the intelligibility of the demand to justify our actions. But there is more to our volitional nature than this. We also have substantive commitments, and certain actions are incompatible with being true to these commitments. Frankfurt’s point is that our all-things-considered normative judgments do not always have the power to weaken or alter our substantive commitments and the associated requirements. Of course, my commitment to being a good enough mother is a commitment to doing what I am (by my own lights) justified in doing. My substantive commitments would make no sense to me if I did not think I was justified in treating them as constraints on my actions; these commitments presuppose that I have (good) reasons to do what I am committed to doing. Nonetheless, being a good enough mother is distinct from, and far more substantive than, being a good enough responder to reasons. Insofar as there is something determinate to which I am committed in being committed to being a good enough mother—something that distinguishes this commitment from my commitment to being a good enough philosopher or friend or wife—this something is independent of whatever is at stake in my commitment to responding to reasons. The substantive commitment is thus a possible source of resistance to my own all-things-considered normative verdicts. Though I can often discover a way to reconcile two apparently conflicting requirements, there is no guarantee that any such reconciliation is possible. The possibility of normative dilemmas is inseparable from the fact that we can retain general ends like the end of being a good (enough) philosopher or mother or person, or of “being considerate” or “treating people kindly,” even while acting in a way that we know to be incompatible with these ends. Without being confused, someone can experience a conflict between (i) an end that is not fully responsive to reasons and (ii) the end of responding to reasons. This means that the metaphysics of human agency supports the possibility of normative dilemmas: human beings can be subject to more than one incompatible requirement at once.5 Without being confused, a person can believe that she is simultaneously required to 309

Sarah Buss

do X (because otherwise she would not be a good [enough] mother) and required not to do X (because, given what will happen to other people if she does X, she has overriding reason not to do X). * The experience of conflicting ends can take no other form: someone can experience a conflict among her commitments/ends only insofar as she experiences herself as subject to conflicting substantive requirements; and these substantive requirements can conflict with the specific ends manifest in her intentions because—and only because—they can conflict with what she takes herself to be required to do, all things considered. In defending this claim, I will first argue that one cannot regard any substantive normative requirement as overriding one’s all-things-considered normative verdicts. I will then turn to the more promising suggestion that a person can experience a volitional conflict by knowingly defying her allthings-considered normative verdicts without taking herself to be complying with a conflicting normative requirement. If such defiance were possible, it would be a case in which the agent experienced a conflict between her end of responding to reasons and the specific end of her action. I will challenge the possibility of any such conflict by calling attention to the contours of agency that rule it out. Even though we might knowingly pursue an end while believing that we have overriding reason to act otherwise, the necessary conditions of agency prevent this lack of intra-personal coherence from being a conflict within our agency. Given what it is to attribute ends to oneself, one cannot take oneself to have conflicting ends without taking oneself to be subject to conflicting requirements. Hence, given the relation between substantive and all-things-considered requirements, the necessary conditions of agency rule out the possibility of experiencing any volitional conflicts that involve deliberately violating one’s contemporaneous all-things-considered normative verdict. According to Frankfurt, an agent subject to volitional necessity “experiences himself as having no choice but to accede to the force by which he is constrained even if he thinks it might be better not to do so” (Frankfurt 1988, 86). Such a person is, Frankfurt claims, like a compulsive agent in being aware of defying his own reason, but unlike a compulsive agent in wholeheartedly endorsing this defiance. But is such a conflicted self-relation really possible? The experience of volitional necessity is the experience of certain actions as to be done. Our question thus becomes: Is it possible to regard a substantive requirement (which one acknowledges as such) as overriding what one is required to do, all things considered? In determining what one ought to do, all things considered, one considers any substantive commitments (and associated requirements) one may have. If one’s all-things-considered conclusion (“I ought to torture this human being”) is at odds with an acknowledged substantive requirement (“I am obligated to refrain from torturing human beings”), then given that one has taken the relative significance of the substantive requirement into account in reaching the all-things-considered conclusion, one has concluded that one must defy this requirement. There is no intelligible alternative: one can never intelligibly conclude that a substantive commitment imposes a more stringent requirement than one’s commitment to responding to the weight of one’s reasons. Though the latter commitment need not be a more fundamental aspect of one’s identity than any of one’s substantive commitments, what one is committed to insofar as one is committed to being guided by reasons has implications for which of one’s commitments one is capable of defying. To be guided by reason is to be responsive to the relative weight of all the “things” one considers. Since these considerations include the fact that one has various substantive commitments, which ground various 310

Agency and autonomy

substantive requirements, one’s all-things-considered normative verdict is a verdict about what these commitments and requirements imply under these circumstances. It follows that one cannot take these commitments and requirements to override this verdict. This is a point about normative conflict: though it is possible to experience normative dilemmas, it is not possible to experience an acknowledged normative requirement as overriding one’s all-things-considered normative judgment; there is no way to make sense of any such “overriding” relation. But isn’t it possible for us to deliberately defy such assessments without taking ourselves to be subject to any conflicting requirements? This suggestion appears to gain support from an all-too-familiar phenomenon. We must carefully examine this phenomenon in order to discover that the appearances are misleading. * Human beings often do things that appear to be in tension with some of their general commitments; for example, they believe that there is a very strong case against supporting factory farming, yet they order a hamburger. One plausible interpretation of this behavior is that, though these people may think it would be a good thing if factory farms were to close, and though they may even think it would be a good thing if they and others were to help bring about these desirable results, they have not really made any commitment to trying to help. Before I examine this suggestion, I want to rule out the possibility that such cases involve the very sort of normative conflict we have discussed. Imagine, then, that I take myself to be committed to doing “my part” to help stop global warming, but I have not yet done anything that I regard as responsive to this goal. Our discussion of dilemmatic conflict suggests that my inaction might reflect the fact that I have a second general end that is incompatible with the end of doing my part to slow global warming. I might, for example, also be committed to seeing the world. If this were my situation, then when I was faced with the prospect of either refraining from taking a greenhouse-gas-emitting flight or refraining from exploring various charming spots on the far side of the Atlantic, I would not be capable of living up to both of my substantive commitments. Accordingly, I might seem to experience a conflict of the dilemmatic sort. Is this a plausible interpretation? The requirements associated with the commitments that ground the experience of normative dilemmas include “perfect” requirements to do certain things, independent of any evaluation of how the reasons add up. There are certain things we must do in order to “make good on” these commitments: torturing one’s child is not a possible way of being a good mother. In contrast, the requirements that figure in the sort of case we are now considering appear to be “imperfect”; there is nothing one must do on any given occasion in order to satisfy them. If I take a flight to Norway this month, this gives me no reason to think that I am betraying my commitment to help slow global warming. After all, I am simply committed to doing something to help, at some point in time. And what if I never do anything—even though I have ample opportunity? Then, clearly, I did not have the end after all—or I gave it up at some point. In either case, on this way of looking at the two stipulated commitments, they cannot possibly generate an inner conflict at any given time. Either it seems to me that I have two perfectly compatible ends, or it seems to me that I really have only one. This reflects two facts about what it is to have an end. First, for all but our most specific intentions, there are many different ways we can achieve our ends.6 And though certain actions can betray certain commitments (as in the case of normative dilemmas), there are many 311

Sarah Buss

commitments that no single action can betray. Second, we cannot have an end without making some effort at some point to achieve it, assuming that we have been given ample time. Taken together, these points imply that no agent can experience herself as subject to synchronously conflicting ends without taking her action to violate a self-imposed substantive requirement. Hence, no agent can intelligibly attribute any means-end incoherence to herself at a time unless she regards this incoherence as the effect of conflicting requirements (to take the end and not to take the necessary means) or of a conflict between a requirement (to take the end) and a permission (not to take the necessary means). Again, if the end someone can intelligibly attribute to herself does not ground a requirement to refrain from acting as she does, then when she fails to take any steps to achieve this end, it is either (i) general enough for her to be justified in thinking that she will take the necessary steps at some point in the future or (ii) specific enough that nothing would count as achieving the same end at a different time, under different circumstances, and so she has reason to conclude that she does not really have this end after all (however much she might want to have a very similar end in the future, or to be the sort of person who would have had this very end now). Of course, if one is mistaken about which steps one must take toward one’s end, one may never even try to take these steps. But under these circumstances the “conflict” among one’s ends is nothing more than a misunderstanding about what counts as aiming to achieve them. It is not internal to one’s self-understanding; one does not experience oneself as conflicted. Since one cannot have an end if one believes one will never be able to achieve it, one can no more have two simultaneous avowedly conflicting intentions regarding what one will do at any given moment, under any given circumstances, than one can set off in two different directions at the same time. But what about intending to walk off in one direction while being convinced that it would be far better to head off in the other? Isn’t this what is going on in the case we are exploring? If one takes oneself to have the goal of helping to slow global warming, can one really think that one is justified in continually giving priority to the goal of seeing the world? If one continues to put off taking any steps to help slow global warming, wouldn’t this be a symptom of conflict between one’s appraisal of one’s options and one’s actions? Isn’t this a classic case of weakness of will? This brings us, finally, to the alleged possibility that a volitional conflict can take the form of a conflict between one’s end in acting and one’s end of responding to reasons. There are two ways of interpreting the suggestion that frequent flyers who fail to respond to their acknowledged reasons experience a conflict between the specific ends they set themselves in acting and their general end of responding to reasons. One of these interpretations identifies a genuine possibility. But it is not the possibility of a conflict within the agent. On the second interpretation, there is a (stipulated, experiential) conflict within the agent. But precisely for this reason, this interpretation does not describe a genuine possibility. I will elaborate on each of these points in reverse order. Imagine someone who intends to be guided by her all-things-considered judgment regarding what she has most reason to do. If she does not abandon this intention, then she intends to do X if she believes she has overriding reason to do X. After all, she intends to do whatever she has most reason to do; and she believes that X is the action that satisfies this description. How can we inject the experience of conflict into this situation without turning it into a dilemmatic situation? The only option is to attribute to this agent an end that, by her own lights, is incompatible with the end associated with her all-things-considered normative

312

Agency and autonomy

judgment, even though this conflicting end is not reflected in a conflicting requirement. As we have seen, a general end will not fill the bill; she will need to intend to act contrary to her present all-things-considered normative judgment, under this description. Aware of her intention to be guided by this very judgment, she will thus be aware of intending to do X and intending not to do X—at the very same time. But no one can make sense of herself under this description. So, she cannot experience herself as thus conflicted. This experience is no more intelligible than the experience of walking in two different directions at once. The source of the trouble, clearly, is the agent’s commitment to being guided by her all-things-considered normative verdict. Eliminating this part of the story would prevent her from knowingly having two inconsistent intentions. She would simply have a single intention which she would appraise from a detached, third-person perspective—a perspective she would have no intention of treating as a guide to action. By stipulation, a person in this condition has only one practical end; her end of appraising her possible actions is entirely independent of whatever ends she sets herself in determining her will. But this means that there is no conflict internal to her agency. There is an important sense in which anyone in this condition lacks unity: what she does and her appraisal of what she does are two disconnected enterprises. By stipulation, however, she does not aim at maintaining a connection between them; she is not opposed to the dissociation of her normative appraisals from her own agency. And, again, there is nothing conflicted about her agency itself. * This sort of dissociation is possible only because one can do something without giving any thought to one’s assumption that this is what the circumstances call for. If we could not operate on “automatic pilot,” there would be no exercise of agency which qualifies as such, independent of our reflections on the desirability of so acting. And so, it would not be possible for such desirability assessments to be dissociated from our actions. Not only is this sort of unreflective condition possible, it appears to be a standard feature of human life. Most of the things most of us do in a day—most of the things we do in a life— are direct responses to our apprehension of what our circumstances appear to call for. As John McDowell puts it: “Occasion by occasion one knows what to do, if one does, … by being a certain kind of person: one who sees situations in a certain distinctive way” (McDowell 1979, 73). No “weighing” precedes such responses. No considering of the “things” for and against. It is this deep taking-for-granted that sustains the habits governing our lives at every level. Do we really have good reason to go on in this way? We assume so; our doings seem to make sense. But raising the question is almost always a task for another day. And what if that day finally comes and one’s answer appears to cast doubt on whether one’s actions—and even one’s way of life—can really be justified? If one were to ask this question out of idle curiosity, or out of an exclusive interest in self-understanding, one’s aim would not be to determine whether to alter one’s behavior. And so, even if one discovered a reason to make such an alteration, one would not thereby generate any volitional conflict— even if one made no adjustment whatsoever to any of one’s routines. If it nonetheless seems that going on as before would be an instance of self-defiance, this could only be because one reads into the story the commitment to being guided by one’s all-things-considered appraisal which I have explicitly read out of it. But if we cannot add such a commitment without ruling out the very possibility of deliberately refraining from

313

Sarah Buss

conforming to this appraisal, then what are we to make of the persisting and widespread conviction that volitional conflict can take just this form? I want to bring my investigation of the contours of agency to a close by suggesting an answer that takes us back to “imperfect” requirements: though there is nothing we can do to defy such requirements, we are often justified in suspecting that we ought to be doing more to live up to them. As the case of the frequent flyer illustrates, most of our general ends are not manifest under most circumstances; declining the opportunity to pursue a given end is perfectly compatible with being genuinely committed to pursuing it. Nonetheless, we are sometimes uneasy about our failure to do anything to advance our alleged general ends. After all, there are many circumstances under which it would be pretty easy to take the necessary steps. So, why aren’t we taking advantage of these opportunities? A similar uneasiness can arise about the general commitment to being guided by one’s reasons. This commitment need not take the form of a general end; we need not consider it in determining what we have reason to do. Perhaps it is best conceived as a stable disposition to do things for reasons—a disposition which is essential to our volitional nature and which we experience as such. This general disposition is not reflected in everything we do. Nonetheless, it is such that we tend to feel uneasy if we suspect that we are being negligent when it comes to our responsiveness to reasons. To be generally disposed to do things for reasons is to be aware that we can have insufficient reason to assess our options as we do, and it is to be capable of suspecting that we may have fallen short in this respect. This is not the experience of volitional conflict. It is, rather, the experience of a weakly grounded will. No exercise of agency is “ungrounded” in the sense that it has no basis. If someone purchases meat which she knows was produced in a factory farm, and if she does this despite believing she has very compelling reason not to do so, this is because on this occasion, she does not regard this particular normative judgment as relevant to her present interest in determining what to buy for dinner. To this extent, she exemplifies the sort of self-dissociation I characterized earlier: she maintains a certain normative stance with no intention of treating her appraisal as a guide to action. This is, of course, compatible with the fact that she relies on other normative judgments to determine her will. Indeed, it is not as though she can make no sense of the fact that her hand is reaching out for the package of hamburger. (She does not experience this movement as the movement of an “anarchic” or “alien” hand.7 ) Nonetheless, she is aware that there are facts relevant to her choice which she is not bringing to bear on it. And she is aware of not having identified any good reason for failing to treat these facts as constraints on her will. This is what she experiences when it seems to her that her will is weak. She suspects that she is operating on weak normative ground—not because she takes herself to be defying her reason, but because she is aware of having foregone the opportunity to consider whether certain additional facts ought to guide her exercise of agency. If I am right about this, then a weak will and a strong will have more in common than has generally been thought. They have more in common insofar as someone appreciates how little she has done to assess the normative assumptions that guide her actions. Even though it is possible for a rational agent to engage in idle reflection on the desirability of what she is doing, human agents usually face a trade-off between (i) reflecting on whether their normative assumptions are really justified and (ii) forming an intention that is guided by the way things normatively appear at the moment. This is not merely because it is difficult to split one’s attention between each of these mental activities. It is,

314

Agency and limits of volitional conflict

more importantly, because a careful investigation of the normative appearances has the tendency to undermine one’s confidence in these appearances, and thereby to threaten one’s ability to favor one course of action over another and to form an intention on this basis. This is why dissociating one’s agency from one’s capacity to reason is not only a way of depriving one’s will of a sufficiently strong grounding, but is also the only way of avoiding volitional paralysis. * I have argued that human beings are vulnerable to a form of volitional conflict to which no human being should want to be invulnerable. This conflict reflects the fact that the deepest aspects of a human being’s identity are constituted by heterogeneous commitments that are not grounded in the capacity to reason. I have also argued that we cannot experience volitional conflict in any other form. But I have cautioned against concluding that we are invulnerable to any other form of synchronic incoherence. Our inability to experience more than one form of volitional conflict is inseparable from the fact that, in addition to having substantive commitments that can conflict with our commitment to responding to reasons, we are capable of dissociating our agency from our reason. Given this capacity, the more important challenge we face as agents is not the challenge of maintaining volitional coherence. It is, rather, the challenge of providing our will with an appropriate grounding without unduly interfering with our capacity to will.

Related entries Agency and identification; Agency and autonomy; Agency and personal identity; Intentional agency.

Notes 1 There is widespread agreement that one cannot “bootstrap” oneself into having a reason in this way. For one of many statements of the view, see Raz (2005). For my critique of these views, see Buss (2021b). 2 For some discussions of whether such irresolution is constitutive of diachronic irrationality, and if so why, see Hedden (2015), Titelbaum (2015), Woodard (forthcoming). 3 For an influential discussion of the relationship between imperatives and identity, see Christine Korsgaard’s discussion of “practical identity.” Korsgaard (1996, 101ff.) 4 There is considerable disagreement over whether coherence requirements owe whatever normative force they have to the requirement to respond to reasons, whether the order of explanation goes the other way around, or whether there are two distinct kinds of normative requirement. For some contributions to this debate, see Kolodny (2003), Kolodny (2005), Kiesewetter (2017), Loyd (2018), and Worsnip (unpublished manuscript). 5 There is a vast literature on moral dilemmas. For two examples of those who argue that there can be no conflicting moral requirements, see Conee (1982) and Ross (1939). For two examples on the other side, see Marcus (1980) and Tessman (2015). For an elaboration of the point I make in the text, see Buss (2021a). 6 As I note in Buss (2020, 207), no matter how precisely we describe a given action, there will be many different ways to satisfy this description. (Even if one’s intention is as specific as the intention to “go upstairs now,” there are any number of ways one might execute this intention.) But this does not undermine the point I am making in the text. 7 For a brief summary of this phenomenon, see Wikipedia, “Alien-Hand-Syndrome.”

315

Sarah Buss

Further reading Frankfurt, Harry. 1992. The Importance of What We Care About. Cambridge: Cambridge University Press. A collection of essays which focus on the possibility and significance of psychic conflict. Mason, H.E., ed. 1996. Moral Dilemmas and Moral Theory. New York: Oxford University Press. A wide-ranging collection of essays on the topic of moral dilemmas. Worsnip, Alex. Fitting Things Together. Unpublished manuscript. A sustained exploration of what is problematic about incoherent beliefs and intentions.

References Buss, Sarah. 2020. “Reflections on the Relation between Reason and the Will.” In Routledge Handbook of Practical Reason, in Ruth Chang Kurt Sylvan (eds.). New York: Routledge: 196–213. Buss, Sarah. 2021a. “Personal Ideal and the Ideal of Rational Agency.” Unpublished Manuscript. Buss, Sarah. 2021b. “Norms of Rationality and the Superficial Unity of the Mind.” Unpublished Manuscript. Conee, Earl. 1982. “Against Moral Dilemmas.” Philosophical Review 91(1): 87–97. Frankfurt, Harry. 1988. “The Importance of What We Care About.” Chap. 7 in The Importance of What We Care About, 80–94. Cambridge: Cambridge University Press. Hedden, Brian. 2015. “Time-Slice Rationality.” Mind 124(494): 449–491. Kiesewetter, Benjamin. 2017. The Normativity of Rationality. Oxford: Oxford University Press. Kolodny, Niko. 2003. “How Does Coherence Matter?” Proceedings of the Aristotelian Society 58(3): 229–263. Kolodny, Niko. 2005. “Why Be Rational?” Mind 114(455): 509–563. Korsgaard, Christine. 1996. Sources of Normativity. Cambridge: Cambridge University Press. Loyd, Errol. 2018. The Importance of Being Rational. Oxford: Oxford University Press. McDowell, John. 1979. “Virtue and Reason.” The Monist 62(3): 331–350. Marcus, Ruth. 1980. “Moral Dilemmas and Consistency.” Journal of Philosophy 77(3): 121–136. Raz, Joseph. 2005. “The Myth of Instrumental Rationality.” Journal of Ethics and Social Philosophy 1(1): 1–19. Ross, W. D. 1939. Foundations of Ethics. Oxford: Clarendon Press. Tessman, Lisa. 2015. Moral Failure. Oxford: Oxford University Press. Titelbaum, Michael. 2015. “Continuing On.” Canadian Journal of Philosophy 45(5–6): 670–691. Wikipedia. “Alien-Hand-Syndrome.” https://de.wikipedia.org/wiki/Alien-Hand-Syndrom Woodard, Elise. Forthcoming. “A Puzzle about Fickleness.” Nous. Worsnip, Alex. 2021. Fitting Things Together. Unpublished manuscript.

316

29 AGENCY AND THE EMOTIONS Carla Bagnoli

The relation between emotions and agency seems straightforward. Agents ordinarily refer to emotions to explain or justify what they do and turn to emotions to figure out how to act. Fear of germs explains that and why A wears a mask at the airport. B appeals to nostalgia to justify her homecoming. In deciding how to reach his destination, C consults his emotional reactions to the sublime marine landscape, which inclines him toward renting a car rather than flying there. On some occasions, emotions guide moral or immoral action. Anger for being unfairly treated prompts D to file a complaint. Professional envy spurs E to engage in unpleasant competitions with his colleagues but inspires F to strive for loftier achievements. Emotions infuse our lives as subjects and agents, and specifically, as agents living in a community. Human agents are situated in a web of social relations, on which they mutually depend. Correspondingly, emotions are (at least partially) shaped by cultures and traditions. In the philosophical reflection, emotions are generally thought to generate action. Indeed, the etymology from Latin ex movere suggests that emotions stand in a tight connection with motives, and thereby with agency. It is a matter of disagreement how to conceive of such a connection, however. Emotions appear to exert a normative pressure, which is in itself neither rational nor irrational, neither moral nor immoral. They can be co-opted in service of moral reasoning, but they also can be a stumbling block that obstructs or sidetracks rational deliberation. Some emotions are malleable and may become habits that explain routine actions or traits of character aligned with reason. Others are not: they resist correction and are insensitive to judgment. To assess the rational and moral appropriateness of specific emotions in particular patterns of practical deliberation is beyond the scope of the chapter. The task I undertake here is more general. I will illustrate some complexities in the relation between emotions and agency discussed by various theories of emotions. The main upshot is that our capacity to feel—our susceptibility and vulnerability to emotions—contributes to defining the stance of (rational and moral) agency. Unlike passive and un-reflexive experiences, emotions can be reflexively engaged, reoriented, and retracted. Because of these features, emotions can perform a crucial role in coordinating agency across individuals and over time. As such, they figure prominently in shared and moral agency, and can be recruited in service of rational deliberation. In sum, emotions are to be listed among the key agentive powers, that is, among the capacities that enable agents to (successfully) engage in their activities. DOI: 10.4324/9780429202131-36317

Carla Bagnoli

1 How to understand emotions? Different theories offer contrasting views of the nature, phenomenology, and functions of emotions (Rorty 1980; Roberts 2003; Tappolet 2010; Bagnoli 2011, 1–36; Deonna and Teroni 2012). Hence, they disagree about the nature of the relation between emotions and (rational and moral) agency. Typically, at least four kinds of elements are taken as constituents of emotions: (i) bodily (ii) the phenomenological or experiential, (iii) motivational, and (iv) intentional or cognitive. Different theories of emotions usually emphasize some of these elements at the expense of the others, and it has been forcefully argued that the category of emotions is neither homogeneous nor does it denote a natural kind (Griffith 1997). The general tendency to dismiss emotions in rational and moral agency is associated with the tacit work of the so-called simple view of emotions, which treats them as blind sensations, often stressing a primitive motivational function.1 Accounts of emotions as sensations have singled out the motivational role of emotions in generating action, thereby denying or discounting their role in other dimensions of rational deliberation and cognition. By contrast, cognitivist theories of emotions have insisted on the intentional and cognitive aspects of emotions as akin to judgments, in order to capture their significance for moral discernment, social cognition, and practical reasoning; but they have often overlooked the distinctive affectivity of emotions and their phenomenology. Recently, however, a broad convergence has formed on the view that emotions have intentional contents both capable and in need of rational assessment, and that emotions can play a central role in cognition and rational agency. Yet, such roles are importantly carried out in virtue of the distinctive experiential and phenomenal quality of emotions (Greenspan 1995). A plethora of hybrid theories attempt to vindicate the interplay of motivational and cognitive features of emotions.2 The enactivist theory seems particularly well suited to account for the complex relations between emotions and agency, insofar as it conceives of emotions as intrinsically actionoriented and unfolding in time (Hutto 2006, 2012). This theory emphasizes the active role played by the agents in their relation with the external world and recognizes a prominent role of emotions in cognitive and rational processes, which are characterized as embodied, embedded, and extended. Because of this rich characterization of the emotions, enactivism is capable of making sense of large varieties of ways in which emotions relate to agency.3

2 Emotions as springs of action Emotions are often represented as conative states which motivate and drive actions, and as such they enter in large parts of the empirical research programs. But their contribution to action can be conceptualized in different ways.

2.1 Emotions as motivational packages Wandering in the desert, G spots a javelina moving fast in his direction, screams and runs away in fear. Fear is felt, and can be registered by tracking bodily reactions, such as the elevated pulse, the sweat, and the contracted and tense muscles. The presence of fear explains G’s screaming and running. But fear explains also that such reactions are rationally justified: it makes sense to run from a danger that cannot be fought; the cry for help may help G find protection and alerts others of a proximal danger. More often, however, the role of emotions in accounting for rational action is subtler and not immediately identifiable as a drive of action. For some theories, we ought to study the 318

Agency and the emotions

way in which emotions interact with rational capacities and support the practical function of reason (Helm 2001; Jones 2004; Bagnoli 2011). For others, emotions contribute to action generation, not only by exerting a motivational pull but also by providing relevant information (Railton 2016). According to the perceptualist model, emotions are affective modes of perception, by which agents discern the salient features of the context of choice (Döring 2003, 2007; Tappolet 2011). On a similar view, emotions convey what matters by directing attention onto some specific aspects of the situation.4 Focalizing attention is a cognitive function which can be intentionally modulated. Sustained attention can alter the import and quality of emotions, and consequently their impact on action. Thus, emotional attention is part and parcel of rational deliberation, and may acquire moral significance. The moral agent is required to pay attention to the right-making features of action. Often, moral progress consists in the effort to attend at others in the proper way. Murdoch (1970) calls this form of moral achievement upon sustained effort “loving attention” (Browning 2018).

2.2 Emotions as action-oriented How does attention to details of the situation leads to performing a particular action? The question naturally arises for those cognitivist theories that conceive of emotions as akin to judgments. On the enactivist model, instead, there is no missing step: emotions generate possibilities of action, given that emotions are embodied and felt. Emotions dispose the body to undertake action as prefigured in a given way.5 Such action possibilities are envisioned once the emotion is in place (Shargel and Prinz 2018, 119). For instance, H has an ancestral fear of fire. When the hotel fire alarm goes off, it triggers a bout of fear. H guesses right that it is just a security check, and all evidence she quickly gathers shows that this is most probably the case (e.g., nobody around is worried, there is no sign of smoke, and all is quiet). Nonetheless, her fear persists and grows, and H feels an increasing urge to look for a safe escape (e.g., avoiding elevators). That is, her fear helps her identify opportunities for escape, which H would not have realized without it. Fear pushes H toward the next staircase exit, and yet she is not passive with respect to her fear. In fact, the enactivist theory emphasizes the active role played by agents in their relation to the external world and recognizes a prominent role of emotions in cognitive, rational, and agential processes. In vindicating such as role, enactivism rejects the view that emotions are representational stances or modes of discernment.6 H’s case illustrates that fear does not represent a proximal danger but it offers a selection of opportunities for action. Thus understood, emotions can be seen as contributing to practical rationality and agency. But this contribution often is the result of complex operations of the mind. The relation between emotions and opportunities for action can be more mediated than it appears at first. Morton (2013) suggests that emotions prepare for action by shaping one’s imagination. H’s fear grows into terror as she imagines the fire spreading in the building, and she recalls filmed images of people throwing themselves out of the windows. To be sure, the opportunities for action that are prepared by specific emotions enhanced by imagination are personalized in any one specific scenario. However, they also exhibit some canonical features. According to De Sousa, for instance, emotions are activated in paradigmatic scenarios, within which they predispose toward specific courses of action. In the case of fear, any agent is typically prepared to flee, fight, or freeze. The trigger mechanisms in paradigm scenarios are biological scripts, which are enriched with social and cultural norms (De Sousa 1987, 182). This theory explains how emotions guide deliberation pre-reflectively, that is, before 319

Carla Bagnoli

agents actively engage in reasoning and other reflective activities, by framing the situation in a pre-constituted manner.7 While emotions activated in paradigm scenarios are not under volitional control (both in their inception and in their unfolding), the emotional patterns and social paradigms allow for large varieties of agential responses, and thus this theory does not detract from diversity nor does it rob agents of the responsibility to choose what to do and how to do it. Granted that emotions participate in the generation of action, do they also provide reasons for action?

2.3 Emotions as normative prompts To capture the normative role of emotions in rational agency, it is useful to introduce the notion of normative pressure. Emotions exert a pressure which is normative in the basic sense that it demands attention. The pressure looms large: it is exerted on one’s thoughts, desires, conjectures, fantasies, perceptions, and on thinking in general (Morton 2013, 37–39). Suppose A fears germs because she is mildly hypochondriac. This is not a state of mind separate from her other states. Insofar as she is hypochondriac, A is sensitive to thermal excursions during the day and is especially good at detecting wind drafts and tracing their sources. She is careful with dressing appropriately in winter so as to avoid getting sick. She is prudent in a particular way: when traveling, she wears a mask. She also collect evidence about how traveling increases the risk of contracting a virus, and she is always the first in line to get vaccinated. Wearing a mask can be explained by referring to a generic fear of illness. For A, this is a policy, not a simple action unreflectively “driven” by fear in the same way in which fear explains G’s running from the javelina. A’s fear of becoming ill is more pervasive and less motivationally evident: it pervades the way A plans her day, thinks about the environment, gathers information, and forms beliefs. Additionally, the relation of fear to action is not captured by adding one more step that bridges the gap between psychological pressure and action. Rather, in this kind of cases, emotions appear to permeate the process of deliberation, and thus they do not function as (mere) motives or drives. If so, it is a mistake to localize the contribution of emotions univocally at the level of motivation. The emphasis on the indirect and pervasive role of emotions in the processes of action generation accords with the subtle phenomenology of emotions. Escaping out of fear is an action marked by a distinctive experiential quality, which is also manifested in the way the action is executed, and hence observable from an external stance. We can express this claim by saying that emotion is a qualifier of the action that it generates: escaping out of fear comports with the experience of escaping in fear, that is, escaping fearly, so to say. Sharing time with a friend out of love comports with sharing time lovingly. Love is not the impulse driving the action; it qualifies the way in which a particular action is executed. Arguably, it is because of their distinctive experiential quality that actions communicate to others the emotions that motivate them. Normally, people can tell whether you love spending time with them or not, and whether you celebrate their glorious victory or grudgingly resent their success. The emotional-communicative aspect of agency counts enormously in social interactions. For this reason, norms about how to regulate the expressive components of actions are uniformly present in all human cultures (Gibbard 1990).8 Such norms for the appropriateness of emotions represent fundamental component of social interactions and interact with norms of recognition which regulate the boundaries of the community (Strawson 1962; Darwall 1977). This reveals an important sense in which emotions and agency are socialized: social recognition is necessary for emotions to be validated. To this 320

Agency and the emotions

extent, emotions associated with normative statuses are susceptible to relations of powers and thus qualify as modes of social validation, political empowerment or discrimination. Shame is a case in point, since it constitutively implicates the gaze of others as paragons of virtue or authoritative sources of moral demands (Williams 1993). In this perspective, shaming can be a constructive tool for the development of moral agency and help faulty agents to align with public morality (Calhoun 2004); but it can also work as a means of repression, oppression, and disempowerment, when it is aimed at silencing disagreement, engendering autonomy, and suppressing diversity. Correspondingly, shame can be elicited by the understanding of one’s failures as a moral agent, but it can also be maladaptive and submissive. Finally, and relatedly, the suspension of reactive emotions (such as e.g., resentment, blame, and forgiveness) may signal that the agent is excused and cannot be held accountable, but it may also signal the withdrawal of recognition of their normative status. When unjustified, the suspension of reactive emotions is obstructive and constitutes a serious moral wrong.

3 Emotions in rational deliberation All theories acknowledge that our susceptibility to emotions often represents a liability, which in some cases might threaten to undermine our status as rational and moral agents. However, a careful investigation of how emotions represent obstacles and hindrances to rationality reveals that our emotional vulnerability is also a vital resource for rational and moral agency, rather than a recurring source of possible errors, lapses, and failures.

3.1 Emotional engagement Emotional vulnerability allows us to engage both with our own emotions and those of others. Social emotions such as shame and guilt show that this kind of engagement is not only affective and conative, but also cognitive. Emotional engagement constitutes a distinctive way of acquiring some fundamental varieties of knowledge of ourselves and others, and of standing in normative relations to ourselves and to others. In deliberative contexts, agents are able to engage with themselves and with their surroundings via emotions (Solomon 2004; Hutto 2006; Brady 2013). This claim can be spelled out in three different ways. First, emotions may play a hermeneutical role: they are interpretative schemas that help agents understand their situation and respond appropriately to it. Through a particular emotion, an agent interprets the situation in a specific way, which might demand a specific kind of action in response. For instance, D is enraged with his boss’ patronizing attitude in handling a promotion bonus. The lingering anger makes D realize that he feels wounded and wronged, and thus leads him to investigate whether this feeling is warranted. Second, emotions may be taken to play a defining role: they help characterize the agent’s action in relation to its specific circumstances. For instance, D’s anger defines the context of his action as one in which he suffered a wrong and contributes to his conceiving of appropriate ways of redressing past wrongs. Third, emotions are reflexive: thinking about their warrant helps agents understand themselves as embodied (e.g., in interoception) and embedded in specific (personal, social, and institutional) interactions. Through and because of his anger, D is brought to reflect on the nature of his emotional reaction, given the circumstances. In this case, anger favors self-understanding. Successful agents do not only have competences that allows them to correctly interpret and understand their emotions. They also have a special kind of authority with respect to claims they make about their own minds, in particular about their own intentional 321

Carla Bagnoli

attitudes. This is not an epistemic privilege of accuracy in one’s self-ascription, but a capacity to shape or determine one’s own states of mind, in ways that importantly count as an exercise of agency (Moran 2001; McGeer 2008; Lambie 2009).

3.2 The problem of emotional occlusion Arguably, the functions of emotional engagement play a large role in deliberation. Indeed, the success of deliberation depends on the appropriate emotional engagement with the circumstances of action, with one’s self and one’s surrounding (including other agents and their emotions). But this is not to say that emotions always facilitate rational or moral deliberation. Emotional occlusion is a paradigmatic case in which (rational and moral) deliberation fails because of a defective emotional engagement that obstructs the agent’s understanding of the circumstances of action. The occluding mechanism might lead the agent to ignore or discount considerations that she would consider salient, if she were not in the grip of the emotion. Emotional occlusion is both cognitively and normatively detrimental. It interferes with the correct formation of belief and with the correct formation of reasons for action. It can be morally dubious, self-serving, self-deceptive, and prone to making exceptions in one’s favor. For instance, envy might make E blind to his co-workers’ achievements so that he routinely underappreciates their skills and misses the opportunity of a valuable cooperation with them. Envy undermines E’s capacity for judgment and his capacity to form reasons for action, by exerting an undue normative pressure on him. As a result, E’s understanding of his surrounding and of his place in the relevant cooperative scheme is seriously compromised, his skills for team-working remain underdeveloped, and his leadership is hindered. This emotional occlusion results in a loss for all involved in the scenario, including the agent.

3.3 The recalcitrance of emotions Because of the relational and communicative aspects of emotions, people can be held accountable for their misplaced emotion or the dysfunctional emotional dynamics. For instance, E might be taken to target by his relevant interlocutors and demanded to revise his team assessment. Suppose he eventually comes to realize that envy distorted his past assessment of the team performances and resolves not to commit this mistake any longer in the future. Yet, as the occasion arises, envy still blinds E, making him impermeable to criticism, blame, and even self-blame. Recalcitrant envy posits a problem in that it seems to involve a rational failure or rational conflict, but it is unclear how so. Recalcitrant emotions may be said irrational insofar as they incline the subject to accept an evaluative construal of the situation that the subject has already rejected (Brady 2009). Furthermore, they seem the source of acratic action, as in E’s case. Some forms of emotional recalcitrance can be characterized in terms of failure of self-understanding and explained by a lack of appropriate well-balanced self-regarding emotions (such as, e.g., self-trust and self-confidence), rather than by the silencing effect of the recalcitrant emotion (as in E’s case). While the phenomenon of recalcitrance is largely investigated as a failure of individual rationality and self-governance, recent studies relate it also to social conditions, such as political oppression (Mackenzie 2012). In peculiar social conditions, some kinds of emotional reactions (e.g., fear of failure in competition) may be so deeply rooted that they make the agent insensitive to reasons and alter the weight and significance of evidence (e.g., the prospects of success).

322

Agency and the emotions

4 Emotions and moral agency Some emotions target agents as such, and others target actions as expressive of the agent’s character or will. In both cases, emotions are related to agency not as drives or normative prompts but insofar as they address agents as having normative standing. Among the emotions that target and respond to rational agents for their normative standing, self-respect and respect for others occupy center stage (Dillon 1994). These emotions are felt individually, but they also convey social validation about what constitute the standards of agency. Correspondingly, there are emotions sanctioning wrongful agency, such as blame and self-blame, guilt and shame. In some circumstances, such “moral emotions” count as modes of withdrawing recognition of normative status (Strawson 1962). Insofar as these emotions react to others as accountable agents, they exert a significant normative pressure and mediate complex normative relations of political authority and social power (Rorty 1998). Finally, there are emotions that target subjects who stand in special personal relations, such as love and hatred. While there are different accounts of love, it is widely recognized that love generates reasons for action and shapes agency in profound ways. For instance, Frankfurt (2004) argues that love generates contingent and yet categorical reasons for action, akin to volitional necessities. Nozick (1989) and Rorty (1987) have emphasized that love makes lovers mutually permeable, shapes their character, and thus tracks their historical trajectory. Love transforms and expands individual agency in ways that reflect the mutual permeability of lovers, mutually enhancing them but also making them mutually vulnerable, since it creates bonds and patterns of mutual dependencies (Baier 1991).

5 Emotional resonance, cooperative schemes, and shared actions Love is one of the most powerful sources of reasons for engaging in cooperative and shared actions. Several other emotions also enable agents to coordinate and thus create complex forms of agential interdependency (Gibbard 1990, 138). For instance, trust is key to generating action and executing complex shared activities (Baier 1986). When combined with hope, trust engenders distinctive dynamics between the trustor and the trustee, and explains why trusting beyond evidence can be not only robustly rational but also empowering (McGeer 2008). Furthermore, there are psychological mechanisms that allow for emotional resonance, and, by this route, fosters coordination. The simplest of these mechanisms is emotional contagion, understood as a basic affective mechanism by which an agent synchronizes her physiological and behavioral states with those of another. Emotional contagion is almost automatic, and it and may occur sub-personally; in fact, it does not presuppose even awareness of the self as distinct from that of others. Emotional contagion is rapidly effective via salience network structures and through the activation of visceromotor mirroring mechanisms (Decety and Jackson 2004). Thus, it constitutes a direct route to cooperative agency, in alternative to sharing intentions and plans. Since it is present also in infants, emotional contagion can be regarded as a rudimentary form of affect sharing, which is probably at the root of more sophisticated forms of empathy (Darwall 2006). More complex forms of shared agency based on empathic resonance require reflective perspective-taking and are highly socialized. Empathic mechanisms which allow us to take the perspective of others are epistemically relevant, shape mutual expectations, and can be the basis of strategic interactions and cooperative activities (Zahavi 2014).

323

Carla Bagnoli

A more controversial claim is that such empathic capacities are constitutive of moral agency, especially because they presuppose social grouping and bounded communities. For some, empathy, and particularly the vicarious sharing of the sympathetic feelings of an observed agent toward the target of his or her actions, are the root of moral judgment (Darwall 2006; Slote 2007). But it is debated whether empathy plays a significant role in generating moral action (Zahavi 2014; Prinz 2018). A safe conclusion is that different forms of emotional resonance produce different forms of agency.

6 Emotional transmutation and temporal agency Emotional vulnerability is a constitutive feature of embodied rational agents such as humans. Together with embodiment, temporality is an important dimension of emotions which bears on agency. Though the temporal dimension of emotions is still relatively under-discussed in philosophy, it forms a large part of emotion research in psychology (Colombetti 2014). Rather than episodic and scattered states, emotions are better viewed as temporally extended and temporally structured processes, which arise, unfold, and decline over time. Attention to the temporal structure of emotions sets aside the major source of reservations against placing emotions within the realm of agentive powers, that is, the claim that emotions change abruptly and without rationale, as moods and whims do, so that agents are helplessly in their grip. By contrast, understood as temporally structured processes, emotions can be seen as crucial agentive powers, which enable us to intelligently engage with the world, revisit our past and its claims, and prepare for future action. Understood as temporal processes, emotions are apt to different forms of regulation, which invoke practice and habituation (Snow 2006; Gross 2015; Annas et al. 2016). This claim allows for different diagnoses of the phenomenon of emotional occlusion and recalcitrance introduced in Sections 3.2 and 3.3, respectively. Emotional occlusion and recalcitrance can be conceived as failures of self-management and self-regulation of emotions, which can be overcome by retraining and habituation, rather than by self-mastering and willpower. In addition, taking into account the temporal dimension of emotions enables us to make sense of the complex relations among different emotions. Elster (1999) points out that the decline of an emotion often determines its transmutation into another. For instance, D’s anger and its agential upshot can be seen as related to a complex network of self-directed emotions, such as shame, self-respect (Borgwald 2012), and pride (Taylor 2011). One may speculate that the action of redress, while prompted by anger and shame associated with the experience of being wronged and publicly humiliated, is ultimately based on self-respect, and directs D to regain his legitimate place in social interactions and move toward moral repair (Novitz 1998; Walker 2006). Finally, the emphasis on the temporality of emotions helps explain the complex interplay of other-directed emotions such as love, hate, and trust (Gibbard 1990; Faulkner and Simpson 2017; Browning 2018). In particular, it points out that the forms of agency informed by such emotions should be assessed in a dynamic perspective.

7 Conclusion Emotions matter for agency in a number of ways. First, they intervene at various stages in the processes of action generation, and hence, they can be plausibly counted among the springs of action (Section 2). Second, emotions are reflexive: emotional engagement with one’s own emotions is a way of gaining awareness and important varieties of self-knowledge (Section 3). Most notably, emotional engagement with one’s surroundings helps agents understand 324

Agency and the emotions

what the situation requires and, consequently, situate their action in its circumstances, which is a crucial step in (rational and moral) deliberation. Arguably, reflective agency is not dispassionate but appropriately emotionally engaged. Third, there are emotions that target agents in virtue of their normative standing (Section 4). Some of these, such as agent-regret and self-respect, mark the stance of agency and can be experienced solely in the first-person perspective. Others, such as recognition-respect, blame, and resentment also depend on social practices of recognition and validation. Empathic emotions and various mechanisms of emotional resonance enable us to expand the sphere of agency by joining others as partners in schemes of instrumental and cooperative interaction and in exercises of shared agency (Section 5). Finally, emotions exhibit a temporal dimension, which puts (rational and moral) agency in perspective (Section 6). The emotional features of our embodiment, rather than being characterized as predicaments and liabilities for agency, should be recast as positive agential powers.

Related topics Agency and autonomy; Agency and responsibility; Agency and identification.

Notes 1 This is the traditional view, since Plato, Aristotle, Spinoza, Descartes, Hobbes, Hume, and Locke take emotions to involve sensations and feelings. An alternative idea of emotions as feelings is offered by James (1890). 2 Among the many cognitivist and hybrid theories of emotions, see in particular Solomon (1996) and Goldie (2000). For an action-oriented theory of emotion, Frijda (1986). Cf. Deonna and Teroni (2012, ch. 7). Griffiths also treat emotions as action-oriented representations, which evolved as strategic responses to socially significant situations; see Griffiths (2004), Griffiths and Scarantino (2009), and Gibbard (1990). For a richer taxonomy of hybrid theories, see Scarantino and De Sousa (2018). 3 There is no one shared account of the relation between emotions and action, but see Hufendiek (2016), Slaby and Wüschner (2014), Shargel and Prinz (2018), Colombetti (2014). 4 Deonna and Teroni (2012, ch. 7). For a critique of the perceptualist model, see Brady (2011). 5 Shargel and Prinz (2018), Griffiths and Scarantino (2009), Hufendiek (2016). For an alternative action-oriented theory of emotion, see Frijda (1986). 6 Varela et al. (1991); Hutto and Myin (2013); Gallagher (2017). The claim that emotions are action-orientation can be defended independent of the claim about representation. For instance, Griffiths treats emotions as action-oriented “representations,” which evolved as strategic responses to socially significant situations; see Griffiths (2004) and Griffiths and Scarantino (2009). By contrast, Gibbard (1990) treats the conative features of emotions within an expressivist, nonrepresentationalist theory. 7 De Sousa (1987, 195); see also Mackenzie (2002, 198). Solomon defends a theory of emotions as essentially active, deliberate strategies or strategic choices; see Solomon (1976). 8 Darwin (1872) emphasizes the adaptive significance of the expressive components of emotions and explains it in relation to communication and social intercourse; compare Hutto (2006).

Further reading 1 Scarantino Andrea. (2017). “Do Emotions Cause Actions, and If So How?” Emotion Review, 9(4): 326–334. The essay criticizes two arguments against the view that emotions cause action, and argues that they can help us better understand what a theory of the causal connection between emotions and actions should explain. 2 Tappolet, Christine. (2016). Emotions, Values, and Agency. Oxford: Oxford University Press. This is a defense of the perceptual theory of emotions, according to which emotions are perceptual experiences of values. It argues that emotions play a crucial role in our grasp of practical reasons, and

325

Carla Bagnoli more generally in human agency, and provides an account of the way emotions are tied to moral responsibility that departs from the response-dependent varieties of the Strawsonian approach to moral responsibility. 3 Hufendiek, Rebekka. (2016). Embodied Emotions: A Naturalist Approach to a Normative Phenomenon. London: Routledge. This study explores emotions as embodied, action-oriented representations, providing a non-cognitivist theory of emotions that accounts for their normative dimensions. 4 Frijda, Nico. (2004). “Emotions and Action.” In A. Manstead, N. Frijda, and A. Fischer (Eds.), Feelings and Emotions: The Amsterdam Symposium (Studies in Emotion and Social Interaction: 158– 173). Cambridge: Cambridge University Press. This chapter discusses the relationships between emotion and action. First, it argues that motivation for action is part and parcel of what we mean by emotions. Second, it establishes the conditions under which those motivations do and do not actually lead to action. 5 DeLancey, C. (2002). Passionate Engines: What Emotions Reveal about Mind and Artificial Intelligence. Oxford: Oxford University Press. This is a defense of the affect program theory that purports to shed light on the rationality of emotions and their role in autonomous agency, by using AI, and argues that emotions suggest a new approach to the challenges of the study of the mind and AI, whose primary topic is autonomy.

References Annas, Julia, Narvaez, Darcia and Snow, Nancy E. (eds.) (2016). Developing the Virtues: Integrating Perspectives. Oxford: Oxford University Press. Bagnoli, C. (ed.) (2011a). Morality and the Emotions. Oxford University Press. Bagnoli, C. (2011b). Emotions and the Categorical Authority of Moral Reasons. In Bagnoli (2011a): 62–81. Baier, Annette. (1986). Trust and antitrust. Ethics 96 (2): 231–260. Baier, Annette. (1991). Unsafe Loves. In The Philosophy of Erotic Love, ed. R. C. Solomon and K. M. Higgins. Lawrence: Kansas University Press, 1991: 433–450. Borgwald, Kristin. (2012). Women’s Anger, Epistemic Personhood, and Self-respect: An Application of Lehrer’s Work on Self-trust. Philosophical Studies 161 (1): 69–76. Brady, Michael S. (2009). The Irrationality of Recalcitrant Emotions. Philosophical Studies 145 (3): 413–430. Brady, Michael S. (2011). Emotions, Perceptions, and Reasons. In Bagnoli (2011a): 135–148. Brady, Michael S. (2013). Emotional Insight: The Epistemic Role of Emotional Experience. Oxford: Oxford University Press. Browning, Gary (ed.) (2018). Truth and Love. Heidelberg/New York: Springer. Colombetti, G. (2014). The Feeling Body. Affective Science Meets the Enactive Mind. Oxford: Oxford University Press. Darwall, Stephen. (1977). Two kinds of respect. Ethics 88 (1): 36–49. Darwall, Stephen. (1998). Empathy, Sympathy, Care. Philosophical Studies 89(2–3): 261–282. Darwin, Charles. (1872 [1998]). The Expression of the Emotions in Man and Animals, Introduction, Notes and Commentaries by Paul Ekman. London: Harper Collins. Decety, J. and Jackson, P.L. (2004). The Functional Architecture of Human Empathy. Behavioral and Cognitive Neuroscience Reviews 3(2): 71–100. Deonna, Julien A. and Teroni, Fabrice. (2012). The Emotions: A Philosophical Introduction. Abingdon: Routledge. De Sousa, Ronald. (1987). The Rationality of Emotion. Cambridge, MA: MIT Press. De Sousa, Ronald. (1988). Emotion and Self-deception. In Perspectives on Self-Deception, ed. Brian P. McLaughlin and Amelie O. Rorty. Berkeley, CA: University of California Press. Dillon, Robin S. (ed.) (1994). Dignity, Character and Self-Respect. New York: Routledge. Döring, S. (2003). Explaining Action by Emotion. Philosophical Quarterly 53(211): 14–30. Döring, S. (2007). Seeing what to do: Affective perception and rational motivation. Dialectica 61 (3): 363–394. Elster, Jon. (1999). Alchemies of the Mind: Rationality and the Emotions. Cambridge: Cambridge University Press. Faulkner, Paul and Simpson, Thomas W. (eds.) (2017). The Philosophy of Trust. Oxford: Oxford University Press.

326

Agency and the emotions Frankfurt, H. (2004). Reasons of Love. Princeton, NJ: Princeton University Press. Gibbard, A. (1990). Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Goldie, P. (2000). The Emotions: A Philosophical Exploration. Oxford: Clarendon Press. Greenspan, Patricia S. (1995). Practical Guilt: Moral Dilemmas, Emotions, and Social Norms. Oxford and New York: Oxford University Press. Griffiths, Paul E. (1997). What Emotions Really Are: The Problem of Psychological Categories. Chicago, IL: University of Chicago Press. Griffiths, P. E. and Scarantino, A. (2009). Emotions in the Wild: The Situated Perspective on Emotion. In The Cambridge Handbook of Situated Cognition, ed. P. Robbins and M. Aydede. Cambridge: Cambridge University Press: 437‒453. Gross, J. J. (2015). Emotion Regulation: Current Status and Future Prospects. Psychological Inquiry 26: 1–26. Helm, Bennet W. (2001). Emotional Reason: Deliberation, Motivation, and the Nature of Value. Cambridge: Cambridge University Press. Hutto, Daniel. (2006). Unprincipled Engagement. Emotional Experience, Expression and Response. In Radical Enactivism, ed. R. Menary. Amsterdam: Benjamins: 13–38. Hutto, Daniel. (2012). Truly Enactive Emotion. Emotion Review 4(2): 176–182. James, W. (1890/1983). The Principles of Psychology. Cambridge, MA: Harvard University Press. Jones, Karen. (2004). Emotional Rationality as Practical Rationality. In Setting the Moral Compass: Essays by Women Philosophers, ed. Cheshire Calhoun. Oxford: Oxford University Press, ch. 18. Lambie, J. A. (2009). Emotional Experience, Rational Action, and Self-Knowledge. Emotion Review 1: 272–280. Mackenzie, Catriona. (2002). Critical reflection, self-knowledge, and the emotions. Philosophical Explorations 5 (3): 186–206. McGeer. (2008). Trust, Hope and Empowerment. Australasian Journal of Philosophy 86(2): 237–254. Moran, Richard A. (2001). Authority and Estrangement: An Essay on Self-Knowledge. Princeton, NJ: Princeton University Press. Morton, Adam. (2013). Emotions and Imagination. Cambridge: Polity Press. Murdoch, Iris. (1970). The Sovereignty of the Good. London: Routledge and Kegan Paul. Novitz, David. (1998). Forgiveness and Self-Respect. Philosophy and Phenomenological Research 58(2): 299–315. Nozick, Robert. (1989). The Examined Life: Philosophical Meditations. New York: Simon & Schuster. Railton, Peter. (2016). “At the Core of Our Capacity to Act for a Reason: The Affective System and Evaluative Model-Based Learning and Control,” Emotion Review, (9/3): 1–8. Roberts, R. C. (2003). Emotions: An Essay in Aid of Moral Psychology. Cambridge: Cambridge University Press. Rorty, Amélie O. (ed.) (1980). Explaining Emotions. Berkeley: University of California Press. Rorty, Amélie O. (1987). The Historicity of Psychological Attitudes: Love is Not Love Which Alters Not When It Alteration Finds. Midwest Studies in Philosophy 10: 399–412. Rorty, Amélie O. (1998). The Political Sources of the Emotions: Greed and Anger. Philosophical Studies, 1998: 143–159. Scarantino, Andrea and De Sousa, Ronald (2018). Emotion. The Stanford Encyclopedia of Philosophy (Winter 2018 Edition), ed. Edward N. Zalta, URL = https://plato.stanford.edu/archives/win2018/ entries/emotion/ Shargel, Daniel and Prinz, Jesse. (2018). An Enactivist Theory of Emotional Content. In The Ontology of Emotions, in Hichem Naar and Fabrice Teroni (eds.). Cambridge: Cambridge University Press: 110–129. Slaby, Jan, and Wüschner, Philipp. (2014). Emotion and Agency. In Emotion and Value, eds. S. Roeser, and C. Todd. Oxford: Oxford University Press: 212–228. Slote, Michael. (2007). The Ethics of Care and Empathy. Abingdon: Routledge. Snow, Nancy E. (2006). Habitual Virtuous Actions and Automaticity. Ethical Theory and Moral Practice 9(5): 545–561. Solomon, Robert C. (1999). What Emotions Really Are: The Problem of Psychological Categories. Philosophical Review 108(1): 131. Solomon, Robert C. (2004). Emotions, Thoughts, and Feelings: Emotions as Engagements with the World. In Thinking about Feeling: Contemporary Philosophers on Emotion, ed. R. C. Solomon. Oxford and New York: Oxford University Press: 76‒88.

327

Carla Bagnoli Strawson, P. (1962/1974). Freedom and Resentment. Chap. 1 in Freedom and Resentment and Other Essays. London: Methuen: 1–25. Tappolet, Christine. (2010). Emotion, Motivation and Action: The Case of Fear. In Oxford Handbook of Philosophy of Emotion, ed. Goldie Peter: 325–345. Tappolet, Christine. (2011). Neo-Sentimentalism’s Prospects. In Bagnoli (2011a): 117–134. Taylor, Jacqueline. (2011). Moral Sentiments and the Sources of Moral Identity. In Bagnoli (2011a): 257–274. Varela, Francisco J., Thompson, Evan and Rosch, Eleanor. (1991). The Embodied Mind: Cognitive Science and Human Experience. Cambridge. MA: MIT Press. Walker, Margaret Urban. (2006). Moral Repair: Reconstructing Moral Relations After Wrongdoing. Cambridge: Cambridge University Press. Williams, Bernard. (1993). Shame and Necessity. Berkeley, CA: University of California Press. Zahavi, Dan. (2014). Self and Other: Exploring Subjectivity, Empathy, and Shame. Oxford and New York: Oxford University Press.

328

PART 7

Agency and time Introduction to Part 7 Luca Ferrero

Not only does our acting take place in time, but it takes time as well—often a lot of time. Even an action with the shortest possible duration, an instantaneous action, makes a difference (or prevents one) across the interval of two adjacent instants. But most of what we do takes much longer than that. One of the characteristic features of our agency is that we can sustain activities that last for a long time, possibly a lifetime. Many of our pursuits take time because we do not have the power to bring about our goals just by a god-like ‘fiat.’ We need to take smaller steps, using the limited resources available at each particular time and location, and building on these steps to reach our distant goals. In other words, we need to structure and manage the instrumental progression of many of our deeds. We can do so thanks to our characteristic planning capacities. These capacities also allow us to settle in advance the initiation of future pursuits, that is, to have future-directed intentions, plans, and policies. In virtue of their contents, our intentions and plans can, in a sense, immediately reach into the future. But our executive powers cannot. Not only are we unable to get everything we want by an instantaneous act, but we cannot immediately act-at-a-distance either. Anything that we want to accomplish in the future, we need to accomplish either by the mere causal effects of our present actions or by relying on our future contributions to an extended activity. That is, by relying on a sequence of our future momentary actions, each performed by the exercise of our local executive powers at each of those future moments. The basic challenge faced by our diachronic agency is the reconciliation of, on the one hand, the limitations of our resources and the impossibility of acting at-a-distance, with, on the other hand, our capacity to conceive of and care for extended activities and distal goals. Because of this challenge, we need to figure out plans as the appropriate sequences of momentary steps to take in light of risks and uncertainties about the future. But there is an additional challenge: we need to both plan and sustain extended pursuits in the face of possible future changes in our motivational profile and practical standpoint. These changes are very likely to occur as time goes by, especially as we get further and further away from the time when we initiated a pursuit or adopted a future-directed plan. The chapters in this section deal with the various philosophical questions and challenges that arise for diachronic agency because of our temporal predicament. Here is a sample of the philosophical questions discussed in these chapters: What are the possible structures of DOI: 10.4324/9780429202131-37329

Luca Ferrero

our extended pursuits? What is the nature and import of our planning capacity? How are the demands of rationality affected by different conceptions of our diachronic agency? How does our diachronic agency relate to our temporal identity as persons and agents, including the role played by narrative explanations and our mortality?

Diachronic agency In ‘Diachronic agency,’ Luca Ferrero discusses the distinctive features of our temporally extended agency. We do not have the power to act directly at a temporal distance, so any of our temporally extended projects must be sustained over its temporal unfolding by momentary actions. We need the capacity both to organize these momentary steps in light of a synoptic overview of the extended activity as a whole and to sustain our motivation to continue to pursue the extended activity. Hence, the distinctive mode in which we act over time is that of ‘planning agency’—which is discussed at length in the next chapter. When we engage in an extended activity, we normally see ourselves as one and the same agent across that activity, but this requires more than mere psychological continuity. This is because continuity is compatible with drastic changes in practical standpoint, which might make it very difficult to sustain extended activities. What is required is rather the stronger relation of transtemporal identification. Another important issue about our diachronic agency is that temporally extended actions and activities can have two different kinds of internal structure. The two basic dimensions are those of continuity versus unity and telic versus atelic pursuits. First, a merely continuous activity (unlike a unified one) only requires the sequential concatenation of momentary actions with no sense of an overall extended structure—compare following a breadcrumb trail or amassing a pile of stones to building a house. Second, a telic action has an end that is at some distance from the action, and one engages in the action to reach the distant end, at which point the action succeeds and terminates—for instance, house-building is supposed to continue up until the house is built. An atelic activity, by contrast, is completed as one engages in it—if one is walking, one has thereby walked. There are also ‘dialectical’ activities, whose point and nature are always under an ongoing and potentially never-ending reconsideration, as it happens for certain (and better) kinds of friendships and personal relationships. The structure of our diachronic agency also affects what we might find valuable both in individual pursuits and in our lives, as shown by the taxonomy of different kinds of ‘temporal goods’ presented in this chapter. Following a suggestion by Kieran Setiya, Ferrero discusses how attention to the temporal structure of goods and values (and of the activities that bring them about) might affect the general structure and orientation of our existence, including how we might manage the risks of some crisis of meaning, such as the so-called ‘mid-life crisis.’ Likewise, the chapter briefly touches on the question of how our mortality might make a difference to the structure of what we find valuable and how this relates to the temporal features of our agency. In the closing section, Ferrero discusses the effects of the moving temporal location of the agent (with her local executive powers) relative to her extended pursuits, including the asymmetries generated by the distinction between past and future stages of the activity and the moving horizon of what the agent can anticipate and (indirectly) control.

Planning agency A distinctive feature of human agency is its diachronic practical organization, both at the individual level (e.g., growing food in one’s garden) and at the social one (e.g., playing in a 330

Agency and time

string quartet). In ‘Planning agency,’ Michael Bratman argues that what plausibly lies behind these basic forms of practical organization is our capacity for planning agency, an agency that is structured and cross-temporally organized by future-directed practical commitments. Bratman begins by articulating a model of individual planning agency. Planning agency is a kind of goal-directed agency that involves future-directed intentions: commitments that settle what one is to do at a later time. These intentions are embedded in larger, coordinating (partial) plans with a hierarchical, means-end structure. Plan states provide a background framework for practical reasoning and exhibit characteristic stability over time. The psychological tendencies (and associated norms of rationality) toward diachronic stability and synchronic means-end coherence and consistency help support the cross-temporal organizing roles of plan states. This role is supported by the cross-temporal referential connections across plans induced by the planning structure. These referential interconnections, together with the normal stability of intentions, support a coordinated flow of activity over time, and a structure of interwoven plan states that provides the background framework for further reasoning and action. A properly functioning planning agency would normally support the distinctive cross-temporal organization in our diachronic agency, including our capacity to guide our activities in light of our grasp of their place in the larger structure of our intentions and commitments. As Bratman argues, we ultimately bother with planning because it supports the distinctive cross-temporal organization of our agency. The metaphysical and rational resources already at play in individual planning agency can be leveraged to model small-scale shared intentional activity (like string quartet playing, which sits between the case of mere walking alongside strangers and the more complex activities embedded in a web of promises). According to Bratman, we can offer a sufficient construction of these joint activities in terms of the public interdependence and interlocking of the individual plan states of the participants, with contents that are in certain ways social but which continue to function in accordance with the dynamics of individual planning agency (including responsiveness to the rational pressures involved in individual planning agency). Finally, Bratman illustrates the potential role of planning structures in our self-governance both at a time and over time. To avoid homuncular accounts of self-governance, one needs to find attitudes whose role in the agent’s psychic economy constitutes the agent’s direction of action and practical thinking in the mode of genuine agential governance. This takes two steps. First, self-governance at a time is provided by plan states which help constitute and support the agent’s persistence over time as the same agent and shape her practical reasoning appropriately. Second, self-governance over time is a matter of governing one’s life over time as the same agent with a stable standpoint. Bratman’s solution appeals to the metaphor of acting together with oneself over time: the cross-temporal interconnections within the planned temporally extended activity of an individual agent are analogous to the interpersonal interconnections characteristic of shared intentional activity. Bratman concludes by pointing out the fecundity of planning agency: this agency shapes the temporality of individual agency, the sociality of agency, and self-governance at a time and over time (whereas further research is needed to see whether it can also extend to more complex forms of organized agency).

Agency, time, and rationality Most actions and activities are executed over time by temporally extended agents. Because the execution takes place over different moments, there are multiple opportunities for 331

Luca Ferrero

reconsideration and no guarantee that the agent will be fully unified over the time span of the action. The possibility of this fragmentation suggests that individual diachronic agency might be best modeled as a sequence of momentary agents who (i) perform a series of momentary actions, (ii) may not fully identify with each other, and (iii) may not equally benefit from the extended action. As explored by Chrisoula Andreou in ‘Agency, time, and rationality,’ the possibility of this fragmented model raises interesting and complicated philosophical questions about diachronic rationality. First, there is the question of whether there are any rational constraints about the discounting of future costs and benefits. We often exhibit a ‘bias toward the near’: we give more weight to rewards closer to us in ways that go beyond compensating for uncertainty about the future. It is controversial whether this discounting is rationally permissible. But even among those who accept its permissibility, it is only exponential discounting that is usually taken as permissible. Hyperbolic discounting, by contrast, is seen as problematic because it can induce the temporal reversal of the rankings of one’s options. Andreou wonders, however, why these reversals are seen as problematic, especially for those who accept a model of instrumental rationality in terms of fragmented agents. Second, there is a question about rational constraints on the adoption of goals that can be carried out only in the distal future. The problem arises because, at the future time of action, one’s rankings might have changed. According to so-called ‘sophisticated rationality,’ one should plan on the assumption that rational agents invariably choose in accordance with their rankings at the future time of choice. But some philosophers disagree and suggest that, even when one anticipates a future change in rankings, rationality might allow one to form an intention and follow through with it. This is because at the future time one should either recognize the temporary nature of the change in rankings or resist reconsidering her options. Given that reconsidering usually carries some costs (starting with those of the new deliberation), there might be a default presumption against reconsideration. But this might conflict with the fact that an actual reconsideration might prompt a justified change of mind ( justified in light of a change in circumstances or in the agent’s preferences). However, as suggested by Richard Holton, the point of having resolutions is to avoid the reevaluation (and the expected change of mind) that would be prompted by temptation. Or one might opt for Michael Bratman’s account, in which one might be open to reevaluation at the time of temptation but still resist changing one’s mind at that time because one recognizes the temporary character of one’s current preferences (that is, these preferences are not taken as representative of one’s extended temporal standpoint). However, Andreou argues, the model of fragmentation raises questions about whether one should prioritize the values of an earlier momentary self over one’s present self. Given that instrumental rationality is supposed to take the agent’s values as given, why should a momentary self either disallow reconsideration or reconsider from an extended temporal standpoint? Finally, there is the question of whether there are any rational constraints regarding abandoning previously adopted goals. The debate is between two families of views. The first option is ‘straightforward maximization,’ which calls for maximization of what best serves one’s current rankings—including abandoning prior plans. The second is ‘constrained maximization,’ which calls for constraining one’s maximizing by those options compatible with prior plans, even when they conflict with one’s current rankings. Constrained maximization can support the rationality of acting resolutely to avoid self-defeating patterns of behavior and instabilities so as to secure diachronic self-governance. This view appears to rely on the model of the agent as committed to being a kind of unified self, a model that is, however, challenged by the fragmented model of diachronic agency. 332

Agency and time

Artificial and machine agency As discussed by Richmond H. Thomason and John Horty in ‘Artificial and machine agency,’ planning and problem-solving are two areas where much work has been done in the development of agent architecture within artificial intelligence (AI) research. An agent architecture is a functional division of a cognitive system. Different specialized subsystems are dedicated to the various functions (such as perception, memory, and symbolic reasoning) and their interactions. This chapter traces the initial history of agent architecture on means-end reasoning and the more recent developments concerned with planning. A plan is a sequence of action that leads from an initial state to the goal state. Planning amounts to searching for the sequence to pursue among the possible ones. There are different degrees of planning complexity. In producing its plans, a minimal planner only uses goals and beliefs about its current and hypothetical states (modulo its resources and competencies) but receives its beliefs and goals from a user and does not interact with its environment. There are several dimensions along which one could make this planner more complicated. Thomason and Horty focus especially on the following: adding sensors and effectors, thereby embedding the agent in the environment and requiring belief revision by monitoring the effects of its actions, accounting for limited rationality and the need for rough-and-ready plans in real-time planning, and adding mechanisms for self-acquisition of high-level goals. One important complexity is added by Intelligence Resource-bounded Machine ­architecture (IRMA), a BDI (belief–desire–intention) architecture which introduces future-directed intentions inspired by Michael Bratman’s theory of planning agency for resource-bounded agents. Planning is subjected to various constraints and trade-offs because of limited resources. Different styles of planning can be explored experimentally in the IRMA architecture (including trade-offs between over- and under-planning, the different styles of intention revision, corresponding to different personality types: stubborn, opportunistic, or even capricious). Finally, the authors discuss Thomason’s modification of the IRMA architecture, which incorporates two sorts of beliefs and desires (distinguishing prima facie from all-things-considered attitudes), uses prioritized defaults to represent informal arguments that mix beliefs and desires, and deals with conflicts between prima facie attitudes. This architecture offers a more complex picture of the interactions between mental attitudes and is a more faithful model of human practical reasoning.

Agency and personal identity The close connection between agency and personal identity is the topic of Marya Schechtman’s ‘Agency and personal identity.’ The basic intuition is that some piece of conduct is ‘agential’ if it flows from what the agent truly is, which is then taken to imply that the identity of the person is constituted by the motivations that give rise to the actions correctly attributed to her. According to this ‘agential view’ of personal identity, both agency and identity require a consistent, unconflicted, unified, and stable motivational profile. The agential view has been defended as a response to two issues. First, a movement counts as an action rather than a mere occurrence when it flows from motivations that are endorsed by the agent. Endorsement requires complete unification in motivational profile (no conflict, no inconsistency, no unsettledness), and thus unification is also required of identity. Second, there is the question of whether one should be held responsible for past actions once one has undergone a radical change in one’s motivations. According to a ‘forensic’ conception of 333

Luca Ferrero

identity, identity tracks attributions of responsibility (rather than the other way around), and one can be held responsible only for what one did under a stable motivational profile. The agential view contrasts with a different picture of personal identity, where identity is independent of agency and requires much less motivational consistency and stability. Despite the intuitive appeal of the agential view, the dynamics of human psychology suggest that personal identity is a much messier affair: first, it is common to encounter subjects who have recurring, and often inconsistent, patterns of thought, behavior, experiences, and commitments; second, people develop and change over time. These recurrences, inconsistencies, and changes might actually be taken as distinctive features of our identities. In response to this challenge, the agential view of identity might try to suggest that some fluctuations and radical changes in motivational profile might reveal the strengths of a commitment to an endorsed motivational profile, with no genuine inconstancy or ambivalence. But this reply does not take seriously that some of these vicissitudes and changes might be considered by many as inherent to their identities, rather than adventitious. The agential identity view might still be made compatible with the denial that personal identity requires a strongly unified motivational profile. To do so, the agential identity view would need to give up on at least one of two assumptions. First, it might suggest that the notion of the identity of persons is ambiguous and there is no inconsistency in personal identity, in spite of the radical changes in motivations over time. A strongly unified motivational profile is only necessary for the attribution of conduct to a person as their action, that is, for agential rather than personal identity. The second, more radical response is to accept a less-unified conception of personal identity and to complicate the attendant picture of agency. This response maintains the connection between agency, identity, and responsibility, but it no longer requires that these three notions be all-or-nothing; rather, it allows for them to be messy and complex.

Agency, narrative, and mortality The relation between agency and personal identity is further explored, through the notion of narrative, by Roman Altshuler in ‘Agency, narrative, and mortality.’ The appeal to narrative in accounts of both identity and agency arises in opposition to reductionist views, which reduce both identity and agency to their components but seem to miss the unity provided by the whole. By contrast, according to narrative views, the unity of both life and action is provided by narratives, which give meaning to the constituents of life and action both by incorporating them into the whole of the narrative and by shaping them accordingly. Although there is much dispute about the specific account of narrative, Altshuler focuses on two general features: the components of a narrative acquire their meaning from the broader context of the story, and the meaning of earlier events is partly fixed by the later ones. Altshuler argues that narrative accounts, as metaphysical accounts of personal identity, face two problems: accretion (the past might just be an accretion of events that fail to cohere with each other and thus gives rise to a future by mere causal force) and fragmentation (these accounts necessarily exclude elements of our identity that do not fit in the narrative). In response, narrative views have emphasized the role of narrative in practical rather than metaphysical identity. According to practical identity views, although all elements of one’s past play some role in shaping one’s metaphysical identity, one can shape one’s agency by choosing which elements are to be included in a coherent and unified narrative. Altshuler argues that narratives shape our agency in a variety of ways: the emotional structure of narratives can help with the structuring of our planning agency, narratives can help with our understanding of the contexts in which we exercise our agency, and they can 334

Agency and time

help us see how our individual actions might fit within wider projects or lives, and they can guide by providing self-understanding (if not even self-constitution) in terms of our histories. Narrative can also help the internal structuring of actions, to the extent that actions themselves might have a narrative form (which seems plausible given the structural parallels between how narratives are shaped by their conclusions and how actions are shaped by their aims). Altshuler argues that, once we put all the functions of narrative together, we can see how narratives do more than changing the emotional significance of past events: they can actually retrospectively determine (and possibly even reconfigure) our motivational and psychological past. This ability to genuinely shape our past is a distinctive feature of our agency. Finally, Altshuler discusses the connections between narrative and mortality. If life has the form of a narrative, only finite lives might be meaningful. Much of the debate about the meaning of mortal and immortal lives proceeds as if the contribution of narrative in shaping and guiding a life were intrinsic to agency as such, regardless of its temporal duration. But Altshuler argues that narrative might be a tool that is exclusively adapted to mortal lives. As he concludes, “perhaps we derive meaning from narratives within our mortal lives, but it is only mortal lives that make room for such narratives.”

335

30 DIACHRONIC AGENCY Luca Ferrero

1 Temporal agents With the possible exception of divine agency, all agency is temporal: it takes place in time and over time. Even the shortest imaginable action, an instantaneous or punctuate one, is a matter of producing some change (or preventing one) from one moment to the next. But our agency is diachronic in a stronger sense: we exist for, act over, and care about much longer temporal intervals than the simple momentary transitions of instantaneous or punctuate actions. However, our executive powers are temporally constrained: we can exercise them only at the present time and with effects that reach immediately only in the proximal future. We do not have the power to act directly at a temporal distance. Hence, any temporally extended action or activity that stretches beyond a single instant must be sustained over time by the continuous exercise of our momentary executive powers throughout its unfolding.1 The temporal ‘locality’ in the exercise of our executive powers is a core feature of our temporal condition. But there is more to this condition. First, our actions are not causally isolated—they have both immediate proximal causal effects and mediated distal ones. We rely on these effects to build more complex extended actions so that we do not have to restart from scratch at each and every moment. Second, resources and opportunities for action are scarce and heterogeneously distributed over space and time. These restrictions limit both what we can pursue and how we can pursue it. In particular, they set constraints on the ordering of the momentary steps needed to make progress in our pursuits. These steps must be appropriately arranged to take timely advantage of the available opportunities. Depending on our goals and circumstances, the initiation and sequencing of these steps can be more or less rigid or urgent. As a result, our diachronic agency requires a rather sophisticated ability for the time-management and coordination of momentary steps, both within and across temporally extended actions and activities. The organization of diachronic agency requires the possession and exercise of various capacities, attitudes, and concepts, including the following: (a) some sufficiently reliable expectations about the future, so that we do not act blindly; (b) some (semantic and episodic) memories, so that we do not have to relearn at every moment who we are, where we are, and what we are able to do; (c) the capacity to conceive of the internal structure of some extended temporal intervals, so that we can understand synoptically how to temporally organize the

336

DOI: 10.4324/9780429202131-38

Diachronic agency

various momentary steps; (d) some desires, cares, and concerns that are relatively stable and not exclusively bound to the present, so that we can be motivated both to initiate and sustain temporally extended activities. The scarcity and heterogeneity in the distribution of resources and opportunities affect not only the material conditions of execution but also the functioning of our psychological capacities (including our practical deliberation) and the formation, retention, and revision of the relevant cognitive and conative attitudes. The management of time and resources, therefore, affects not only our executive capacities but also the psychological work that precedes and accompanies the unfolding of our extended activities. A distinctive feature of our diachronic agency is our planning capacity, our ability to organize and sustain the temporal unfolding of our deliberations and actions. According to Michael Bratman’s influential account, the planning is made possible by the contribution of our future-directed intentions. Details of the workings of planning agency can be found in a separate entry in this volume (Bratman 2021). Here I am only going to point out the variety of ways in which intentions and, more generally, plan-states are supposed to contribute to our diachronic agency. Plan-states do all the following: (1) They partially articulate a ‘plan’—a recipe or a blueprint—for structuring the momentary steps instrumental to or constitutive of the intended action. (2) They settle practical matters by closing deliberation and framing further deliberation. (3) They contribute to the division of deliberative labor across time, saving the agent from the costs of re-deliberating about matters that have already been settled. (4) They guide the implementation of the plan by governing future conduct. (5) They sustain the action throughout its unfolding (either by their direct operation or by dint of the agent’s propensities, habits, and dispositions). (6) They help overcome or counteract temptations, irrational preference reversals, and the like. This list highlights many of the dimensions of our temporal condition. For agents like us, future-directed intentions and plan-states might play all these roles. But these roles might be partially independent of each other since they address different and separate dimensions of our diachronic agency. Another important aspect of our diachronic agency is the temporal extension of our own existence. We are not just momentary agents, who only exist for, act in, and care about the present moment. When we engage in an extended activity, we normally see ourselves as one and the same agent—at the very least, for as long as that activity is supposed to take. Temporal identity as the selfsame agent might require more than psychological (or bodily or animal) continuity. The problem is that, over a sufficiently long time, mere continuity is compatible with massive changes in the agent’s practical standpoint. Changes in practical standpoint make it difficult for agents to sustain extended activities without recourse to some strategies, possibly manipulative ones, to secure the collaboration of future reluctant selves. Hence, it seems that, in standard instances of extended agency, the agent must see herself not just as the same by mere continuity but also by dint of a stronger transtemporal identification, which preserves the stability of her practical standpoint (see Ferrero, 2022). Finally, our diachronic agency is affected by our mortality. First, we are mortal in the sense of having a necessarily finite life, which puts an upper limit to the extension of our activities and is an unavoidable source of scarce resources and opportunities for action. Second, 337

Luca Ferrero

we are mortal in the sense of being always liable to death: we constantly need to engage in active self-maintenance throughout our existence. The activity of self-maintenance might thus constitute the most basic, necessary, and pervasive exercise of our diachronic agency.

2 Kinds of temporal agents What I have described above are the major factors and dimensions of our diachronic agency. Individual agents might be differentially impacted, given their circumstances, by some of these factors. In some extreme cases, the effects might be rather dramatic and make notable qualitative differences in the functioning of diachronic agency. Consider, for instance, the effects of extreme scarcity of resources or the drastic shortening of the time horizon of one’s planning when one is constantly concerned about potential or actual imminent deadly threats (say, under the Hobbesian state of nature, which does not allow for the flourishing of diachronic agency). Even in less dramatic scenarios, the effects of scarcity and instability might be so pervasive as to affect the applicable norms of practical rationality (see Morton 2017). It might be useful briefly to compare our diachronic agency to the diachronic agency of subjects who are under very different constraints. Take nonhuman animals, for instance. Like us, they are constrained by the locality of executive powers, the scarcity of resources, and mortality. However, it seems that these agents lack the psychological and conceptual capacities to conceive synoptically of both their existence and their activities as temporally extended. If so, they can only engage in extended activities in the mode of mere continuity; that is, via sequences of momentary or short-term actions that are prompted by present-directed desires and are responsive only to local cues. These sequences add up to extended and complex activities but in a way that is unbeknown to their agents (although these activities can benefit the organisms as extended creatures and, as such, be ‘visible’ to natural selection—a fact that can help the organisms acquire the capacities, including various instincts, to engage in these extended activities although still only in the mode of mere continuity). At the other end of the spectrum, some agents might be only under some but not all of the constraints that characterize our diachronic agency. Let’s consider agents who, like us, cannot directly act at a temporal distance and thus need to spread the execution over an extended interval. Let’s also imagine that these agents have the psychological and motivational capacities to conceive and care about engaging in extended activities but do not otherwise suffer from some of the characteristic limitations of our temporal predicament. For instance, these agents might be immortal, or suffer from no scarcity of deliberative resources or information (‘frictionless deliberators,’ Bratman 1987), or be guaranteed a stable practical standpoint (for instance, by being immune from temptation). Reflecting on the diachronic agency of these subjects, even if they are only fictional, can be methodologically helpful. To begin with, some of these agents might be taken to set standards or regulative ideals for our own diachronic agency. For instance, we might articulate specific standards of rationality under conditions of full information and no deliberative costs and then qualify these standards on account of our limitations. Alternatively, we might consider the absence of temptation as the ideal case of extended agency, an ideal against which we can then come to appreciate the contribution of any remedial techniques—such as resolutions in Holton (2009)’s sense or so-called pre-commitments—that help us overcome temptations.

338

Diachronic agency

3 The structure of extended actions and activities Temporally extended actions/activities are not just extended in time. They can have different kinds of internal structure. Two basic intersecting dimensions along which we can classify activities are given by the pairs of continuity versus unity, and telicity versus atelicity.

3.1 Continuity versus unity A merely continuous activity is produced by nothing other than the concatenation of momentary or short-term actions, which are not undertaken in light of any appreciation or concern for their overall structure and long-term effects. The simplest kind of continuous activity is the outcome of the sequential iteration of momentary actions of the same type. For instance, a leisurely walk is just the concatenation of steps taken from one moment to the next without heeding the overall configuration or effects of the extended walk that happens to be so produced. A more complex continuous activity might be composed of different types of momentary actions, which might add up to some more structured extended conduct. But this activity would still count as merely continuous if the agent is just busy with stringing the momentary steps together from one moment to the next in response only to local cues and concerns. This is, for instance, the standard mode in which complex instinctual behaviors of nonhuman animals appear to unfold over time. Contrast mere continuous activities with those that are temporally unified or integrated, that is, those where the agent takes each momentary step in light of her synoptic appreciation and endorsement of their overall extended structure. For instance, in house-building, the agent needs to take each step out of her appreciation and care for the overall organization of the intended sequence of steps. This organization is ultimately imposed by the structural requirements of its intended product (the house): a mere continuous activity can result in the amassing of a heap of stones or something akin to a bird nest or a beehive (which ultimately has a repetitive structure achievable by continuity), but not in the building of a house. In paradigmatic instances of temporally unified activities, the agent continues to sustain the activity at each and every moment on account of a stable appreciation and endorsement of the original merits of the case in support of the activity rather than of the manipulative effects of her earlier steps (including what Bratman calls the snowballing effect) or of some auxiliary device of pre-commitment (such as ‘tying oneself to the mast’ or making sidebets on one’s future conduct). In other words, a paradigmatic temporally unified activity respects the agent’s diachronic autonomy (Ferrero 2010: 15): throughout the activity, the agent normally takes herself to have a stable standpoint that continues to support the unified activity as choiceworthy at each and every moment, and in the absence of any manipulation (although not independently of the effects of what the earlier steps contributed to the actual progression of the activity). Contrast this case with that of an agent who expects some resistance to or reluctance in her future contributions to the extended activity and, as a result, needs to adopt some strategy to force, cajole, or manipulate her future momentary contributions. The distinctive property of a unified activity is that the different momentary steps are organized in a structured way. They have to fit properly together. At issue is not just the continuity from moment to moment but their overall structure. The structure does not necessarily impose a strict and fixed ordering of all the steps; some might occur earlier or later, depending on the circumstances (building the foundations of a house is to be done before 339

Luca Ferrero

building the roof, but many of the other steps might be taken in many different possible orders). Eventually, however, all the steps must ‘click together’ by fitting into the overall organization demanded by the intended outcome. The intended outcome might be, like in house-building, an item or state of affairs external to the activity itself. In many of these cases, the activity and its unity are only instrumental to the production of the item and only valued as such. But there are cases where the end is the very engagement in the activity in question. Sometimes, the agent values the instrumental process even in the absence of its expected product; sometimes the agent values the process as (partly) constitutive of that very activity. An example of the latter case is engagement in a conversation. Although there might be valuable products of the conversation (say, gaining understanding about the topics of the conversation), one can value the dynamically unified process of engaging in a conversation as such (for further discussion of continuity versus unity, see Ferrero 2009b).

3.2 Telic versus atelic pursuits The distinction between continuous and unified extended actions should not be confused with the distinction between telic and atelic pursuits. The latter distinction, although it also pertains to the temporal structures of activities, concerns specifically the temporal relation between the activity and its conditions of success. A telic action only succeeds when its end or ‘telos’ is reached. Usually, reaching the telos counts both as the culmination and the termination of the action. The telos of a particular instance of house-building is a house. Once the house is built, the house-building has reached its success and terminus. It would be problematic if one were to continue with house-building despite having reached the required structural closure. A telic action might be terminated earlier than its successful completion: the fact that an agent is, at some point in time, engaged in that action is no guarantee that the action is ever going to be completed. The fact that an agent was, is, or will be building a house leaves open whether the agent will have ever built a house. (In linguistic terms, the progressive aspect description does not entail the corresponding description in the perfective aspect.) An atelic activity, by contrast, is both completed and successful as one engages in it. It has no completion or culmination external to or separate from its very exercise. The agent who is walking has thereby walked. (In linguistic terms, the description in the progressive aspect entails the corresponding description in the perfective aspect.) There is no terminus that one must reach at a later time in order to complete the atelic activity. This does not entail that trying to engage in an atelic activity is necessarily successful. In trying to walk, one might trip over oneself and fail. Thus, in order to walk one needs to take the appropriate steps. Atelic activities, as much as telic actions, have standards that one needs to meet to engage in them. But if one meets the conditions for engagement in an atelic activity, one ipso facto meets the success conditions of that activity. In other words, when one is actually engaged in a telic action, one is not thereby guaranteed either contemporaneous or future success, whereas actual and correct engagement in an atelic activity guarantees contemporaneous success at it. This feature gives atelic activities a different temporal orientation from telic actions. Telic actions are, by their very nature, directed toward the distal future. Hence, they can suffer from premature interruption. Atelic activities, instead, are oriented toward the ongoing present, that is, they are directed at their own contemporaneous performance. As such, they might be stopped at any time and still be successful. However, the orientation toward their contemporaneous success does not prevent the atelic activities from continuously propelling 340

Diachronic agency

themselves into the proximal future, in a way that might make them extend indefinitely. That is why atelic ends are sometimes called ‘infinite’ by comparison to ‘finite’ telic ends (Setiya 2014, Rödl 2010, 2011).2 As a structural matter, the interruption of a generic atelic activity is never premature and, as such, never a failure, even if the activity has a default propensity toward an indefinite (and potentially infinite) continuation.3 An interesting kind of atelic activities is that of what Brewer (2009) calls ‘dialectical activities’: the point and nature of a dialectical activity are always under an ongoing and potentially never-ending reconsideration. Take friendship. In its dialectical form, the pursuit of friendship is not unified by an end that is given in advance in a fully specified form. Rather, the activity is unified by the ongoing clarification of what continues to make the activity desirable and valuable, something that might come in forms that continue to change over time. Similar to living, a dialectical activity is atelic, in that one necessarily succeeds at it as one engages in it, rather than in the future. But this activity also looks forward to its continuation, propelled by the desire and the need further to clarify its point and nature (see Millgram 2016).4 Although typical examples of unified actions are telic (e.g., house-building), and standard examples of continuous activities are atelic (e.g., walking), the two classifications should not be conflated. The distinction between continuity and unity concerns the absence versus the presence of a synoptic structure in the activity; the atelic versus telic distinction concerns the temporal relations between the conditions of engagement in a pursuit with the conditions of its success. Some activities might complicate the classification. Consider the typical case of a telic action where the instrumental progression—‘the process’—toward the telos might be considered valuable in an atelic form: engaging in the process is successful as one engages in it even if the process might never culminate in the intended final product (although the intended product matters as far as setting up the constraints on what counts as engaging in that kind of process). For instance, someone might engage in some art-making because one primarily values the process of the art-making even if one never completes the intended piece of art (compare the discussion of the ‘arts of action’ in Nguyen 2020). Another interesting example is a continuous and atelic activity that comprises unified strands of telic components. For instance, consider the playing of a perpetuum mobile—a potentially unending progression of musical steps that relate to each other in non-local ways (such as harmonic modulations). The shorter segments have a unified internal structure. To get the segments right, the performer needs to have a synoptic view over these segments to make sure that they properly fit together. Playing any of these segments is a telic action: the meaningful musical relations take time to unfold and, until that culmination, they have not been successfully resolved. Yet none of these culminations amount to the termination of the atelic activity. Quite the opposite, the resolution of the telic demands propels the music forward, toward a new set of telic and structured demands. The culmination of any of the telic segments is not a terminus but a push toward a potentially indefinite continuation of the musical movement. When observed over a longer period, the playing of the perpetuum mobile has the features of a continuous, ongoing, and atelic activity, since one is indeed successfully engaging in it as one is playing it and not at some later moment of eventual culmination. However, the contribution of the telic elements is not accidental: the very atelic and continuous character of the activity considered in its long-term unfolding depends constitutively on the dynamic interweaving of the telic segments. This complex structure seems to be common to several valuable extended activities, including the ‘dialectical’ ones. 341

Luca Ferrero

4 Temporal goods and values The structure of diachronic agency affects what we might find valuable both in individual pursuits and in the overall shape of our lives. Here is a quick taxonomy of the ‘temporal goods’ that might be pursued or brought about by various forms of diachronic agency. •

•

•

• •

•

Goods of mere continuity: The value of not being (permanently) interrupted, regardless of any other feature of the activity (or a life). Mere continuity is valuable at a particular moment regardless of any prior or future history of the activity (or life) in question. Some paradigmatic examples: the goodness of surviving from one particular moment to the next, the continuation of present pleasure (and conversely, the badness of continuing to suffer a present pain). Temporally additive goods: The value accumulates over time in a simple additive away, that is, with no regard for the history of its accumulation (for instance, financial wealth, where only the amount but not necessarily the provenance of prior wealth affects its future accumulation; hence the saying pecunia non olet—the value of money is not tainted by its origins). Structured, non-processual goods: We value them because of their non-additive nontemporal structure, although a temporal activity might be necessary to bring them about and/or maintain them. Example: a house. Historical goods: We value them because they bear the marks of some of their prior vicissitudes. Example: the value of ruins or historical artifacts. Temporal positional goods: We value them because of the relative or absolute time of their occurrence or existence. Example: the value of being the first of a kind in a temporal sequence (beating a world record, say). Or the value that we assign to some goods relative to their order of appearance in our lives (e.g., when we prefer that gains follow losses, rather than the other way around). Processual goods: We value them based on their dynamics or their intrinsic relation to a process with a dynamic. First, we might value attainments, accomplishments, and achievements as such, that is, as the successful culmination of telic actions (in addition to whatever non-processual goods the action might also bring about)—we value completing the house, not just the finished house. Then there are processual goods produced by the dynamics of the actual engagement into a telic action: making progress, securing the integration of the various steps, adjusting the course of action in the face of perturbations and setbacks, the excitement of investigation and discovery, the tensions produced by the uncertainty of eventual success and the risk of failure, the drama, and so forth. (Notice that we can also enjoy and value these processual good ‘parasitically’ by being vicariously exposed to the dynamics of telic actions in our consumption of narratives.) These processual goods are also present in atelic activities as long as these activities have internal telic components, as in the previous example of the perpetuum mobile. Atelic activities might also give rise to distinctive kinds of processual goods associated with atelic success as contemporaneous to the engagement with the activity and its propensity toward indefinite continuation. For instance, this might be one valuable aspect of the experience of so-called ‘flow’ (Csikszentmihalyi 1990).

Most of these temporal goods can only be enjoyed, appreciated, and pursued by diachronic agents like us. These goods also play a significant role in shaping our conduct and lives, and

342

Diachronic agency

in making them meaningful (see Kauppinen 2015). Unless one could argue that these goods are all reducible to or commensurable with a single kind of value (most likely of an additive sort), their structural variety, including their different temporal profiles, accounts for much of the complexity in determining what is valuable about our diachronic agency. These temporal goods also play an important role in our temporal lives. If we care about our distinctive kind of temporality, we need to recognize the unique role played by historical, positional, and processual goods in our lives. For these goods can be truly appreciated only by beings with the capacity to conceive and engage in temporally unified actions and activities (see Kauppinen 2020 for a discussion of the relation between diachronic agency, values, and the teleological significance of actions). This is especially so for processual goods, which can be brought about only by actual engagement in the corresponding actions and activities (see Nguyen 2020). It might be tempting to think that, among the temporal goods, we should give some priority to those produced by our telic actions, since these actions are paradigmatic examples of our distinctive capacity for future-oriented and temporally unified conduct. But Setiya (2014) warns us that structuring our lives primarily around telic actions can have unwelcome effects—as shown, for instance, by the so-called ‘mid-life crisis.’ According to Setiya, if we engage primarily in actions with telic or finite ends, we face the prospect of a procession of projects that, even if successful, ultimately feel empty and meaningless. The cause is the structural absurdity of pursuing primarily telic ends, whose value is constitutively self-destructive: telic ends might give purpose and direction to our conduct, but their achievement immediately and necessarily extinguishes their power to guide. Engaging with telic ends makes sense only as long as they have not yet been achieved. This absurdity can be avoided only by engaging primarily in atelic activities, that is, activities with infinite ends—such as ‘going for a walk, hanging out with friends, studying philosophy, or living a decent life’ (Setiya 2014:13). What is valuable about these projects is gained as we pursue them. They are valuable because the process of being engaged in them is valuable. This value is by its nature inexhaustible, given that, unlike the value of telic actions, it does not disappear as we succeed in our continuous engagement in any atelic activity (for a discussion of Setiya, see Kauppinen 2022; for a different take on inexhaustible goods via repetition, see Fischer 1994).5 Our mortality is another aspect of our condition as temporal agents that makes a difference to what we find valuable. A worry is often expressed that the impermanence of our deeds and lives deprives them of any meaning and value. In response, some have argued that only finite lives can be meaningful because of the role played by narrative closure in giving significance to a life. But this position appears to conflate our need for some temporally structured values with the necessity of a culmination as a narrative terminus in our deeds and lives. This is not to deny that, as Scheffler (2013) argues, “the aspects of life that we cherish most dearly—love and labor, intimacy and achievement, creativity and humor and solidarity and all the rest—all have the status of values for us because of their role in our finite and bounded lives.” However, the finitude in question need not be that of mortality as the inevitability of a temporally finite life. Instead, what appears necessary is the finitude generated by the scarcity of opportunities for action. This scarcity is a central feature of our temporal existence as mortal beings, but it might not be sufficient to prove the undesirability of all forms of immortality (for it seems possible to conceive of a kind of immortal life that might be affected by a similar scarcity; see Ferrero 2015 and Altshuler 2021).

343

Luca Ferrero

5 The dynamics of the practical standpoint The previous discussion was concerned, to use the linguist’s terms, with the ‘aspectual’ dimension of temporal conduct—the characteristic features of its internal structure. Let’s now consider issues of ‘tense,’ the effects of the (moving) temporal location of the agent relative to her extended conduct (on the distinction between tense and aspect, see Dyke 2013). Of particular interest is the situation of an agent located amidst an ongoing pursuit, when some portions of the pursuit are now in her past, whereas others still lie in her future. Because of the locality of executive control, there is a marked asymmetry between the extent of the agent’s control of the stages of her pursuit, since she can no longer control the past stages and has only indirect control, if any, of the future ones. This asymmetry affects the structure of the agent’s plans. The sequencing of momentary steps reflects their position relative to the agent’s temporal location. Usually, there is also a corresponding asymmetry in the agent’s knowledge about these stages: there is usually less knowledge about the future, which is why the agent needs to handle risk and uncertainty in her plans (for an introduction to questions about diachronic rationality, see Andreou 2021). Notice that agents might approach risk and uncertainty with different psychological profiles (such as different degrees of risk aversion) and different ‘styles’ of planning (including different styles of time-management, such as the particular kind of diachronic irrationality exemplified by procrastination, see Tenenbaum 2010). A temporal asymmetry might also be reflected in how the agent values the relative location of the temporal goods she might gain throughout her pursuit. In particular, the agent might discount the values of costs and benefits relative to her present position. She might discount future goods, or past ones, or both – not necessarily in the same manner. The nature of the temporal goods might also matter as far as both discounting and rationality are concerned (for instance, honoring sunk costs might indeed be a fallacy for merely additive goods but not necessarily for more structured temporal goods, where past stages contribute to the significance of the future ones; see Kelly 2004 and Kauppinen 2020). It is very much an open question what rationality recommends about discounting (see Hare 2013; Dorsey 2017; Kauppinen 2018; Sullivan 2018; Andreou 2021). Here I have only outlined the effects of the passage of time on prudence, but related problems arise about morality as well (see Bykvist 2013). The passage of time might affect the stability of the agent’s practical standpoint. One’s desires, preferences, saliences, cares, and values might change over time, either temporarily or permanently. If this change occurs within a particular pursuit, it raises the issue of whether the agent should (and if so, how) continue to sustain the project through its completion (for a discussion of temptation, see Bratman 1999: ch.3–4; 2007: ch.12; Holton 2009). Especially challenging for practical rationality are cases in which the changes in practical standpoint are extensive and permanent, regardless of whether they occur quickly (as it happens in many transformative experiences, Paul 2014) or gradually (either by drift or via what Callard 2018 calls ‘aspiration’). In closing the discussion of effects of ‘tense’ on diachronic agency, it is important to remember that all the asymmetries generated by the agent’s relative temporal position within a particular pursuit are not static. When I first illustrate them, I focused on a particular moment in the ongoing project as the source of asymmetries: the agent’s present moment. But in conceptualizing the phenomena of tense and the correlated rational demands, we must always consider the dynamic of the moving center of the asymmetries. The passage of time 344

Diachronic agency

relentlessly recenters the source of the asymmetries. This has a compounding effect on the difficulties that we already encounter, at a single static moment, when we are trying to account for our different relations to past, present, and future stages of our activities and their associated temporal goods. The problem is particularly acute when the changes are nonlinear, including the inexorable change of temporal status that, sooner or later, affect all stages as they recede from the future into the present and eventually into the past.

6 Temporal integration of actions, activities, and the agent For brevity’s sake, this entry focuses on particular pursuits of individual agents. In closing, I want to mention some further complications (while remaining within the confines of individual agency). First, rationality demands that multiple extended pursuits be integrated not just at a time but also over time. The simplest way to avoid conflicts is to schedule them at different times, but this does not always work. Integration often requires that conflicting pursuits be either ranked—by giving priority to some while subordinating others—or qualified—by making some ends and the corresponding intentions either partial (Holton 2009), conditional (Ferrero 2009a) (Ludwig 2015), or disjunctive (Ferrero 2016). Second, extended actions and activities are of the same extended agent. The agent’s temporally extended existence is subject to the same diachronic constraints of their agency. Hence, we need to sustain our existence as agents in the face of the locality of executive powers, the scarcity of resources, and mortality. Besides, the distinction between continuity and unity also applies to the structure of our existence. To the extent that we engage in temporally integrated pursuits, it is not sufficient that we just survive from moment to moment in the mode of mere continuity. We also need a relatively stable practical standpoint and some kind of temporal identification to underwrite our engagement in extended and integrated pursuits that respect our diachronic autonomy. This stability and sense of identity are not a given. They rather need to be secured in the face of the constraints imposed by our predicament as temporal beings. Last but not least, both the stability and the sense of identity are not static phenomena: they need to be secured in light of the same dynamical asymmetries between past, present, and future that already complicate our engagement in temporal actions and activities (see Ferrero, 2022).

Related entries Agency and practical reasoning; Planning agency; Intentional agency; Agency, time, and rationality; Agency and personal identity; Agency, narrative, and mortality; Agency, events, and processes; Agency and games

Notes 1 In this entry, I will use ‘action’ and ‘activity’ interchangeably, although ‘activity’ can sometimes refer more specifically to extended pursuits with so-called atelic ends; see discussion below. 2 The distinction between telic and atelic pursuits is often presented as matching Aristotle’s distinction between poiesis (as a species of kinesis) and praxis (as a species of energeia) 3 This is not to deny that there might be atelic activities in which termination counts as a failure, even if the activity has necessarily succeeded up to the point of termination. Take the activity of living. Although it is true that, if one is living at any particular moment, one has thereby succeeded in having lived up to that moment included, the termination of this activity seems to be a failure.

345

Luca Ferrero For living is supposed to secure at least one’s proximal survival, that is, the continuation of one’s living from one moment to the next. Unlike walking, living might be constitutively oriented toward its continuation, at least into the immediate future. If this is so, from the point of view of the atelic end of living, death is always premature. This is not to say that the activity of living is trying to maximize its overall temporal extension. The maximization is not the guise under which living propels itself into the indefinite future. It is nonetheless possible to pursue any atelic activity in that form, but, when so, this is via a second-order telic end of maximization of the extension of the pursuit of a first-order atelic one. In this scenario, the interruption of the pursuit of the atelic end might not necessarily be a failure of the atelic activity as such, but it would be a failure of the pursuit of the second-order telic end of maximization. 4 Dialectical activities should not be confused with telic activities whose ends are not yet fully specified, such as a specificationist deliberation that aims to figure out one’s concrete objective before engaging in the corresponding action (see Millgram 2020). Likewise, they differ from ‘aspirational’ activities, which aim to acquire a new practical standpoint and are successful only when the new values have been fully acquired (on aspiration, see Callard 2018). 5 It is noticeable that most examples of valuable atelic activities offered by Setiya are highly structured activities, which either include subordinate telic ends on the model of the perpetuum mobile or might fit the mold of Brewer’s dialectical activities or both.

Further reading Bratman M. (1987). Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Bratman M. (2007). Structures of Agency. New York: Oxford University Press. Bratman, M. (2018). Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. A series of seminal books on planning agency as our characteristic form of diachronic agency; see also the ‘Planning agency’ entry in this volume. Kauppinen, A. (2022). “Against Seizing the Day.” Oxford Studies in Normative Ethics 11. Setiya, K. (2014). “The Midlife Crisis.” Philosophers’ Imprint 14(31): 1–18. Two papers debating the implications of the distinction between telic and atelic ends for the meaningfulness of our pursuits and existence. Ferrero L. (2015). “Agency, Scarcity, and Mortality.” The Journal of Ethics 19(3–4): 349–378. Hägglund, M. (2014). “Chronolibido,” in A. Mukherjee and L. Marcus (Eds.), A Concise Companion to Psychoanalysis, Literature, and Culture: 312–327. Scheffler S. “Fear, Death, and Confidence,” In S. Scheffler and N. Kolodny (Eds.), Death and the Afterlife. New York: Oxford University Press: 83–112. Three papers discussing the relationship between our diachronic agency, temporal goods, and mortality.

References Altshuler, R. (2021). “Agency, Narrative, and Mortality,” this volume. Andreou, C. (2021). “Agency, Time, and Rationality,” this volume. Bratman, M. (1987). Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Bratman, M. (1999). Faces of Intention. Cambridge: Cambridge University Press. Bratman, M. (2007). Structures of Agency. Oxford: Oxford University Press. Bratman, M. (2021). “Planning Agency,” this volume. Brewer, T. (2009). The Retrieval of Ethics. Oxford: Oxford University Press. Bykvist, K. (2013). “Time and Morality.” In H. Dyke and A. Bardon (Eds.), A Companion to the Philosophy of Time. Hoboken, NJ: Wiley-Blackwell: 549–562. Callard, A. (2018). Aspiration: The Agency of Becoming. Oxford: Oxford University Press. Csikszentmihalyi, M. (1990). Flow. The Psychology of Optimal Experience. New York, NY: Harper and Row. Dorsey, D. (2017). “Future-Bias: A (Qualified) Defense.” Pacific Philosophical Quarterly 98: 351–373. Dyke, H. (2013). “Time and Tense.” In H. Dyke and A. Bardon (Eds.), A Companion to the Philosophy of Time. Hoboken, NJ: Wiley-Blackwell: 328–344.

346

Diachronic agency Ferrero, L. (2009a). “Conditional Intentions.” Noûs 43(4): 700–741. Ferrero, L. (2009b). “What Good Is a Diachronic Will?” Philosophical Studies 144(3): 403–430. Ferrero, L. (2010). “Decisions, Diachronic Autonomy, and the Division of Deliberative Labor.” Philosophers’ Imprint 10(2): 1–23. Ferrero, L. (2015). “Agency, Scarcity, and Mortality.” The Journal of Ethics 19(3–4): 349–378. Ferrero, L. (2016). “Pro-Tempore Disjunctive Intentions.” In R. Altshuler and M. J. Sigrist (Eds.), Time and the Philosophy of Action. Abingdon: Routledge: 108–123. Ferrero, L. (2022). “The Structures of Temporally Extended Agents.” In C. Bagnoli (Ed.), Time in Action: The Temporal Structure of Rational Agency and Practical Thought. Abingdon: Routledge. Fischer, J. M. (1994). “Why Immortality Is Not So Bad.” International Journal of Philosophical Studies 2(2): 257–270. Hare, C. (2013). “Time – The Emotional Asymmetry.” In H. Dyke and A. Bardon (Eds.), A Companion to the Philosophy of Time. Hoboken, NJ: Wiley-Blackwell: 507–520. Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press. Kauppinen, A. (2015). “Meaningfulness.” In G. Fletcher (Ed.), Routledge Handbook of the Philosophy of Well-Being. Abingdon: Routledge. Kauppinen, A. (2018). “Agency, Experience, and Future Bias.” Thought: A Journal of Philosophy 7(4): 237–245. Kauppinen, A. (2020). “Prudence, Sunk Costs, and the Temporally Extended Self.” Journal of Moral Philosophy 17(6): 658–681 Kauppinen A. (2022). “Against Seizing the Day.” Oxford Studies in Normative Ethics, 11. Kelly, T. (2004). “Sunk Costs, Rationality, and Acting for the Sake of the Past.” Noûs 38(1): 60–85. Ludwig, K. (2015). “What Are Conditional Intentions?” Methode: Analytic Perspectives 4(6): 30–60. Millgram, E. (2020). “Practical Reason and the Structure of Actions.” In Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2020 Edition). https://plato.stanford.edu/archives/ win2020/entries/practical-reason-action/. Morton, J. M. (2017). “Reasoning under Scarcity.” Australasian Journal of Philosophy 95(3): 543–559 Nguyen, C. T. (2020). “The Arts of Action.” Philosophers’ Imprint 20(14): 1–27. Paul, L. A. (2014). Transformative Experience. Oxford: Oxford University Press. Rödl, S. (2010). “The Form of the Will.” In Tenenbaum, S. (ed.), Desire, Practical Reason, and the Good. Oxford: Oxford University Press: 138–160 Rödl, S. (2011). “Two Forms of Practical Knowledge and Their Unity.” In A. Ford, J. Hornsby, and F. Stoutland (Eds.), Essays on Anscombe’s Intention. Cambridge, MA: Harvard University Press: 211–241. Scheffler, S. (2013). Death and the Afterlife. New York: Oxford University Press. Setiya, K. (2014). “The Midlife Crisis.” Philosophers’ Imprint 14(31): 1–18. Sullivan, M. (2018). Time Biases: A Theory of Rational Planning and Personal Persistence. Oxford: Oxford University Press. Tenenbaum, S. (2010). “The Vice of Procrastination.” In C. Andreou and M. White (Eds.), The Thief of Time. Oxford: Oxford University Press: 151–164.

347

31 PLANNING AGENCY Michael E. Bratman

1 Human practical organization One question in the philosophy of agency is as follows: What is it to be an agent? A plausible strategy here draws on an idea from Harry Frankfurt (Frankfurt 1988a). Agency involves behavior that is under the control of a guidance mechanism, one that can track an end and adjust the behavior in its direction, and one whose operation constitutes the activity of the agent, not merely of a subsystem. The last clause raises issues of circularity, but is needed to distinguish the activity of the agent from the digestive activity of her stomach. This first question focuses on agency quite generally, including the agency of nonhuman animals. A second question focuses on a special case: What is it to act for a reason? This question has occupied center stage in the literature framed by the work of Elizabeth Anscombe (2000) and Donald Davidson (1980a). And much of this literature supposes that there is a close connection between acting for a reason and acting intentionally. One issue here is whether, as Davidson argued, the relevant connections between thought and action are not only rational but also in an important sense causal. A second issue is whether, as Anscombe argued, there is a distinctive kind of self-knowledge involved. A philosophically informed theory of our human agency needs to answer these questions. But it also needs to address further matters about human, practical organization. One striking feature of our human agency is its organization over time. Think about growing food in one’s garden. If successful, this will involve complex forms of cross-temporal coordination and organization. And these forms of diachronic organization will be to some extent grounded in the thought of the agent, in contrast with, say, the development of an acorn into an oak. A second striking feature of our human agency is its social organization. Here, as a first step, we can focus on small-scale cases of shared intentional activity. Examples include walking together (Gilbert 1990), playing a quartet together, painting a house together. When we, for example, walk together, there is not simply the public responsiveness of each to each that is involved in walking alongside each other without a collision. There is a further kind of sharing, one that involves distinctive forms of social organization that are to some extent grounded in the thinking of the participating agents. A theory of our human agency needs not only to answer our questions about agency and acting for a reason, it also should provide resources for understanding these basic forms of

348

DOI: 10.4324/9780429202131-39

Planning agency

human practical organization and their interrelations. In the end these ideas about agency, reason, intention, knowledge, time, and sociality need to fit together. When we focus on these latter issues about human practical organization, it is plausible to turn to our planning capacities. My planning helps explain the cross-temporal organization of my growing food in my garden, and our planning helps explain how we successfully paint the house together. This supports the thought that our capacity for planning agency lies behind these forms of practical organization, both cross-temporal and social. And this then makes room for the idea that this capacity also underlies other important forms of human practical organization. These thoughts about the core role of our capacity for planning agency support a focus on such agency. A first step in developing these ideas is to articulate a model of individual human planning agency.

2 Planning and the diachronic organization of individual agency A basic thought is that planning agency is a distinctive kind of goal-directed agency, one that involves attitudes of intention, many of which are future-directed and settle matters about what to do later. These intentions involve a commitment to future action that is not just ordinary expectation: after all, I might expect that I will later give into a certain temptation but not now intend to do so. Davidson (1980b) sees this commitment to later action as a judgment that, as things now look, this is what it would be best to do then. But this is problematic: I might so judge but still see the question of whether to A as open and not settled; I might judge that A and B are equally best but intend A and not intend B; I might (albeit, irrationally) settle on A even while judging it will not be the best option (Bratman 1999b). Future-directed intentions are normally embedded in larger, coordinating plans. These larger plans normally have a hierarchical, end-means structure, and these plans will typically be partial in the sense that they do not yet specify all the steps needed for each intended end (Harman 1976; Bratman 1987). Such partiality of our plans seems inevitable given our limits (Simon 1983). We normally do not know enough or have the mental resources to fill in all the details of our plans prior to putting them in place as guides to thought and action. We need to settle in advance on partial plans and then fill them in as needed and as time goes by. This involves a kind of self-trust. And it induces a need for characteristic forms of further reasoning aimed at appropriately filling in prior, partial plans: prior partial plans provide a background framework within which relevant reasoning–including the weighing of reasons–takes place. Intentions are states in this planning system (Harman 1976; Bratman 1987; but see Tenenbaum 2016). Though subject to revision, these plan states have a characteristic stability over time. They are responsive to pressures to fill in partial plans as needed with specifications of means and the like. And they are responsive to pressures for consistency of the different things one intends with each other and with what one believes (but see Núñez 2019). These tendencies toward diachronic stability and synchronic means-end coherence and consistency help support the cross-temporal organizing roles of such plans. And these tendencies correspond to associated rationality norms of stability, coherence, and consistency, norms whose (implicit) acceptance is at work in the functioning of a planning agent. These roles and norms help distinguish intentions from ordinary desires and beliefs. After all, desiring things that are not by our lights co-possible is all too human. And a belief that one will be doing something later need not rationally require that one settle on means to doing it. 349

Michael E. Bratman

Some think that while intentions are not ordinary beliefs, they are nevertheless a special kind of belief or belief-like attitude (Harman 1976, Velleman 1989, Setiya 2004, but see Alonso 2021). It might be thought that this appeal to belief is needed to explain the norm of means-end coherence (Velleman 2014, but see Bratman 2018b). And it might also be thought that this appeal to belief is needed to explain what is involved in planning for a further future on the basis of a prior intention concerning a nearer future (Velleman 2014, but see Bratman 1987: 37–39; Bratman 1999a; Alonso 2020). My own approach avoids such a cognitivist model of intention and instead sees intentions as distinctive practical attitudes that are characterized by their roles in our rational plan dynamics (Bratman 1987; Paul 2009). Returning to that plan dynamics, if all goes well, planning structures induce crosstemporal referential connections that are both forward and backward looking. My present plan to go to Boston next month at least implicitly refers to my later, then-present-directed intention to go by getting on the airplane, and my later intention at least implicitly refers back to my earlier intention. And the normal stability of such intentions over time helps support a coordinated flow of activity over time. These cross-temporal constancies and referential interconnections help support a temporally extended structure of partial plans that provides a background framework for further reasoning aimed at filling in these plans as need be and as time goes by. And this further reasoning is shaped in part by the cited rational pressures. In these ways, a planning agent’s activity over time is typically embedded within structures of interwoven, partial, referentially interlocking, hierarchical, and more or less stable plan states, and a rational dynamics within which plan states provide a background framework for ongoing thought and action. This idea of cross-temporally stable and referentially interlocking attitudes is familiar from Lockean approaches to personal identity over time (Parfit 1984; Yaffe 2000). The standard functioning of planning agency involves such Lockean cross-temporal ties, and thereby plays a role in the Lockean persistence of the agent over time. Note three initial forms of support for stability of intention over time (Bratman 1987). First, acting on an intention normally changes the world, and these changes may make it increasingly sensible to continue to act on that intention. There can be such a snowball effect because in acting on an intention one gets closer to its target. But there can also be social snowball effects. Once it is public that one has certain intentions, there may be social costs – including impacts on social coordination and on reputation – in later abandoning them. Second, reconsidering a prior intention takes time, uses mental resources, and may require rethinking other courses of action on which one had earlier settled. So, there is frequently reason not to reconsider, both because of the costs of reconsideration and because of risks of undermining coordination previously forged. And in the absence of reconsideration, a prior intention will tend to persist. Third, given our limited mental resources, we frequently depend on general, nondeliberative habits about when to reconsider. And it seems that habits that to some extent favor non-reconsideration would tend to support the overall effectiveness of our resource-limited agency. I will later add to these ideas by considering our self-governance over time. For now, however, we have said enough about the functioning of planning agency to see how that functioning will normally support important forms of cross-temporal organization of our temporally extended agency. The capacity for temporally organized intentional agency involves the capacity to guide one’s activities in light of one’s grasp of their location in a larger, temporally extended structure of what one has intentionally been doing and what one is committed to 350

Planning agency

doing (Ferrero 2009). The psychic economy of a planning agent realizes this capacity. The interwoven, interlocking, and somewhat stable structures of partial and hierarchical plan states normally guide present activity as an element in larger activities favored by these plan states, and it normally involves an explanatorily relevant grasp of salient relations between temporally larger activities and their temporal subparts. The conjecture, then, is that our temporally organized intentional action is normally grounded in some such planning psychology and its embedded intentions (Compare Lavin 2016). This leaves open the precise relation between what one intends and what one does intentionally. In many cases, one’s intentionally A-ing will be grounded in one’s intention to A. But there can be cases in which it is grounded in an intention in favor of a related but different action. One example of this breakdown in a “simple view” of the relation between intentional action and intention will be certain cases of intentionally bringing about significant, expected, but not, strictly speaking intended, side effects (Harman 1976; Bratman 1987: ch. 8; Knobe 2003). This support of this overall planning psychology for temporally organized intentional agency is no surprise. If we ask why bother with planning – why not just cross our bridges when we come to them? – the commonsense answer will appeal to its support for the cross-temporal organization of our agency. We have been exploring the deep structure of this plan-infused way in which the human mind supports this cross-temporal organization (Bratman 2018d). Turn now to small-scale sociality.

3 Planning and the organization of shared intentional activity Earlier I contrasted our walking together with mere public interdependence of intentional activities of each. There is also a contrast with a case in which we exchange promises to walk with each other, thereby incurring obligations to perform. Such exchanges, and their associated obligations, are not sufficient for shared intentional activity: after all, each might promise insincerely. Nor are such promises necessary for our walking together (Hume 1978: 490). In pursuit of a model of shared intentional activity (SIA) – and associated shared intentions – we need to thread a path between walking alongside strangers and a web of promissory obligations. Here we can again turn to our planning agency and pursue a strategy of construction: We consider plan states of each participant, states whose contents are in certain ways social. We identify relevant interpersonal interrelations across those plan states. And we describe ways in which these plan states of each, in functioning in accordance with the rational dynamics of individual planning agency, interdependently work their way through to joint action. We thereby provide a plan-theoretic construction of such forms of social organization. Suppose, for example, that you and I paint the house together, and that this is a SIA. We can develop a plan-theoretic construction of this phenomenon by highlighting six conditions. First, each of us intends not just to paint, but that we paint. Second, each intends that we paint in part by way of the other person’s intention that we paint: in this sense the intentions of each referentially interlock. Third, we each intend that we paint by way of sub-plans of each of us that mesh in the sense of being co-compatible. Fourth, there is interdependence in the persistence of these intentions of each. Fifth, these intentions of each lead to our painting the house by way of mutual responsiveness of each to each, mutual responsiveness that tracks the intended joint activity. And sixth, these mental structures are out in the open (Bratman 2014). 351

Michael E. Bratman

I argue (Bratman 2014) that these public, interdependent, and interlocking intentions of each will, in responding to the rational pressures involved in individual planning agency, function together in ways that, if all goes well, constitute SIA. A key is that if you and I are involved in such an interpersonal planning structure, I do not just intend to do my part while expecting you to do yours, as I might intend to walk to my left knowing that the stranger is walking to my right. Rather, I intend that we act in part by way of your analogous intention and meshing sub-plans of our intentions, and intending is not merely expecting. This means that the rational pressure on me to make my plans coherent and consistent – pressure built into individual planning agency – ensures rational pressure on me to mesh with and, as needed, support your relevant plans, and vice versa. This will frequently involve rational pressure on each to help the other if needed. In this way, rational pressures on the individual planning agents – given suitable contents of, and interrelations between their plans – induce rational pressures in favor of forms of social coherence and consistency that are characteristic of shared intentional activity. These pressures of social rationality depend on the presence of relevant intentions of each participant, intentions whose continued persistence is supported by pressures for stability. And there will be these rational pressures even when – as is common – each participates in the shared activity for different reasons. The claim is that this plan-theoretic construction is sufficient for such socially organized shared activity. While this leaves open the possibility that there are other realizations of these forms of social organization, the provision of such sufficient conditions points toward a broad individualism concerning SIA. This plan-theoretic construction of SIA draws primarily on conceptual resources from the domain of individual planning agency (though there are complexities about the out-in-the-open condition). It sees the metaphysics of SIA as a construct of metaphysical resources already in play in individual planning agency. And it sees certain rational pressures characteristic of shared intentionality as rooted in rational pressures central to individual planning agency. This assumes we can make sense of the idea of my intending that we act in certain ways, and of the idea that such intentions of each are characteristic of SIA. These ideas can be challenged (Baier 1997; Velleman 2000; Stoutland 2002). I respond to these challenges in Bratman (1999c, 2014: ch. 3). Here let me note that a basic rationale for the introduction of these ideas is their theoretical fecundity: we thereby make progress with our questions concerning the interrelation between the cross-temporal and the social organization of our human practical lives.

4 Planning and self-governance We have explored ways in which our capacities for the cross-temporal and small-scale social organization of action can be grounded in our capacity for planning agency. Turn now to potential roles of planning structures in the psychological organization central to our self-governance. What is self-governance? Well, in understanding our individual self-governance, we need to avoid appeal to a little person inside who does the work. We seek instead a nonhomuncular model, one that appeals to the relevant “operation … of the systems we are” (Frankfurt 1988a: 74). Here a basic strategy is to suppose that the agent has a practical standpoint, one that consists of attitudes that constitute her stance with respect to relevant practical issues. When her practical standpoint appropriately guides, the agent governs (Frankfurt 1988b). The idea is that certain attitudes play roles in the agent’s psychic economy such that their guidance constitutes the agent’s direction of action. Further, this guidance involves practical thinking of a sort that qualifies this agential direction of action as agential governance. 352

Planning agency

What attitudes are these? My proposal is that attitudes whose guidance can be the agent’s direction are attitudes that provide the Lockean backbone of that agent’s persistence over time – attitudes that help constitute and support interlocking, Lockean structures of agency and mind of that very person over time. And for this to count as self-governance these attitudes need to play these Lockean roles in part by way of their roles in relevant practical reasoning (Bratman 2007, 2017; but see Millgram 2014). Note next that relevant plan states satisfy these design specifications. As we have seen, plan states have it as a characteristic function to constitute and support relevant Lockean cross-temporal ties. Further, plan states help shape characteristic forms of practical reasoning. This includes both the framing of end-means reasoning and deliberation shaped by policy-like attitudes in favor of giving certain considerations certain weights (Bratman 2007; Murray and Buchak 2019). This gives us a preliminary model of a planning agent’s synchronic self-governance, one that highlights guidance of thought and action by a plan-infused standpoint. Such a planinfused standpoint needs to avoid plan inconsistency or incoherence, since that would normally block the condition, for non-homuncular self-governance, of a clear place where the agent stands. So, a violation of the synchronic planning norms of consistency and coherence is a pro tanto violation of a condition of relevant self-governance (Bratman 2018c). What about self-governance over time? What is it to govern one’s life over time and not just at each time along the way? Even if the person governing at each time is one and the same person, that person might lurch erratically from one standpoint to another, thereby – or so it seems – undermining self-governance over the time period. So, what else is needed? My proposal is that a planning agent’s self-governance over time, given that she is self-governing at relevant times along the way, can be understood by appeal to the metaphor of acting together with oneself over time (Bratman 2018e; but see Nefsky and Tenenbaum forthcoming). The idea is to highlight cross-temporal interconnections within the planned temporally extended activity of an individual agent, interconnections that are analogous to the interpersonal interconnections characteristic of shared intentional activity. For example, just as in SIA each intends that the activity proceed by way of relevant intentions of each other, so in diachronic self-governance the agent at t1 intends that the temporally extended activity proceed by way of her associated intentions at t2, and the agent at t2 at least implicitly intends to be following through with her relevant intentions at t1. This brings our model of diachronic self-governance together with our plan-theoretic model of SIA. This is a version of an important parallel between the cross-temporal organization of an individual’s activity and interpersonal, social organization. This has implications for issues about willpower and about stability in the face on incommensurability. Suppose you know you will be tempted to drink heavily tonight at the party. You now think that would be a mistake. However, you know that at the party your evaluation will shift in favor of drinking more. You also know that if you did drink heavily you would, at a yet later time, regret that. So, this morning you resolve to drink only one glass tonight (Holton 2009; Paul 2014). What we can now say is that following through with your resolve can, in certain such cases, be a way of avoiding a breakdown in diachronic self-governance. This is in part because it can be a way of retaining relevant cross-temporal interconnections of intention. (For further complexities see Bratman 2018a: 10–11.) In a second case, one decides in the face of incommensurable temporally extended options and then, in the process of follow through, is later faced with continued incommensurability. Broome (2001) offers the example of Abraham, who must choose whether to follow what 353

Michael E. Bratman

he sees as a divine command to sacrifice his son. He plausibly sees this as a choice between incommensurable values. Suppose that he decides to proceed but while in transit reconsiders and sees that the incommensurability remains. What we can now say is that following through with such a decision can, in certain such cases, be a way of avoiding a breakdown in diachronic self-governance. This is, again, in part because it can be a way of retaining relevant cross-temporal interconnections of intention (Bratman 2018e). In these ways, planning structures lie behind forms of organization that are central to our self-governance, both synchronic and diachronic. And this may help us better understand the grounds for the norms of consistency, coherence, and stability that play a basic role in our planning agency (Bratman 2018a; also see Ferrero 2014; Millgram 2019).

5 Fecundity of planning agency Let’s take stock. The step from present-focused goal-directed activity to planning agency supports important forms of temporally organized intentional agency. It also provides – given appropriate contents, contexts, and interrelations – a backbone for shared intentional agency. Basic norms of individual intention rationality – norms of consistency, coherence, and stability – frame much of our practical thinking. In the context of SIA they also help induce conformity to associated norms of social rationality. Because of their Lockean roles in temporally extended agency, and their related roles in practical reasoning, planning structures can help constitute our self-governance, both at a time and over time. In these interconnected ways, the fact that we are planning agents helps shape the temporality and sociality of our agency, our governance of our lives, and the structure of our rational guidance. Does this fecundity extend yet further? Does the plan-theoretic support for small-scale cases of SIA extend to the complex forms of organization involved in larger, organized institutions such as a neighborhood association, a corporation, a legal system, or a democracy? Versions of a positive answer are suggested (Shapiro 2011; Chapman 2020; Bratman forthcoming), but this remains a question for further research. Again, Allan Gibbard proposes that we understand normative judgments in terms of plans and planning (Gibbard 2003). Justin Snedegar argues that within partial-plan-framed practical reasoning normative reasons are relativized to deliberative questions (Snedegar 2018). And there are deep issues about how the plan-theoretic appeal to future-directed practical commitments interacts with the idea, central to most work in decision theory, that the basic parameters of practical thinking are provided by utilities and probabilities (Bratman, Israel, Pollack 1988; van Hees and Roy 2008; Gold 2018; Murray and Buchak, 2019). But, again, these are matters for further research.

Related entries Diachronic agency; Intentional agency; Agency, time, and rationality; Artificial and machine agency; Agency and autonomy; Agency and personal identity.

Further reading Bratman, M.E. 1987. Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Develops planning model of individual temporally extended agency. Bratman, M.E. 2014. Shared Agency: A Planning Theory of Acting Together. New York: Oxford University Press. Develops planning model of shared intentional agency.

354

Planning agency Harman, G. 1976. “Practical Reasoning,” Review of Metaphysics 29/3: 431–463. Introduces the importance of planning to intention and practical reasoning. Millgram, E. 2014. “Segmented Agency,” in M. Vargas and G. Yaffe (eds.) Rational and Social Agency: The Philosophy of Michael Bratman. New York: Oxford University Press: 152–189. Challenges the planning theory’s approach to the unity of our agency. Velleman, J.D. 2014. “What Good Is a Will?” in M. Vargas and G. Yaffe (eds.) Rational and Social Agency: The Philosophy of Michael Bratman. New York: Oxford University Press. Challenges the idea that intentions are practical commitments that are not belief-like attitudes.

References Alonso, F. (2020) “Planning on a Prior Intention,” Journal of Ethics and Social Philosophy 18:3: 229–265. Alonso, F. (2021) “The Limits of Partial Doxasticism,” Philosophical Quarterly. doi: 10.1093/pq/pqab027 Anscombe, G.E.M. (2000) Intention, Second Edition. Cambridge, MA: Harvard University Press (Originally published 1957) Baier, A.C. (1997) “Doing Things with Others: The Mental Commons,” in L. Alanen, S. Heinämaa, and T. Wallgren, eds., Commonality and Particularity in Ethics. London: MacMillan. Bratman, M.E. (1987) Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. (Re-issued CSLI Publications 1999.) Bratman, M.E. (1999a) “Practical Reasoning and Acceptance in a Context,” in M.E. Bratman ed., Faces of Intention. Cambridge: Cambridge University Press. Bratman, M.E. (1999b) “Davidson’s Theory of Intention” in M.E. Bratman ed., Faces of Intention. Cambridge: Cambridge University Press. Bratman, M.E. (1999c) “I Intend that We J,” in M.E. Bratman ed., Faces of Intention. Cambridge: Cambridge University Press. Bratman, M.E. (2007) “Three Theories of Self-Governance,” in M.E. Bratman ed., Structures of Agency. New York: Oxford University Press. Bratman, M.E. (2014) Shared Agency: A Planning Theory of Acting Together. New York: Oxford University Press. Bratman, M.E. (2017) “A Planning Theory of Self-Governance: Reply to Franklin,” Philosophical Explorations 20: 15–20. Bratman, M.E. (2018a) “Introduction: The Planning Framework,” in M.E. Bratman ed., Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. Bratman, M.E. (2018b) “Intention, Belief, Practical, Theoretical,” in M.E. Bratman ed., Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. Bratman, M.E. (2018c) “Intention, Practical Rationality, and Self-Governance,” in M.E. Bratman ed., Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. Bratman, M.E. (2018d) “Agency, Time, and Sociality,” in M.E. Bratman ed., Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. Bratman, M.E. (2018e) “A Planning Agent’s Self-Governance Over Time,” in M.E. Bratman ed., Planning, Time, and Self-Governance: Essays in Practical Rationality. New York: Oxford University Press. Bratman, M.E. (forthcoming) Shared and Institutional Agency: Toward a Planning Theory of Human Practical Organization. New York: Oxford University Press. Bratman, M.E., Israel, D. and Pollack, M. (1988), “Plans and Resource-Bounded Practical Reasoning,” Computational Intelligence 4: 349–355. Broome, J. (2001) “Are Intentions Reasons? And How Should We Cope with Incommensurable Values?’ in C. W. Morris and A. Ripstein, eds., Practical Rationality and Preference: Essays for David Gauthier. Cambridge: Cambridge University Press. Chapman, E. (2020) “Shared Agency and the Ethics of Democracy,” Georgetown Journal of Law and Public Philosophy. 18, Issue 2: 705–732. Davidson, D. (1980a) Essays on Actions and Events. Oxford: Oxford University Press. Davidson, D. (1980b) “Intending,” in D. Davidson ed., Essays on Actions and Events, Oxford University Press. Ferrero, L. (2009) “What Good is a Diachronic Will?” Philosophical Studies 144: 403–430.

355

Michael E. Bratman Ferrero, L. (2014) “Diachronic Structural Rationality,” Inquiry 57: 311–336. Frankfurt, H. (1988a) “The Problem of Action,” in H. Frankfurt ed., The Importance of What We Care About. Cambridge: Cambridge University Press. Frankfurt, H. (1988b) “Identification and Wholeheartedness,” in H. Frankfurt ed., The Importance of What We Care About. Cambridge: Cambridge University Press. Gibbard, A. (2003) Thinking How to Live. Cambridge, MA: Harvard University Press. Gilbert, M. (1990) “Walking Together: A Paradigmatic Social Phenomenon,” Midwest Studies 15: 1–14. Gold, N. (2018) “Putting Willpower into Decision Theory: The Person as a Team Over Time and Intrapersonal Team Reasoning,” in J. Bermúdez, ed., Self-Control, Decision Theory, and Rationality: New Essays. Cambridge: Cambridge University Press. Harman, G. (1976) “Practical Reasoning,” Review of Metaphysics 29(3): 431–463. Holton, R. (2009) Willing, Wanting, Waiting. Oxford: Oxford University Press. Hume, D. (1978) A Treatise of Human Nature. Oxford: Oxford University Press. Knobe, J. (2003) “Intentional Action and Side-Effects in Ordinary Language,” Analysis 63: 190–193. Lavin, D. (2016) “Action as a form of temporal unity: on Anscombe’s Intention,” Canadian Journal of Philosophy 45: 609–629. Millgram, E. (2014) “Segmented Agency,” in M. Vargas and G. Yaffe, eds., Rational and Social Agency: The Philosophy of Michael Bratman. New York: Oxford University Press: 152–189. Millgram, Elijah. (2019) “Review of Planning, Time, and Self-Governance,” Notre Dame Philosophical Reviews 2019.05.15. Murray, D. and Buchak, L. (2019) “Risk and Motivation: When the Will Is Required to Determine What to Do,” Philosophers’ Imprint 19(16): 1–12. Nefsky, J. and Tenenbaum, S. (forthcoming) “Extended Agency and the Problem of Diachronic Autonomy,” in Carla Bagnoli, ed., Time in Action. New York: Routledge. Núñez, C. (2019) “Requirements of Intention in Light of Belief,” Philosophical Studies. https://doi. org/10.1007/s11098-019-01321-0. Parfit, D. (1984) Reasons and Persons. Oxford: Clarendon Press. Paul, S. (2009) “How We Know What We’re Doing,” Philosophers’ Imprint 9(11):1–24. Paul, S. (2014) “Diachronic Incontinence is a Problem in Moral Philosophy,” Inquiry 57: 337–355. Setiya, K. (2004) “Explaining Action,” Philosophical Review 112: 339–393. Shapiro, S. (2011) Legality. Cambridge, MA: Harvard University Press. Simon, H. (1983) Reason in Human Affairs. Stanford: Stanford University Press. Snedegar, J. (2018) “Deliberation, Reasons, and Alternatives,” Pacific Philosophical Quarterly early view doi:10.1111/papq.12262. Stoutland, F. (2002) “Critical Notice of Faces of Intention,” Philosophy and Phenomenological Research 65: 238–240. Tenenbaum, S. (2016) “Reconsidering Intentions,” NOÛS 52: 443–472. van Hees, M., & Roy, O. (2008). “Intentions and Plans in Decision and Game Theory,” in B. Verbeek, ed., Reasons and Intentions. Farnham: Ashgate: 207–225. Velleman, J.D. (1989) Practical Reflection. Princeton, NJ: Princeton University Press (Re-issued CSLI Publications 2007). Velleman, J.D. (2000) “How to Share an Intention,” in J.D. Velleman, ed., The Possibility of Practical Reason. Oxford: Oxford University Press. Velleman, J.D. (2014) “What Good Is a Will?” in M. Vargas and G. Yaffe, eds., Rational and Social Agency: The Philosophy of Michael Bratman. New York: Oxford University Press. Yaffe, G. (2000) Liberty Worth the Name: Locke on Free Agency. Princeton, NJ: Princeton University Press.

356

32 AGENCY, TIME, AND RATIONALITY Chrisoula Andreou

1 Introduction Consider the following description: B faces a choice point where she can either X and get payoff R or not X and get payoff S. Though this description might seem quite mundane, reflection on it raises some interesting issues. Given that virtually all actions or courses of action that we engage in take place over time, rather than being instantaneous, B’s X-ing will occur not at a point in time but over a span of time. As such, B can be capable of (intentionally) X-ing only if B is a temporally extended agent that can have and retain the goal of X-ing (or of a doing that involves X-ing) from the initiation of the action or course of action to its completion. If, however, X-ing takes more than just a brief moment, and there are, practically speaking, multiple potential points of reconsideration, it might be wondered whether what is described as B’s X-ing is better described as a series of momentary agents {B1, B2,…, Bn}, performing a series of momentary actions {X1, X2,…, Xn}. Relatedly, it might be wondered whether what is described as a payoff to B is better described as a potentially unequal distribution of payoffs to some group of momentary agents that may fail to include some or even all of the momentary agents performing {X1, X2,…, Xn}. However things are described, the following interesting questions arise: Are there any rational constraints regarding the discounting of future utility? Are there any rational constraints regarding the adoption of goals that can only be carried out in the future? Are there any rational constraints regarding reconsidering or abandoning previously adopted goals? This chapter will consider these questions and closely related questions that are tied to the following facts: although agents, like actions, are normally at least somewhat temporally extended, they need not fully identify with their past and future selves; moreover, although agents are capable of non-deliberatively following through with goals they have adopted, they are also capable of reconsidering and changing courses.1

2 Discounting Human beings discount future utility. Generally speaking, we give more weight to rewards that will be available soon than to rewards of equal magnitude that will be available in the more distant future. We sometimes even favor smaller rewards over larger later rewards. This makes sense given uncertainty about the future. Since I can’t be sure that I will be alive

DOI: 10.4324/9780429202131-40357

Chrisoula Andreou

tomorrow, it can be quite prudent to frontload rewards. But we seem to discount future utility to an extent that goes beyond compensating for such uncertainty, exhibiting what is sometimes referred to as a ‘bias toward the near.’ Whether such bias is rationally permissible is controversial. Intuitively, it may seem to qualify as simply irrational impatience. After all, insofar as all that is at stake is how my choice will impact my well-being, and the rewards at issue will accrue to me, it seems like all that matters is the total impact of the rewards on my well-being. Insofar as the there are no unaccounted for costs or risks, taking a smaller reward just because it is sooner seems like a mistake. Consider, however, the following possibility. Perhaps, what matters, at least primarily and at least to me now, is not how I do, where the relevant I includes all my past and future selves, but how I* do, where I* includes all the selves in the set of my past and future selves with whom I* strongly identify. I* am presumably temporally extended, but my* extent may be quite limited, and my* well-being not at all impacted by the experiences of my* distant future selves. Am I* rationally required to care about these others and to act to benefit them? And, even if total indifference is not in order, must I* give their well-being as much weight as my* own? According to one prominent line of thought, the answers to these questions are ‘yes’ because the mere fact that a self is temporally distant from my current self does not make that self less worthy of concern―time bias is just a form of arbitrary partiality. Interestingly, similar reasoning has been used to argue that being purely, or even just primarily, self-interested is a form of arbitrary partiality. After all, if the mere fact that a self is temporally distant from my current self does not make that self less worthy of concern, why would the mere fact that a self is other than myself make that self less worthy of concern? As Thomas Nagel casts the matter in his classic defense of the possibility of altruism, the conception underlying altruism is that of oneself as merely one person among others, and of others as persons in just as full a sense. This is parallel to the central element in a conception of oneself as temporally extended: that the present is just a time among others, and that other times are equally real. (1970: 88) The idea that it is rationally impermissible to be at all partial to oneself, discounting the utility of others, is quite striking. Perhaps even more striking is the idea that it is rationally impermissible to discount the utility or disutility not only of one’s future selves, but also of one’s past selves.2 Intuitively, it seems perfectly rational to favor having rewards ahead of one rather than behind one. Relatedly, it seems perfectly rational to favor having burdens behind one rather than ahead of one. Consider Derek Parfit’s famous operation case: I am in some hospital, to have some kind of surgery. Since this is completely safe, and always successful, I have no fears about the effects. The surgery may be brief, or it may instead take a long time. Because I have to co-operate with the surgeon, I cannot have anaesthetics. I have had this surgery once before, and I can remember how painful it is. Under a new policy, because the operation is so painful, patients are now afterwards made to forget it. Some drug removes their memories of the last few hours. I have just woken up. … I ask my nurse if it has been decided when my operation is to be, and how long it must take. She says that she knows the facts about both me and another patient, but that she cannot remember which facts apply to whom. She can tell me only that the following is true. I may be the patient who had his operation yesterday. 358

Agency, time, and rationality

In that case, my operation was the longest ever performed, lasting ten hours. I may instead be the patient who is to have a short operation later today. It is either true that I did suffer for ten hours, or true that I shall suffer for one hour. I ask the nurse to find out which is true. While she is away, it is clear to me which I prefer to be true. If I learn that the first is true, I shall be greatly relieved. (1984: 165–166) My attitude seems perfectly understandable, and yet it involves discounting past disutility and exhibiting a form of time bias. It might be suggested that this form of time bias makes sense for agents, since it makes sense to be more concerned about the future, which is still open, than about the past, which we cannot change. But this point seems to rest on equivocation. It is certainly misguided to be preoccupied with the past in a way that rests on the illusion that one can change it, but it does not follow that it makes sense to discount past (dis)utility.3 Perhaps there really is a blanket prohibition against partiality to only certain selves. Like the view that partiality toward oneself is impermissible, the view that discounting past utility is impermissible because it is a form of partiality deviates from the dominant view that rationality is purely instrumental. Roughly put, according to this view, rationality does not tell agents what to care about, but only how to proceed given what they care about. It thus allows for all sorts of partiality, including partiality to one’s current self. Even among those who accept the rational permissibility of partiality to certain selves and, relatedly, of discounting future utility, not all forms of discounting are seen as equal. In models of rational choice, discounting is generally assumed to be exponential, with the agent’s discount curve reflecting a steady discount rate (e.g., x% per day). Discounting can also, however, be hyperbolic, where hyperbolic curves are more bowed than exponential curves.4 Hyperbolic discounting can cause an agent’s rankings of her options to be inconsistent over time. Consider Figure 32.1 (adapted from Ainslie 2001), where the two (roughly drafted) curves represent the agent’s discounting-influenced valuing of two rewards (one small and the other larger and later) as a function of time. In the scenario represented in this figure, when time t and t* are still a ways off into the future, the agent values holding out for the larger later reward over taking the smaller sooner reward, as is indicated by the fact that, in the left half of the figure, the curve representing the value assigned to the larger later

ASSIGNED VALUE

TIM E

t

t*

Figure 32.1 Two (roughly drafted) discount curves from two rewards, one a small reward with gratification occurring at t (if that reward is selected) and the other larger later reward with gratification occurring at t* (if the earlier reward is passed up).

359

Chrisoula Andreou

reward falls above the curve representing the value assigned to the sooner smaller reward. When, however, t is imminent, the value assigned to the smaller sooner reward spikes up and the curves cross; the agent thus reverses her ranking of the available options and ranks taking the sooner smaller reward above holding out for the larger later reward. This sort of ranking reversal is often generated by hyperbolic discount functions. The same is not true of exponential discount functions. These functions can prompt an agent to rank taking a small reward above holding out for a larger later one; but this ranking will be consistent over time, with the discount curves never crossing one another. See, for example, the (roughly drafted) curves in Figure 32.2. This difference between exponential discount functions and hyperbolic discount functions has received a great deal of attention. It might, however, be wondered why ranking reversals are supposed to be so significant. After all, given that any sort of bias toward the near involves some fragmentation of the self, why, if exponential discounting is accepted as permissible, should the fragmentation hyperbolic discounting generates be frowned upon? One major concern is that an agent whose hyperbolic discounting generates ranking reversals has reason (at least if he will not be dying shortly) to be wary of his hyperbolic discounting, at least relative to the concerns of his current self. For, suppose he faces a series of choices between two options, a sooner smaller reward and a larger later reward. The first choice is between SS1 and LL1, the second choice is between SS2 and LL2, the third choice is between SS3 and LL3, and so forth. Now suppose he currently prefers that he opts for immediate gratification in the first case, but that he holds out for the larger later reward in the remaining cases. It may be that he can be quite confident that hyperbolic discounting will work against most of his current rankings regarding his future options because it will prompt him to opt for immediate gratification in every case, something that he currently, and perhaps at all times, sees as undesirable. An exponential discounter has no reason to expect this sort of frustration of his current priorities for the future. For, an exponential discounter’s rankings will not be reversed by discounting: whether he favors frontloading a reward or showing patience for the sake of a larger later reward, his discounting function will not work against his current priorities for the future being realized. It might be concluded that exponential discounting is instrumentally rational, though hyperbolic discounting is not. Of course, if being instrumentally rational is not all there is to being rational, then, even if the preceding conclusion is correct, there is still room for the view that all forms of discounting generated by partiality are rationally impermissible because they are prohibited by some non-instrumental requirement(s) of rationality.

ASSIGNED VALUE

TIME

Figure 32.2

t

t*

Two (roughly drafted) discount curves from two rewards, one a small reward with gratification occurring at t (if that reward is selected) and the other larger later reward with gratification occurring at t* (if the earlier reward is passed up).

360

Agency, time, and rationality

3 Planning, reconsideration, and resoluteness Like the question of whether there are any rational constraints on the discounting of future utility, the question of whether there are any rational constraints regarding the adoption of goals that can only be carried out in the future is tied to possibility of agents being fragmented over time. According to the dominant view in the literature, rational planning involves ‘sophistication,’ wherein an agent’s plans (1) take into account the possibility that her rankings over different options may change over time and (2) seek to bring about what is best, relative to her current rankings, while accepting that her future self will, rationally, be governed by the rankings she has then. Consider the following situation: an agent must choose between going to a party or staying home. Once at the party, she will face the choice between consuming alcohol and not consuming alcohol. Suppose that she currently ranks going to the party and not consuming alcohol above going to the party and consuming alcohol. Relatedly, while she ranks going to the party and not consuming alcohol above staying home, she ranks staying home above going to the party and consuming alcohol. Suppose also, however, that she correctly anticipates that, if she goes to the party and considers her options, her rankings will change and she will rank consuming alcohol at the party above not consuming alcohol at the party. If the agent is a sophisticated planner, she will plan on staying home even though she prefers going to the party and not consuming alcohol. For, being sophisticated, she will assume that if she goes to the party, she will consume alcohol (in accordance with her updated ranking), and conclude that the best she can do now, given her current rankings and given what her future self would rationally do if she were to go to the party, is stay home. Indeed, it might be argued that if the agent is bound to consume alcohol if she goes to the party, then, despite initial appearances, going to the party and not consuming alcohol is not a genuine option for her at the time the question of whether to go to the party arises.5 Sophistication is closely related to Michael Bratman’s ‘linking principle,’ which is conveniently glossed by Richard Holton as follows: ‘I should not form an intention that I now believe I should, at the time of action, rationally revise’ (2009: 154).6 But, unlike sophistication, the linking principle does not assume that rational agents anticipate invariably choosing in accordance with their rankings at the time of choice. As Holton emphasizes, what exactly the linking principle allows for depends on how exactly it is interpreted. According to Holton (2009: 154), the principle is ambiguous between the following: Strong Link: I should not form an intention that I now believe that if I were, at the time of action, to reconsider, I should rationally revise. Weak Link: I should not form an intention that I now believe I should, at the time of action, rationally reconsider and revise. Interpreted in the first way, the linking principle leaves room for the following possibility: Based on the idea that it is sometimes rational to act contrary to my rankings at the time of choice, I permissibly form an intention that I believe will require me to act contrary to my rankings at the time of choice. Interpreted in the second way, the linking principle leaves room for the following possibility: even though I think it is irrational to act contrary to my rankings at the time of choice, I permissibly form an intention that I believe would require me to act contrary to my rankings at the time of choice if I rationally reconsidered, but that will not require me to act contrary to my rankings at the time of choice because I should not reconsider my intention and form any current rankings; instead, I should non-deliberatively follow through on my intention. 361

Chrisoula Andreou

While Bratman favors (something in the ballpark of ) the first interpretation of the linking principle, Holton favors (something in the ballpark of ) the second. What figures as common ground between them is the idea that, particularly in cases of temptation, it is simplistic to think that being rational involves having and acting on a current/up-to-date ranking of one’s options without heeding any non-current ranking of one’s options. Consider again the case of the agent who has to decide between staying home or going to a party. The agent ranks going to the party and not consuming alcohol above going to the party and consuming alcohol, but anticipates that, if she attends the party and considers her options, her ranking will reverse and she will rank consuming alcohol at the party above not consuming alcohol at the party. According to both Bratman and Holton, it does not follow that it is irrational to form the intention to go to the party and not consume alcohol. To the contrary, both Bratman (1999, 2014) and Holton (2009) allow that there is room for a rational agent to form this intention and follow through. For Bratman, this is because rationality can speak against an agent’s acting in accordance with her current ranking, particularly if she anticipates that it will not be enduring, and so, if she acts in accordance with her current ranking, she will regret it.7 For Holton, this is because rationality can, and in cases of temptation normally will, require an agent to refuse to reconsider her options once the anticipated temptation is at hand. At this point, it should be clear that the question of whether there are any rational constraints regarding the adoption of goals that can only be carried out in the future is closely connected to the question of whether there are any rational constraints regarding reconsidering or abandoning previously adopted goals. Let’s turn to considering this question head-on, focusing first on reconsidering previously adopted goals. Consider first a case in which one has the opportunity to reconsider a previously adopted goal, but reconsidering would be costly, and one has no reason to believe that one will change one’s mind. Suppose, for example, that I am at a restaurant, and, very hungry, I am looking forward to having some food arrive at my table soon. I’ve had plenty of time to look at the menu and settled on Pad Thai. The waiter is now at the table and I can order or I can turn him away and reconsider. If I have no reason to believe that I would change my mind if I reconsidered, it seems clearly irrational to incur the cost of doing so (namely, the extra wait for food). Now consider a case at the other end of the spectrum, in which reconsidering will not cost me anything, and I have good reason to believe that, if I reconsider, I will change my mind. Suppose, for example, that I am at a restaurant, but, in this case, the waiter has not yet arrived at my table, and, though I settled on Pad Thai a while ago, I remember that they recently added a new menu item that a friend encouraged me to try. I have nothing else to do, and I expect that, if I reopen the question of what to have and locate the new item on the menu, I will probably opt for that instead. In this case, it seems quite clear that reconsidering is rational. Based on these two cases, it might be concluded that reasonable policies and/or habits of reconsideration must be sensitive to the costs of reconsidering, which count against reconsidering, and the possibility that one would change one’s mind, which counts in favor of reconsidering. But things are complicated. As suggested above, according to Holton (2009), in cases of temptation, rationality can require one to resist reconsidering one’s options, even if one has good reason to believe that reconsideration would result in a reevaluation of one’s options and a change of mind, and even if reconsidering will not cost one anything (in part because one has nothing else to do and one doesn’t mind the activity of reconsidering). Indeed, according to Holton, the point of planning in the face of anticipated temptation may 362

Agency, time, and rationality

be precisely to avert one’s acting on the reevaluation and ranking reversal that one expects would be generated by deliberation in the face of temptation. But, given that instrumental rationality is supposed to take agents’ values as given, why would it prioritize the values of an earlier time slice and disallow reconsideration? Is there, in cases of temptation, something about the agent’s earlier values that truly capture the agent’s standpoint? Even if this is so, rationality need not prohibit reconsideration. It can simply require the agent to deliberate from the values that capture his true standpoint. This brings us to Bratman’s suggestion that an agent that experiences a ranking reversal in the face of temptation need not resist reconsideration. She can instead re-endorse her original plan by recognizing that, if she deviates and acts in accordance with her current ranking, she will regret it. Recognizing her current ranking as temporary and herself as temporally extended, she need not see her current ranking as representing her standpoint. Rationality can thus advise her to choose against her current ranking (or at least change it and then act accordingly).8 We have now ventured into debate about whether there are any rational constraints regarding abandoning previously adopted goals. The big rift in this debate is between those who think that it is rational to abandon a previously adopted goal whenever one’s concerns would be best served by doing so, and those who think that rationality sometimes calls for ‘constrained maximization,’ wherein one constrains one’s maximizing to options that are compatible with one’s prior plans, despite the fact that doing so requires acting contrary to one’s current rankings.9 In this debate, the focus is less on cases of temptation and more on ‘autonomous benefit cases’ in which an agent can benefit from forming an intention that it would be costly for her to carry out. Consider Gregory Kavka’s ‘toxin puzzle’: [A]n eccentric billionaire…places before you a vial of toxin….[You are provided with the following information:] If you drink [the toxin], [it] will make you painfully ill for a day, but will not threaten your life or have any lasting effects….The billionaire will pay you one million dollars tomorrow morning if, at midnight tonight, you intend to drink the toxin tomorrow afternoon….You need not drink the toxin to receive the money; in fact, the money will already be in your bank account hours before the time for drinking it arrives, if you succeed….[The] arrangement of…external incentives is ruled out, as are such alternative gimmicks as hiring a hypnotist to implant the intention… (Kavka 1983: 33–34) If rationality calls for ‘straightforward maximization,’ which requires that one invariably seek to do what best serves one’s concerns, even if this requires abandoning a prior plan, then rationality will figure as an obstacle to gaining the million in the toxin case. For, knowing that he will not drink the toxin if doing so is a pure cost, as it will be when the time for drinking arrives, the rational agent will not be able to form the intention to drink the toxin and so cannot gain the million. If, however, rationality calls for sticking to prior plans that one benefited from forming, even if this requires constraining oneself and not acting in accordance with one’s current rankings, then the rational agent will be able to form the intention to drink the toxin and so will be able to gain the million. Based on the idea that rationality should not be a curse in autonomous benefit cases, some have concluded that rationality sometimes calls for constrained maximization.10 In light of cases of a different sort, it has been suggested that rationality requires at least some kind of resoluteness, since exhibiting resolve is sometimes necessary for avoiding patterns of choice that are self-defeating relative to one’s enduring rankings. Consider the following case. One consistently ranks hot tea above warm tea, but finds both hot tea and warm 363

Chrisoula Andreou

tea incommensurable with iced coffee.11 Both hot tea and iced coffee are currently available in the break room. Though one is torn, one passes up a hot tea for an iced coffee. Soon after walking away, one sees that the wait for both beverages is now long enough that those toward the end of the line who are waiting for tea will be getting warm tea rather than hot tea. Still torn between tea and coffee, one gets back in line and swaps one’s iced coffee for a warm tea. Though understandable, one’s vacillation is self-defeating relative to one’s stable ranking of hot tea above warm tea, and is thus arguably impermissible.12 According to Bratman (2012), the impermissibility of vacillating in cases of incommensurability is not restricted to cases involving a self-defeating pattern of behavior, at least for agents with an interest in self-governance over time (where such governance is understood as guided by a self that is sufficiently unified over time). To facilitate self-governance over time, planning must generate certain defaults, and this rationalizes a prohibition against ‘brute shuffling.’ In cases of brute shuffling, one settles on a certain alternative, but then changes one’s mind, even though no better option has emerged. For Bratman, even when such vacillation does not lead to a self-defeating pattern of behavior, it undermines the ‘defeasible constancy’ that is central to the sort of identity over time that is needed for self-governance over time.13 Notably, even if this is correct, if an agent is not committed to being the sort of unified self that exhibits the defeasible constancy that is necessary for the form of selfgovernance Bratman is concerned with, brute shuffling might be perfectly permissible.

4 Conclusion In standard cases of action, a temporally extended agent performs an action that must be executed over time. But agents need not be fully unified over time, and when actions take more than a moment, with multiple opportunities for reconsideration, it might be suggested that the agent in play is best modeled as a sequence of momentary agents, who may not fully identify with one another, and who may incur unequally distributed costs and benefits in relation to the action at issue, which can itself be broken down into a series of momentary actions, each performed by a different momentary agent. Whether or not this suggestion is correct, it highlights some interesting complications that raise substantial philosophical questions associated with the following facts: although agents are, like actions, normally at least somewhat temporally extended, they need not fully identify with their past and future selves; moreover, although agents are capable of non-deliberatively following through with goals they have adopted, they are also capable of reconsidering and changing courses. The associated substantial philosophical questions focused on in this chapter include the following: Are there any rational constraints regarding the discounting of future utility? Are there any rational constraints regarding the adoption of goals that can only be carried out in the future? Are there any rational constraints regarding reconsidering or abandoning previously adopted goals? As we have seen, contemporary philosophical debate includes an intriguing array of responses to these questions.

Related topics Diachronic agency; Planning agency.

Notes 1 These facts figure as an important part of the ‘diachronic predicament of temporal agents,’ which is helpfully sketched out in Ferrero (2012). 2 For a relatively recent defense of this view, see, for example, Greene and Sullivan (2015).

364

Agency, time, and rationality 3 For some relevant discussion, see Greene and Sullivan (2015). 4 For extensive discussion of hyperbolic discounting, including discussion of, or related to, the points and figures reviewed in the remainder of this section, see Ainslie (2001). 5 For nuanced discussion concerning identifying an agent’s options, see Portmore (2019, chapters 2 and 3). 6 For the official formulation of Bratman’s linking principle, see Bratman (1999: 64). 7 Bratman’s position in his 1999 piece differs somewhat from his position in his 2014 piece. Roughly put, the change is as follows: in his 1999 piece, Bratman suggests that an agent that anticipates regretting acting in accordance with her current ranking should, other things equal, act contrary to her current ranking; in his 2014 piece, Bratman suggests that an agent that anticipates regretting acting in accordance with her current ranking is, other things equal, under rational pressure to revise her current ranking. 8 See note 7. 9 For a defense of the latter, revisionary view, see, for example, Gauthier (1994). 10 Ibid. 11 In accordance with the conception of incommensurability employed in, for example, Raz (1986) and Broome (2000), the tea and coffee qualify as incommensurable (given the agent’s taste) if they are such that neither is better than the other, nor are they exactly equally good. According to Ruth Chang (2002), alternatives that are incommensurable in this sense need not be incomparable; they may be comparable as ‘on a par.’ 12 For discussion of self-defeating vacillation as one among other forms of ‘self-defeating self-governance,’ see Andreou (2012). 13 For critical discussion of Bratman’s position regarding brute shuffling, see, for example, Ferrero (2012).

Recommended reading Andreou, Chrisoula. (2017). ‘Dynamic Choice,’ in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/archives/spr2017/entries/dynamic-choice/. This entry reviews challenges to rational choice posed by the fact that choices are spread out over time. Schelling, Thomas C. (1984) ‘Ethics, Law, and the Exercise of Self-Command,’ in Thomas C. Schelling (ed.), Choice and Consequence. London: Harvard University Press. This essay explores the idea of self-command or self-management given fragmentation involving “alternating values that are incompatible or uncompromisable” (86).

References Ainslie, George. 2001. Breakdown of Will. Cambridge: Cambridge University Press. Andreou, Chrisoula. 2012. ‘Self-Defeating Self-Governance,’ Philosophical Issues 22: 20–34. Bratman, Michael. 2014. ‘Temptation and the Agent’s Standpoint,’ Inquiry 57: 293–310. ———. 2012. ‘Time, Rationality, and Self-Governance,’ Philosophical Issues 22: 73–88. ———. 1999. ‘Toxin, Temptation, and the Stability of Intention,’ in Faces of Intention. Cambridge: Cambridge University Press. Broome, John. 2000. ‘Incommensurable Values,’ in Well-Being and Morality, ed. Roger Crisp and Brad Hooker. Oxford: Oxford University Press. Chang, Ruth. 2002. ‘The Possibility of Parity,’ Ethics 112: 659–688. Ferrero, Luca. 2012. ‘Diachronic Constraints of Practical Rationality,’ Philosophical Issues 22: 144–164. Gauthier, David. 1994. ‘Assure and Threaten,’ Ethics 194: 690–721. Greene, Preston, and Meghan Sullivan. 2015. ‘Against Time Bias,’ Ethics 125: 947–970. Holton, Richard. 2009. Willing, Wanting, Waiting. Oxford: Clarendon Press. Kavka, G. S. 1983. ‘The Toxin Puzzle,’ Analysis 43: 33–36. Nagel, Thomas. 1970. The Possibility of Altruism. Princeton, NJ: Princeton University Press. Parfit, Derek. 1984. Reasons and Persons. Oxford: Clarendon Press. Portmore, Douglas W. 2019. Opting for the Best. New York: Oxford University Press. Raz, Joseph. 1986. The Morality of Freedom. Oxford: Clarendon Press.

365

33 ARTIFICIAL AND MACHINE AGENCY Richmond H. Thomason and John Horty

1 Introduction What is an AI agent? In the 1990s, a shift from AI applications dedicated to narrow expertise at a single reasoning task (such as chess-playing) to more flexible problem-solving applications, embedded in a changing environment, led to a shift in terminology: ‘expert system’ gave way to ‘intelligent agent’.1 The term ‘agent’ is also used in AI for a participant in a multi-agent system. Essentially, these are game-theoretic agents, capable of managing simple communications protocols. Setting such usages aside, we will concentrate on intelligent agents: the paradigm example is a multi-purpose, autonomous robot. What is an agent architecture? It is a functional division of a cognitive system. Even moderately complex agents will have different cognitive functions to perform, and specialized subsystems dedicated to these functions. Some subsystems are devoted to perception and to motor systems; others may be devoted to long-term memory, to symbolic reasoning, to language, and to other specialized functions of inner cognition; others may serve as interfaces. According to one tradition, associated with Herbert Simon and Allen Newell, the same architecture can be physically realized in different ways, and there is no reason in principle why a biological and an electronic infrastructure shouldn’t realize the same architecture. Accordingly, this tradition doesn’t try to distinguish computer science from cognitive psychology. Laird et al. (2017) provide a useful overview of this field. It would be a misconception to expect an agent architecture to be closely coupled with its biological or electronic realizations. An architecture is a division of the operations a mind needs to perform. Although, for instance, the human brain does appear to divide into neurological subsystems dedicated to specific functions, nothing prevents a realization of an architecture from devoting the same computational resources to multiple cognitive purposes, just as the partition of a high-level computer program into subroutines may be lost when the program is compiled into machine language. Agent architectures are abstractions from their physical realizations that help us to understand and design the mind of an embodied agent. Cognitive modules can be classified as peripheral or central, depending on how closely coupled they are with the environment. Perceptual receptors and motor actuators are peripheral.

366

DOI: 10.4324/9780429202131-41

Artificial and machine agency

Deliberate reasoning, including matching of means to ends and producing and evaluating alternative scenarios, is central. And of course there are intermediate cases. Peripheral systems can’t be designed or understood without taking the environment into consideration and can’t function efficiently unless they are partially decoupled from central cognition. A motor actuator, for instance, must respond to tactile sensations automatically and efficiently, but—sometimes at least—needs to be influenced by high-level goals. Digestion is a complex task, but not at all deliberate (and this example reminds us that the environment can be internal). Although breathing can be deliberate, it must be capable of automaticity in a viable biological agent. The pressures and joint angles required for digital manipulation are automatic, even if the task—say, cracking an egg into a bowl—is under the overall direction of central cognition. Learning a motor task—tying one’s shoes, for instance—is largely a matter of exchanging automatic for deliberate control. Because of these considerations, agent architectures engage a broad spectrum of approaches and methodologies, all of philosophical interest. Here, however, we concentrate on planning and problem-solving. The earliest work on AI architectures was exclusively concerned with these tasks, only broadening when applications in robotics became feasible. As we will see, this AI tradition intersects closely with philosophical work on practical reasoning and agency.

2 Matching actions to goals Planning, in the form of means-end reasoning, was one of the earliest research areas to emerge in AI, accompanied by a parallel and overlapping tradition in cognitive psychology; see Newell and Simon (1963), Miller et al. (1960). Actions induce changes: they lead from one possible state of affairs to another. A potential plan is a sequence of actions leading from an initial state to a goal state, and planning is a matter of searching the space of these potentialities. A conception of reasoning as search was profoundly influential on early work in AI.2 Planning problems can be formalized in logical terms: this provides an innovative and early example of the use of logic in AI. The logic of planning uses models very like K ripke’s relational modals for modal logic, which are familiar to most philosophers. In planning models, possible worlds are instantaneous states and the dynamics is driven by actions. In the single-agent case, an action relates a source state to an outcome state—the state that is produced by performing the action. To take a very simple example, imagine a lower-case typing agent named Archie. An instantaneous state consists of a (possibly empty) sequence of letters and a position in the sequence; the action ‘a’ adds the letter ‘a’ to the current sequence and shifts the position forward; similar actions are available for each lower-case key; and there are delete, move-backward, and move-forward actions. Typing a letter adds the letter to the current position; deleting a letter can only be performed if the sequence is nonempty and removes the current letter without changing position; move-forward can be performed unless the position is final; move-backward can be performed unless the position is initial. If Archie is looking at the ‘s’ in ‘yes’ and wants to replace the word with ‘say’, there are many ways to go about doing this. Archie’s planning problem is to find one of these, perhaps move-backward; move-backward; delete; delete; move-forward; ‘a’; ‘y’. Planning, then, is a matter of finding a model meeting certain constraints—a task that can be specified in terms of logic, although it is not a matter of theorem proving. 367

Richmond H. Thomason and John Horty

This example illustrates the fact that actions have preconditions, so that a planner must ensure that the state in which an action is performed is appropriate. Archie wishes to delete the initial ‘y’ in ‘yes’; but a letter cannot be deleted unless it occupies the scanned position. To achieve the subgoal of deleting the ‘y’, Archie must begin by moving backward twice. Formalizing means-end reasoning is a matter of devising an appropriate language and axiomatizing the causal effects of actions. We will not pursue this matter here; for information on this topic, see Reiter (2001) and Doherty and Kvarnström (2008).

3 Attitudes for a simple planning agent Inquiry about an AI agent begins with a reasoning task and asks about cognitive requirements. What resources and competences must an agent have to perform the task? A minimal planning agent like Archie needs only goals, together with beliefs about current and hypothetical states.3 And it has no need to examine or modify these attitudes. A minimal planner does not interact with its environment, and receives its beliefs and goals from a user. It then produces a plan, which—if it is a good planner—will be more or less optimal. The reasoning task of a planner can be complicated along several dimensions. (1) We can add uncertainty about the current state and about the effects of actions. (2) We can add a dynamic environment—and the environment may contain other agents. Information updates can be managed in many ways; for instance, in the multi-agent case there may be opportunities for communication. (3) We can add sensors and effectors, embedding the agent in a natural or simulated environment, and calling for belief revision and the capacity for monitoring the effects of actions. (4) We can take into account limited rationality and the need for rough-and-ready plans in real-time planning problems. (5) We can try to provide an agent with mechanisms for forming its own high-level goals. Exploring all these alternatives would take us far beyond the scope of this paper. We will discuss one influential trend, involving modifications (3) and (4), as well as the more problematic issues raised by modification (5).

4 The IRMA architecture and its descendants The Intelligent Resource-bounded Machine Architecture (IRMA), first presented in Bratman et al. (1988), is motivated by the need to construct plans tailored to many different goals, and to execute (or at least schedule) them. This complicates the planning process because it must not only take into account the environment and the causal effects of actions, but commitments formed in completed planning episodes.4 The IRMA architecture deals with this complication by adding future-directed intentions, and is thus classified as a BDI architecture—an agent architecture that is organized around a reasoning agent’s beliefs, desires, and, crucially, intentions. The intentions in a BDI architecture serve as constraints on the formation of new plans. An intention is more than a desired end. It is a way of achieving an end, and so can be represented as a series of actions, or of pairings with actions with times, or more generally, with a partial plan. This is complicated by the fact that plans and actions can be represented at various levels of abstraction—plans that appear at a coarse level of granularity may conflict when refined. AI researchers recognize the issue of how to represent levels of abstraction and select an appropriate level, although it is not very well understood. But see Hobbs (1990).

368

Artificial and machine agency

For a planning agent, a granularity decision involves risk. For instance, a very abstract and partial plan for shopping might look like this: (1) Go to Store (2) Shop (3) Return Home You could make the plan less abstract by filling in the means of transportation and the shopping list, and less partial by filling in a route to the store, an itinerary in the store, and a route home. There is a trade-off here. If we overplan—if plans are too detailed—we’ll constantly have to undo and revise them during the execution process. If we underplan and forge ahead, hoping for the best, we run the risk of finding ourselves stymied. For instance, we could arrive at the checkout counter with no wallet. Managing this trade-off is a matter of judgment, involving the risks and dangers as well as the degree of uncertainty and dynamicism in the environment. But it is also a matter of personality: a bold agent will discount risk and forge ahead, trusting that things will work out. A cautious agent may want to plan every detail, and may well produce a plan that will have to be discarded as soon as the execution phase is well underway. An IRMA agent needs to manage goals, beliefs, and intentions, and to mobilize these resources in plan formation. Its reasoning task is challenging because detecting inconsistency in a collection of plans is computationally difficult. This touches on the limited rationality theme that is addressed in Bratman et al. (1988) and that Bratman (1987) offers as a rationale for the utility of intentions in practical reasoning. Accommodating the needs of limited cognitive capacity is inseparable from the motivation and articulation of planning architectures because an unlimited mind would have no need to use previous intentions as guides to action. Future-directed intentions endow a planning agent with a useful sort of inertia. A plan enables an agent to get through its day without having to constantly consult its goals and form entirely new plans based on current information. This provides obvious social benefits in the form of reliability, but Michael Bratman (1987), as well as many AI researchers, believes that it also has epistemic and practical advantages for a resource-limited agent. Bratman’s arguments need to be qualified in light of later developments in robotics that make use of machine learning. An autonomous vehicle, for instance, calculates its actions on the fly by a process of continuous real-time planning, with minimal intentions and with mechanical systems providing the only source of inertia. This is made possible by vast amounts of training data. Of course, no one would claim that such data are available in all circumstances. Nevertheless, it seems that intentions are valuable only for some combinations of cognitive resources with tasks and environments. It is not well understood how to characterize these combinations. Although limited rationality is of considerable philosophical interest, we will have little to say about this topic here; but see Russell and Wefald (1991 and Gigerenzer and Selten (2002) for more information. The ideas presented in Bratman et al. (1998) led not only to increased philosophical interest in intentions, for instance in Bratman’s subsequent work, but in AI. See especially Cohen and Levesque (1990). This contrasts with a philosophical tradition influenced by Anscombe (1958), which concentrates on intentional action at the expense of future-directed intentions.

369

Richmond H. Thomason and John Horty

Only a very simplistic agent would automatically discard a proposed plan on discovering a conflict with current intentions. What to do in such cases has to do with priorities between the goals responsible for the conflict and opportunities for eliminating conflicts by revision of the responsible plans. This sort of reasoning could in principle be done using expected utilities, but these are not available in the spectrum of cases that gave rise to the IRMA architecture—general purpose, autonomous agents capable of sensing and acting in a dynamic and uncertain environment. So the use of global decision theoretic methods has not been vigorously pursued by AI researchers in this context, though of course such methods might be useful in localized special cases. Again, different styles of reasoning about intention revision can be identified, corresponding to personality types. A stubborn agent will tend to favor preexisting intentions. An opportunistic, or worse, a capricious agent will readily discard them when presented with a novel opportunity. Considered as a high-level account of practical reasoning, IRMA postulates three data sources—beliefs, desires, and intentions. Beliefs and desires inform means-end reasoning by providing goals and ensuring the feasibility of plans. Previous intentions, as well as desires in the form of preferences, guide plan selection and intention formation. Four principal reasoning modules manage these tasks. The means-end reasoner constructs causally coherent scenarios. The proposal filter checks these scenarios for coherence with standing intentions. The opportunity analyzer and the deliberator allow an agent to modify its standing intentions if new information provides a sufficiently attractive alternative. The following flowchart shows the interactions between data sources and reasoning services, with arrows indicating flow of information. The diagram shows, for instance, that means-end reasoning draws on beliefs, and that the deliberator manages intentions—adding new ones and perhaps removing old ones. Some reasoning mechanisms are not shown: for instance, the ones that revise beliefs. It is assumed that these processes are influenced by perception, and that intentions produce actions, but the diagram doesn’t show these linkages. To be useful in this scheme, desires must take the form of preferences over alternative plans. Bratman et al. (1988) do not explore how these preferences might be produced; inspired by human neuropsychology, Shanahan (2006) describes how a neural network might provide this service. As mentioned earlier, the IRMA architecture was originally inspired by Bratman’s proposal, developed in Bratman (1987), that for a resource-bounded agent, practical reasoning would best proceed against a background set of intentions. The idea was that these intentions,

Figure 33.1

Interactions between data sources and reasoning services.

370

Artificial and machine agency

or already-constructed plans, could be used to filter options from consideration—if an option conflicts with a current intention, a rational agent could then simply reject it outright, without devoting limited resources to considering whether the option is worth pursuing. Although this seems like a promising idea, it is not clear that it is the only or even the optimal strategy for a limited-capacity planning agent. And even if the IRMA strategy is optimal, there is still the problem of fine-tuning the filter: an agent that rejects too many options might miss important opportunities, while an agent that fails to reject enough options loses the computational advantage supposed to derive from a background set of intentions. One advantage of the IRMA architecture is that it allows us to address these issues experimentally, by exploring different filtering mechanisms in agents designed to work in simplified environments. Initial work along these lines was pursued by Pollack in Pollack and Ringuette (1990), which focused on a simple path-planning agent that could earn points by arriving at certain squares in a grid, while, at the same time, other opportunities opened up for earning points. This work supported the conclusion that a robust filtering mechanism was indeed helpful: an agent that pursued its original plans steadfastly, in spite of new opportunities, earned more points than an agent that was prone to abandon plans to pursue new opportunities. Of course, different parameter settings led to different results: if the new opportunities could be extremely valuable, for example, it was worth abandoning ongoing plans to pursue them. Pollack’s work on experimental evaluation of intention-based filtering strategies is summarized and related to the philosophical background in Pollack (1992); it is developed in a richer environment in Horty and Pollack (2000). A more general discussion of plan, or intention, ‘management’ can be found in Pollack and Horty (1999). A comprehensive presentation of a BDI architecture can be found in Wooldridge (2000). Wooldridge fills in details about how the architecture might fit into an embodied agent, equipped with sensors and effectors. Instead of flowcharts, Wooldridge relies on program sketches (or ‘pseudocode’) to present the ideas. This format is more informative, and even readers who have never programmed may find it more useful. Woodridge’s agents perform an endless series of perception-deliberation-action routines. He pays special attention to the issue of how distractible an agent should be—how willing it should be to seek new information, and how ready to modify its plan-based inertia.

5  Reasoned desires Figure 33.1 reveals an important difference between intention on the one hand, and beliefs and desires on the other. Intentions are influenced by deliberation, while beliefs and desires are not. Whatever mechanisms serve to obtain and revise both sorts of attitudes must be independent of practical deliberation in a BDI architecture. But in fact they are not. Take belief. There is a difference between prima facie and all-things-considered belief. Abductive inference, or reasoning to the best explanation, is a common source of belief. Typically, more than one explanation is available, and it can be a nontrivial matter to decide which is best. And frequently we are confronted with conflicting evidence: some things count in favor of a certain belief, and others against it. Also intentions, among other things, are a source of beliefs about the future. Even though intention-based beliefs are fragile in special ways, because the mechanisms for relinquishing intentions are quite different from those for relinquishing beliefs, they can serve as a source of other beliefs. Of course, even setting aside probabilistic approaches to belief, there is a very extensive literature on the rational mechanisms of belief formation. Much less has been said about 371

Richmond H. Thomason and John Horty

desires, despite striking similarities between the two. There is a familiar difference between prima facie and all-things-considered desires. Just about any nontrivial decision will involve weighing pros and cons. Like sifting conflicting reasons for believing, this is a matter of sifting conflicts between conflicting prima facie desires. And, while abduction is a matter of producing best explanations from evidence-based beliefs, planning is a matter of producing best intentions for action from desire-based goals. Finally, beliefs should have a profound effect on all-things-considered desires, because goals must be realistic. These thoughts motivate the modification of the architecture that was proposed in Thomason (2000). This B2D2I architecture incorporates two sorts of beliefs and desires, distinguishing prima facie from all-things-considered attitudes. It uses prioritized defaults (as described, for instance, in Horty 2012) to represent informal arguments that mix beliefs and desires. Crucially, prima facie attitudes can conflict. Arriving at an all-things-considered attitude is a matter of resolving the conflict. The following simple and elaborated examples illustrate conflicting premises, where B ↪ is an epistemic default, D ↪ a bouletic default. In the first example, the agent would like to have a sweet dessert, believes (unqualifiedly) that she would gain weight if she has the dessert, and wouldn’t like to gain weight. (1) Dessert B ↪ gain weight (2) D ↪Dessert (3) D ↪ ¬ gain weight In this formalization, all defaults are prima facie attitudes. All-things-considered attitudes are conclusions reached from premise sets. Unlike classical logic, default logic does not associate a single conclusion set with a set of premises. Conflicting premises can induce alternative conclusion sets, which enforce consistency while maximizing the information obtainable from the premises. Multiple conclusion sets correspond to alternatives that make logical sense; nothing prevents an agent from using other criteria to choose between these alternatives. If no distinction is made between epistemic and bouletic defaults, the first premises generate three conclusion sets: (1) {dessert, gain weight}, (2) {¬gain weight}, (3) {dessert, ¬gain weight}. (1) and (2) (indulging in dessert despite gaining weight, and not gaining weight) are reasonable alternatives. But (3) is not reasonable—it is a case of wishful thinking, of allowing a desired outcome (not gaining weight) to override the belief that having dessert induces weight gain. Wishful thinking looks like a logical defect, and can be corrected by prioritizing epistemic defaults over bouletic defaults, so that conflicts between the two are resolved in favor of beliefs. For details about prioritized defaults, see Horty (2012). Consider now a second premise set that introduces conflicting beliefs. (1) Dessert B ↪ gain weight (2) D ↪Dessert (3) D ↪ ¬ gain weight (4) Dessert B ↪ ¬ gain weight 372

Artificial and machine agency

Perhaps a friend has told the agent that he regularly eats desserts without gaining weight. If the source of Premise 1 is medical research and the source of Premise 4 is personal anecdote, Premise 1 should have greater force. But as long as beliefs are not prioritized according to credibility, these premises restore the conclusion set {dessert, ¬gain weight}. Although this approach to reasoning about desires provides a formalization and a characterization of reasonable conclusion-making, it does not afford a workable reasoning mechanism. There are many ways to go about doing this—using constraint satisfaction, aggregation techniques, or neural nets, for instance—and this remains an open and largely unexplored issue. A B2D2I architecture enlarges the connectivity between practical attitudes: beliefs can affect (all-things-considered) desires; intentions can create desires (through subgoaling) as well as beliefs, and of course (all-things-considered) beliefs and desires can affect intentions. This provides a picture that is more faithful to human practical reasoning. Although similar architectures have been discussed in the AI literature (see, for instance, Broersen et al. 2005), they have not been deployed and tested in challenging settings, and as of now their value in applications such as robotics is uncertain. The practical reasoning tradition in AI has little or nothing to say about how goals originate. The assumption is that even autonomous robots do not develop their own high-level goals, but obtain these from human users. One source of goals, of course, is emotions, and there is a growing literature on AI architecture that incorporates emotions; see Calvo et al. (2015). But almost entirely, these are concerned with simulating emotions for better interactions with human users and for entertainment, rather than with incorporating them in an architecture for practical reasoning. As we have seen, BDI agent architectures were centered around intentions, as well as simple beliefs and desires, while B2 D2I architectures introduce two levels of beliefs and desires. But attitude proliferation does not stop there—AI researchers continue to investigate more complex architectures. Although we cannot explore any of these in detail, we do want to mention the very interesting BOID architecture (Broersen et al. 2001), which supplements the traditional BDI components with a store of agent obligations. The BOID architecture has been used, in particular, in the study of goal formation; see Broersen et al. (2002). Although goal formation turns out to be a difficult problem in AI, we humans may not be much better off than robots in forming high-level goals. These seem to arise more from natural inclination and maturation, and from social influences, rather than from any rational process. Almost all our day-to-day goals emerge from inertia and habit, and are rarely reexamined. The architectures discussed here are not meant to explain the origin of high-level goals, although they do provide a picture of the rational maintenance of desires in everyday cases of practical reasoning.

6 Conclusion Philosophical discussions of agency often involve assumptions about the architectures of rational agents—that is, the functional organization of their psychological components. Philosophers typically presuppose a simple Humean agent architecture, involving only beliefs and desires, although sometimes they entertain more complex arrangements. Almost always, philosophers allow these assumptions to remain implicit. Many philosophers working in rational agency are aware that there is relevant work on human cognitive architecture in the field of cognitive science; see, for instance, Newell (1992) 373

Richmond H. Thomason and John Horty

and Salvucci (2016). But few know of the rich, explicit, and detailed history of work on agent architecture in the field of AI, which is equally relevant to philosophical concerns. This brief, opinionated introduction to the work may, we hope, help to correct this.

Related entries Planning agency; Diachronic agency; Agency and practical reasoning.

Notes 1 Background on the history can be found in Agre (1995) and Nilsson (2010, ch. 31); this shift in focus is now so firmly entrenched that the leading textbook in AI (Russell and Norvig 2009) is organized around the idea of designing an intelligent agent. 2 See Fikes and Nilsson (1971) for an initial formulation of this idea, and Blum and Furst (1997) for an more recent, very sophisticated refinement; other approaches to planning that have been explored in AI, and that are particularly relevant to philosophical work, include least-commitment partial-order planning (Weld 1994) and hierarchical task network planning (Erol et al. 1996). 3 Of course, the agent must also be able to create and verify plans, but our concern with architecture places algorithmic considerations in the background. 4 These commitments differ logically from beliefs about the future because the conditions for revising them differ. Whether or not I intend to stay in my room all morning because I intend to do so or because I’m locked in, I believe I will stay there. But in the first case I can discard the belief (and the intention that supports it) when I discover a conflicting opportunity; perhaps when I learn that I can go for a walk with a friend, whereas in the second case I can’t discard the belief, no matter how attractive the opportunity.

Further reading Bratman, Michael E., David Israel, and Martha Pollack. 1988 “Plans and resource-bounded practical reasoning.” Computational Intelligence, 4: 349–355. Broersen, Jan N., Mehdi Dastani, Joris Hulstijn, Zisheng Huang, and Leendert van der Torre. 2001. “The BOID architecture: Conflicts between beliefs, obligations, intentions, and desires.” In Elisabeth André, Sandip Sen, Claude Frasson, and Jörg P. Müller, editors, Proceedings of the Fifth International Conference on Autonomous Agents (AGENTS-01): 9–16. ACM Press. Pollack, Martha, and John F. Horty. 1999. “There is more to life than making plans: Plan management in dynamic, multi-agent environments.” The AI Magazine, 20(4): 71–84. Thomason, Richmond H. 2000. “Desires and defaults: A framework for planning with inferred goals.” In Anthony G. Cohn, Fausto Giunchiglia, and Bart Selman, editors, KR2000: Principles of Knowledge Representation and Reasoning: 702–713. Morgan Kaufmann, San Francisco. Wooldridge, Michael J. 2000. Reasoning about Rational Agents. Cambridge University Press, Cambridge.

References Agre, Philip E. 1995. “Computational research on interaction and agency.” Artificial Intelligence 72(1– 2): 1–52. Anscombe, G.E.M. 1958. Intention. Blackwell Publishers, Oxford. Blum, Avrim L. and Merrick L. Furst. 1997. “Fast planning through planning graph analysis.” Artificial Intelligence 90 (1–2): 281–300. Bratman, Michael E. 1987. Intentions, Plans and Practical Reason. Harvard University Press. Bratman, Michael E., David Israel, and Martha Pollack. 1988 “Plans and resource-bounded practical reasoning.” Computational Intelligence 4: 349–355. Broersen, Jan N., Mehdi Dastani, Joris Hulstijn, Zisheng Huang, and Leendert van der Torre. 2001. “The BOID architecture: Conflicts between beliefs, obligations, intentions, and desires.” In

374

Artificial and machine agency Elisabeth André, Sandip Sen, Claude Frasson, and Jörg P. Müller, eds. Proceedings of the Fifth International Conference on Autonomous Agents (AGENTS-01): 9–16. ACM Press. Broersen, Jan N., Mehdi Dastani, Joris Hulstijn, and Leendert van der Torre. 2002. “Goal generation in the BOID architecture.” Cognitive Science Quarterly 2 (3–4): 428–447. Broersen, Jan N., Mehdi Dastani, and Leon van der Torre. 2005. “Beliefs, obligations, intentions and desires as components in agent architectures.” International Journal of Intelligent Systems 20(9): 893–919. Calvo, Rafael A., Sidney K. D’Mello, Jonathan Gratch, and Arvid Kappas, eds. 2015. Oxford Handbook of Affective Computing. Oxford University Press, Oxford. Cohen, Philip R., and Hector J. Levesque. 1990. “Intention is choice with commitment.” Artificial Intelligence 42(3): 213–261. Doherty, Patrick, and James Kvarnström. 2008. “Temporal action logics.” In Frank van Harmelen, Vladimir Lifschitz, and Bruce Porter, eds, Handbook of Knowledge Representation: 709–757. Elsevier, Amsterdam. Erol, Kutluhan, James Hendler, and Dana Nau. 1996. “Complexity results for HTN planning.” Annals of Mathematics and Artificial Intelligence 18(1): 69–93. Fikes, Richard E., and Nils J. Nilsson. 1971. “Strips: A new approach to the application of theorem proving to problem solving.” Artificial Intelligence 2(3–4): 189–208. Gigerenzer, Gerd, and Reinhard Selten, editors. 2002. Bounded Rationality: The Adaptive Toolbox. MIT Press, Cambridge, MA. Hobbs, Jerry R. 1990. “Granularity.” In Daniel S. Weld and Johan de Kleer, eds. Qualitative Reasoning about Physical Systems: 542–545. Morgan Kaufmann, San Mateo, CA. Horty, John F. 2012. Reasons as Defaults. Oxford University Press, Oxford. Horty, John F., and Martha E. Pollack. 2000. “Evaluating new options in the context of existing plans.” Artificial Intelligence 127(2): 199–220. Laird, John E., Christian Lebiere, and Paul S. Rosenbloom. 2017. “A standard model of the mind: Toward a common computational framework across artificial intelligence, cognitive science, neuroscience, and robotics.” The AI Magazine 38(4): 13–26. Miller, George A., Eugene Galantner, and Karl H. Pribham. 1960. Plans and the Structure of Behavior. Holt, Rinehart and Winston, New York. Newell, Allen. 1992. Unified Theories of Cognition. Harvard University Press, Cambridge, MA. Newell, Allan, and Herbert A. Simon. 1963. “GPS, a program that simulates human thought.” In Edward Feigenbaum and Julian Feldman, eds. Computers and Thought: 279–293. McGraw-Hill, New York. Nilsson, Nils J. 2010. The Quest for Artificial Intelligence. Cambridge University Press, Cambridge. Pollack, Martha. 1992. “The uses of plans.” Artificial Intelligence 57(1): 43–68. Pollack, Martha, and John F. Horty. 1999. “There is more to life than making plans: Plan management in dynamic, multi-agent environments.” The AI Magazine 20(4): 71–84. Pollack, Martha, and Marc Ringuette. 1990. “Introducing the tileworld: Experimentally evaluating agent architectures.” In Thomas Dietterich and William Swartout, eds. Proceedings of the Eighth National Conference on Artificial Intelligence: 183–189, Menlo Park, California. American Association for Artificial Intelligence, AAAI Press. Reiter, Raymond. 2001. Knowledge in Action: Logical Foundations for Specifying and Implementing Dynamical Systems. The MIT Press, Cambridge, MA. Russell, Stuart, and Peter Norvig. 2009. Artificial Intelligence: A Modern Approach. Prentice Hall, Englewood Cliffs, NJ, 3rd edition. Russell, Stuart, and Eric Wefald. 1991. Do the Right Thing. The MIT Press, Cambridge, MA. Salvucci, Dario D. 2016. “ACT-R and beyond.” In Susan E.F. Chipman, editor, The Oxford Handbook of Cognitive Science: 15–26. Oxford University Press, Oxford. Shanahan, Murray P. 2006. “A cognitive architecture that combines internal simulation with a global workspace.” Consciousness and Cognition 15(2): 433–449. Thomason, Richmond H. 2000. “Desires and defaults: A framework for planning with inferred goals.” In Anthony G. Cohn, Fausto Giunchiglia, and Bart Selman, editors, KR2000: Principles of Knowledge Representation and Reasoning: 702–713. Morgan Kaufmann, San Francisco. Weld, Daniel S. 1994. “An introduction to least commitment planning.” The AI Magazine 14(4): 27–61. Wooldridge, Michael J. 2000. Reasoning about Rational Agents. Cambridge University Press, Cambridge.

375

34 AGENCY AND PERSONAL IDENTITY Marya Schechtman

A close connection is often drawn between agency and personal identity. The guiding idea is the simple and intuitive one that for an action to be agential it must flow from an agent, reflecting who she truly is. This is frequently understood as implying that a person’s identity is made up of the motivations that give rise to the actions that are rightly attributed to her. Since agency is thought to require unity and consistency of motivational profile, this approach yields a view of personal identity that also requires unity and consistency. When we think about what constitutes personal identity independent of questions of agency, however, we are likely to arrive at a conception of identity that allows for more disunity and change. Here I look at the picture of identity that emerges when it is defined in the context of questions of agency, at the somewhat different picture that emerges when we think about identity on its own, and at the implications of the differences between these pictures.

1 There are many places within philosophy where links are drawn between agency and personal identity. Here I offer as examples two contexts in which this link leads to a strongly unified picture of the self. The first context focuses on the question of what makes a movement in someone’s human history an action as opposed to a mere happening in her body. It is one thing to push someone in anger, for instance, and quite another to knock into them after tripping and falling. Pushing is something I do; in falling I am a passive instrument of the laws of physics. The tricky thing is to find a general way to distinguish between what I do and what happens through me. One widely defended approach makes this distinction in terms of reflective endorsement. Those events which flow from motivations I endorse are my actions, on this view, while those that do not are mere occurrences in my history. There are many versions of this approach. Here I sketch the influential expression of this general view offered by Harry Frankfurt to convey the general sense. While it is obvious that movements that result from the action of irresistible physical forces are not actions that I take, Frankfurt points out that it is not only external forces that can defeat our agency. ‘Certain events in the history of a person’s mind, likewise, have their moving principles outside of him,’ he says. ‘He is passive with respect to them, and they 376

DOI: 10.4324/9780429202131-42

Agency and personal identity

are likewise not to be attributed to him’ (Frankfurt 1988a, 61). He gives the example of an unwilling addict who is driven by his addiction to use drugs, despite his heartfelt desire to stop. When we behave compulsively in this way, Frankfurt argues, we lack agency. The desire for the drug the unwilling addict craves is experienced in his psychological history, but it is not his desire. The same is true more generally in cases of weakness of will where we find ourselves acting in ways we do not wish to, for example, leaning on the horn in bad traffic, eating dessert when we had resolved not to, checking email during time that was to be dedicated to writing. Frankfurt theorizes these observations in terms of higher-order endorsement. We experience a great many desires that pull us in different directions, Frankfurt says, but as humans we also have the capacity to take an attitude toward these desires, forming higher-order desires about which first-order desires we wish to act on. At the level of first-order desires there is just the motivational pull of the desires themselves. Once we take a higher-order attitude to these first-order desires, however, we take a stand about what we want to do. This establishes an agent above and beyond the competing pulls of the first-order desires, and a fact about what constitutes the agent’s will. I may have a desire to be a courteous driver, for instance, and a desire to hit my horn and yell expletives at the person in front of me. If I form a higher-order desire to be courteous and repudiate the desire to display my impatience, I have made it the case that I wish to be courteous. If I manage to refrain from honking and yelling, I have done what I wished to do. If my impatience gets the better of me, I have failed to do what I wanted, and an alien impulse has overpowered my will. The connection drawn between agency and identity is clear and explicit. What is internal to me as a person is what I have endorsed. Repudiated desires, even if they occur in my psychological history, are not internal to me as a person. Frankfurt tells us that when someone decides which desires he wishes to act on, the ‘decision determines what the person really wants by making the desire on which he decides fully his own. To this extent the person, in making a decision by which he identifies with a desire, constitutes himself.’ It is, he says, ‘these acts of ordering and rejection – integration and separation – that create a self out of the raw materials of inner life (Frankfurt 1988b, 170). It is crucial to constituting a self, moreover, that higher-order decisions about which desires we wish to move us to action are entirely wholehearted. There is only a fact about which desires are internal to an agent if there is a decisive commitment to some desires and a repudiation of others. If the competition between first-order desires is reproduced at the second-order or above, there remains no fact about where the person stands on this conflict, and so no fact about who someone really is. Frankfurt tells us that ‘ambivalence, like self-deception, is an enemy of truth,’ because a person’s ‘ambivalence stands in the way of there being a certain truth about him at all’ (Frankfurt 1988b, 170). Determining whether a movement in my history is an action requires knowing if its motivation is internal to me and internality is defined in terms of a wholehearted commitment to acting on this desire. What makes me an agent makes me a self and determines my identity. Since agency requires me to be completely unified in my motivational profile, on this view, so does identity. If I have an unsettled or inconsistent motivational profile, I have not established a well-defined personal identity. Frankfurt’s view is far more nuanced than what I have presented here, and it is, as I said, only one example of a more general approach to questions about agency and identity. This brief and selective sketch is sufficient, however, to demonstrate the central idea that personal identity is constituted by the unification of the will via wholehearted, higher-order attitudes. It is important to note that although this approach grows out of reflection on what it takes for an action at a time to belong to an agent (and so on what it is for a particular motivation or 377

Marya Schechtman

set of motivations to be internal to a person at the time of action), it has diachronic implications. Agency requires that the agent be settled on which motivations she wishes to lead her to action, and someone cannot be settled just for the moment. If wholehearted endorsement of a motivation could amount simply to being unconflicted about it right now, the addict could be wholeheartedly committed to getting straight while sitting in his counseling session and just as wholeheartedly committed to the life of an addict the next day. The idea that one’s identity as an agent, and so one’s personal identity, requires motivational stability over time is thus a key part of this approach. The diachronic aspect of this view serves as the starting point for a second context in which agency and identity are often seen as co-extensive. Questions are commonly raised about whether a person who has undergone radical change should be held responsible for actions undertaken in the past. If we look at cases like the criminal who has undergone a jailhouse conversion, the radical involved in violent protests who has become a conservative pillar of the community, or the politician who displayed unacceptable sexist behavior but has since become a champion of women’s rights, there is often a serious question about whether the person before us now is rightly blamed for what he did in the past. Frequently this is described by saying that he is now a different person from the one who took the pernicious actions. Once again, we see the idea that the identity of the person is co-extensive with the set of motivations and actions that can be rightly attributed to her, as well as the idea that stability and unity of motivational profile is required for sameness of person. This idea is expressed, among other places, in psychological continuity theories of personal identity (e.g., Lewis 1983; Parfit 1984; Perry 1975; Shoemaker 1984), which define personal identity in terms of psychological continuity over time. These theories also draw a very close connection between identity and agency. Starting from Locke’s claim that ‘person’ is a ‘forensic term’ (Locke 1975, 346), they suggest that one way to determine whether someone in the present is or is not the same as some past person is to consider whether the present person is rightly held responsible for what the past person did. If the present individual is the same agent as the past one, in other words, she is the same person. There are many different versions of these views, but all suggest that stability of motivational profile is a key part of what makes for sameness of person. The guiding idea is expressed nicely by David Lewis when he says that to capture the practical significance of personal identity what matters is that ‘[m]y total present mental state should be but one momentary stage in a continuing succession of mental states. … Change should be gradual rather than sudden, and (at least in some respects) there should not be too much change overall’ (Lewis 1983, 17). There are, of course, other views of both agency and personal identity. Nonetheless, the prominence of reflective endorsement views like Frankfurt’s and psychological continuity theories like Lewis’ shows that there is an important strain of thought that sees personal identity and agency as two sides of a single coin, and on which both agency and identity require a consistent, unconflicted, and stable motivational profile. Call this the ‘agential identity view.’

2 While there is a great deal of intuitive appeal to the agential identity view, if we think about personal identity independent of agency, the degree of stability and consistency the view requires seems less plausible. According to the agential identity view we must think of personal identity as a diachronic phenomenon, requiring stability over time, but if we look at some of the dynamics of human psychology, we find reason to think that personal identity is

378

Agency and personal identity

messier than this view implies. Here I look at two common elements of human experience that complicate the agential identity view. The first involves recurring, often inconsistent, patterns of thought, behavior, experience, or commitment. These are common in human life. Most people have at least some characteristic fluctuations in mood, affect, and general outlook, and these are closely associated with fluctuations in motivational profile. If someone suffers from recurring depression, for instance, there may be times when she is convivial, joyful, energetic, and optimistic and, predictably, periods where she is lethargic, experiencing feelings of hopelessness and despair. Less dramatically, someone may typically be burned out and ready to change jobs at the end of each semester, but regularly revitalized and ready to go at the beginning of the next. A devout Catholic may struggle periodically with her faith but always find her way back; someone may swing between periods of intense career ambition and periods where it seems as if the sacrifice of time with family and friends is simply not worth it. These dynamics are familiar. According to the agential identity view, someone who experiences these kinds of fluctuations has either failed to establish a well-formed identity or else changes identity frequently. It is natural, however, to think of such patterns of fluctuation as themselves identity defining. Churchill’s ‘black dog’ is a storied part of his personality, and a biographer seeking to provide a sense of who he really is would make a curious choice to omit mention of it. Similarly, biographies of Mother Teresa have reported that she had continuing and serious struggles with her faith. It certainly seems that we have learned something important about her if we learn this. In describing a friend to you I might tell you that she is the kind of person who regularly falls deeply in love after about two days and then just as quickly out of love again. And if we learn that the hard-driving CEO is periodically gripped by the conviction that it is all empty and not worth pursuing, we know something about her that we did not know before. The idea that this kind of changeability is part of what makes someone who she is serves as a challenge to the idea that identity requires a stable and consistent motivational profile. Before considering how defenders of the agential identity view could respond to this challenge, we can reinforce it by considering a second fact about human life that raises similar themes, the fact that people develop and change over time. It is common to see the trajectory and nature of these changes as part of what constitutes someone’s identity. Human lives have a developmental shape; we expect people to grow, learn, and develop. While the agential identity view does not deny this fact and can, to some extent, accommodate it, it sits awkwardly with the kind of diachronic stability of motivational profile the view requires. To see this, it helps to think about the way in which a person’s story gives context to his motivational profile that is relevant to his identity. The identity of Holly Golightly, the protagonist of Breakfast at Tiffany’s, is inextricably bound up with the fact that she was once Lula Mae Barnes, a small-town country girl, who reinvented herself as an Upper East Side sophisticate, and it is central to Jay Gatsby that he started out as James Gatz. Similarly, it is a critical part of the identity of the Apostle Paul that he was a persecutor of Christ’s disciples before he was struck blind and converted on the road to Damascus. On one familiar understanding of personal identity, to know who someone is you need to know her story, how she got to where she is now, and what happened to her along the way. But if the developmental story is part of identity, we have to conclude that Lula Mae survives her reinvention as Holly Golightly, James Gatz his metamorphosis into Jay Gatsby, and Paul his conversion. These suggest not only that a person can persist through the kinds of radical change that the agential identity view sees as identity transforming but, more importantly, that dramatic

379

Marya Schechtman

changes in motivational profile such as those that occur in conversions, trauma, or selfreinvention can be seen as a central part of what makes someone the person she is, and so as identity-constituting. When we are thinking about personal identity outside of the context of agency, it seems natural to view recurring and sometimes inconsistent motivational profiles and radical changes as elements of personal identity rather than as identity undermining. If this is so, however, then the agential identity view’s insistence that personal identity requires a strongly unified motivational profile is called into question.

3 Agential identity theorists could respond to the challenges just described in a variety of ways. One is simply to reject the intuitions on which they are based, insisting that the cases described are not cases in which individuals have managed to constitute or sustain a wellformed personal identity. It is hard to see exactly how this clash of intuitions could be adjudicated, however, and it is at least jarring to be told that larger-than-life figures like Churchill and Mother Teresa are not fully formed as persons with personal identities, or that we best understand Gatsby by forgetting that he was Gatz or the Apostle Paul by ignoring the fact that he was not always a Christian. A more promising approach is to deny that the examples I have offered as challenges are properly described. Consider first the examples of recurring motivational fluctuation. The agential identity view acknowledges that even after someone makes a wholehearted commitment to one first-order desire, competing first-order desires may remain. The important thing is that she has definitively taken sides against them. While Churchill may have experienced recurring depression, a proponent of the agential identity view might say that he probably resisted the motivational profile associated with it in favor of that he experienced when not depressed. Mother Teresa, similarly, probably fought against her doubts, but not against her faith. The vicissitudes in experienced motivational pulls in these cases serves only to reveal the strength and cost of commitment to the endorsed motivational profile, not that there is genuine inconstancy or ambivalence. Cases of radical change might be dealt with in a similar way. What we see in Gatsby, it might be argued, is an unwavering ambition to get away from his humble origins. What makes him such a well-defined character is precisely the motivational profile that remains constant and takes him from the farm in North Dakota to the mansion on Long Island. It is not obvious that this reply will work. First, while the claim that there is unity of profile may be plausible in the cases just mentioned, it will be much less so in others. The CEO who sometimes thinks of chucking it all in favor of a simpler life may be no more inclined to fight her urge to live a less stress-filled and competitive life when she feels it than she is to fight her ambition when that is what she is experiencing. And if there is a story to tell about how Gatsby maintains a unified motivational profile throughout his reinvention, it is much harder to tell a similar story about Paul’s conversion. The more fundamental point is that the intuition behind the challenge to the agential identity view is not just that people can have and maintain identities throughout fluctuations and radical change; it is that the fluctuations and radical change are inherent to their identities. A Mother Teresa with recurring spiritual doubts is a different person from one with unwavering faith, and Gatsby is different from the other denizens of West Egg precisely because he started out so differently. To say that that only one of the sets of conflicting motivations or one temporal portion of the transformed individual represents his true identity while the other is irrelevant does not do full justice to the intuition. 380

Agency and personal identity

When we consider questions of identity independent of questions of agency, we tend to think of people as immensely complicated beings whose identities are constituted by the intricate dynamics of their psychological lives. People are multifaceted and changeable and the more we know about their different facets and about the path they have traveled to get to where they are now, the more we know about who they are. This is why events like weddings or funerals which gather people from different phases or contexts of a person’s life are so often seen as important windows into their identities. Agential identity theorists thus need to take the challenge raised here seriously, and to consider further how they might be addressed.

4 The agential identity view’s argument for a strongly unified picture of personal identity is based on two assumptions. The first is that a person’s identity is constituted by the set of motivations that are internal to her. The second is that agential action requires a strongly unified motivational profile. From these assumptions the conclusion that personal identity also requires a strongly unified motivational profile follows quickly. I have pointed to two phenomena which suggest a familiar way of thinking about personal identity according to which identity inherently involves dynamics among changing and often inconsistent motivations. If the conclusion that personal identity requires a strongly unified motivational profile is rejected, it is necessary to give up or modify at least one of the two assumptions that lead to it. One possibility is to qualify the first assumption. The challenge I have raised suggests that there is a sense of personal identity according to which someone’s identity is constituted, at least in part, by conflicting and changing motivations, and so is neither as consistent nor as stable as the agential identity view suggests. This does not mean that there is not also another sense of personal identity which does require motivational unity. We use the term ‘personal identity’ in many different ways; there is agential identity, social identity, the ‘identity’ of identity politics, and others. Agential identity theorists can thus allow that there might be all kinds of ideas of personal identity with respect to which our conflicts and ambivalence are identity constituting. Their claim, however, concerns agential identity, the notion of identity at work when we are trying to determine whether a movement in a person’s history is an action. In this context, the argument would go, the relevant type of internality does require strong unity of motivational profile, and so does the concept of identity. This is a plausible response. There is no denying that we do use the phrase ‘personal identity’ both in the sense I have described, which rests on complexity and allows multiplicity, and in the way that is evoked in the agential identity view, which requires strong unity, among many others. One might say, for instance, ‘I always thought her faith was unwavering, but it turns out that she wrestled with doubt her whole life. I guess we never really knew her.’ But we might also say, ‘I know your best friend stole the jewelry you inherited from your mother and sold it for drug money, but you have to understand that it was not really she who did it, it was the addiction.’ One way of reconciling the phenomena I described with the arguments of the agential identity view is thus to claim that although they yield inconsistent pictures of personal identity, there is no real conflict between them since they are defining different senses of an ambiguous term. This response modifies the first assumption by qualifying the sense of personal identity it employs. This is a valuable result in itself. The account of identity offered by agential identity theories is often employed beyond the strict bounds of questions about agency, and a clear recognition that this is not warranted by the arguments is instructive. There is 381

Marya Schechtman

reason, however, to consider an even more radical response. This first reply leaves intact the assumption that agency requires strong unity because the notion of personal identity relevant to agency is presumed to be the one the agential identity view develops. It is worth asking, however, whether this is necessarily so, and whether the messier conception of identity that emerges from the challenge I raised to this view might not do better. To be clear, I am not denying that we do have multiple senses of personal identity that apply in different contexts. I am instead asking whether we should take it for granted that the one agential identity theorists offer is the best for the context in which it is employed. Consider the above example which was put forth as a case in which we are using ‘personal identity’ in the sense relevant to the agential context: ‘It was not she who stole your jewelry; it was her addiction.’ We do talk this way, but if we think about our actual practices of attribution and blame, things look more complicated. Even focusing exclusively on forensic concerns, we do not treat this kind of behavior as simply an occurrence in a person’s human history. There is indeed a kind of passivity when someone is driven by a powerful desire of the sort encountered in addiction, and this sort of passivity can mitigate blame or serve as a reason to forgive. Still, stealing from your best friend for drug money is not the same as falling into her when the crowded bus you are both on comes to a sudden stop. In the latter case, a simple ‘sorry’ suffices, while the former calls for a more complicated form of apology. It is, moreover, clearly unwarranted to have my feelings hurt even if I am badly injured when friend falls into me on the bus; it is less evidently unjustified to feel hurt if she sells the only mementos of my deceased mother to obtain drug money. A defender of the agential identity view might insist that the two cases are actually completely parallel, and that the tendency to treat them differently rests on a misunderstanding of the nature of addiction. While this is a question far too complicated to be tackled here, there are reasons to resist this analysis. Recall that according to the agential identity view, it is not only a resolute rejection of the impulse to steal that prevents it from being an action of the addict; deep ambivalence also prevents agency. Actions that flow from an impulse about which someone is ambivalent speak only for the first-order desire that motivates them and have no authority to speak for the person. In the kind of case I am imagining where a friend struggling with addiction is conflicted, wanting both to be a loyal friend but also to procure money for drugs, there is no fact of the matter about what she wants, and her theft is not a true action. The call for stability of motivational profile means, moreover, that if my friend ultimately overcomes her addiction, the sober person is not the same person who stole from me. These results are strange and inconsistent with practice. My friend’s addiction may certainly call for understanding and support, but it is also reasonable to be hurt and angry if she takes my precious mementos and sells them, even if she is conflicted about it. And at the happy time that she gets sober, surely it is appropriate for her to take responsibility for the pain she caused me during that time, rather than to simply experience relief that it was not, after all, she who did this to me. Acknowledging what one has done is typically taken to be an essential step to repairing relationships damaged by addiction. It is possible for the agential identity view to accommodate these observations by saying that the question of when blame should be assessed and the question of whether an action is truly attributable to someone are not the same. We might have reasons for blaming people for actions that are not properly ascribed to them as agents, as well as to absolve them of blame for actions that are. Addiction, in fact, may provide an example of both kinds of

382

Agency and personal identity

divergence. Insofar as this view seeks only to pronounce on the question of agency and not of blame, it might be argued, the concerns I raise do not apply to them. There is, however, another possibility which arguably captures the phenomena better. This is to maintain the three-way connection between agency, identity, and blame, but to allow that all three are complex and messy. In the case of my friend’s addiction, perhaps it is most accurate not to give a simple ‘yes’ or ‘no’ answer to the question of whether the desire to betray me is part of who she is, or is something she did as opposed to something that happened to her, or whether she is rightly blamed for what she did. Answering any of these questions will require a longer explanation, one which refers to her degree of conflict and to the history that led her to her addiction and addictive behavior. The messiness that constitutes identity in the sense I described earlier is going to be directly relevant to questions of agency and blame on this view. Put differently, the strict unity called for in the agential identity view is required only if we think that attribution and agency must be all- or-nothing, but there is reason to think that they are not, and so reason to consider a looser and more complicated understanding of both agency and identity. We have considered one prominent strand of thought about agency and identity, which argues that both require strong consistency and stability in motivational profile. Reflection on a view of personal identity that involves inconsistency and change suggests that, minimally, we will need to qualify the type of identity described in this approach or, more radically, to rethink the idea of agency it presupposes. This latter suggestion requires much more development, but it represents an intriguing direction of research.

Related topics Agency and responsibility; Agency and self-knowledge; Agency and autonomy; Agency, narrative, and mortality; Agency and identification; Agency, time, and rationality.

Further reading Korsgaard, C. M. 1989. Personal Identity and the Unity of Agency: A Kantian Response to Parfit. Philosophy and Public Affairs, 18(2), 103–131. Korsgaard is one of the central defenders of the agential identity view, who offers a slightly different version and defense of the view than Frankfurt does. Korsgaard, C. M. 2009. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press. This book offers a detailed argument for Korsgaard’s mature view of the nature of agential unity and provides a strong example of the agential identity view. Marino, P. 2011. ‘Ambivalence, Valuational Inconsistency, and the Divided Self,’ Philosophy and Phenomenological Research 83(1), 14–71. This paper is a good example of a strain of philosophical thought that rejects the claim that agents must be strongly unified for slightly different reasons than I do.

References Frankfurt, H., 1988a. ‘The Importance of What We Care About.’ Chap. 7 in The Importance of What We Care About: Philosophical Essays. Cambridge: Cambridge University Press: 80–94. Frankfurt, H., 1988b. ‘Identification and Wholeheartedness.’ Chap. 12 in The Importance of What We Care About: Philosophical Essays. Cambridge: Cambridge University Press: 159–176. Lewis, D., 1983. Philosophical Papers Vol. i (Vol. 1). Oxford: Oxford University Press.

383

Marya Schechtman Locke, J., 1975. Essay Concerning Human Understanding, P. H. Nidditch, ed. Oxford: Oxford University Press. Parfit, D., 1984. Reasons and Persons. Oxford: Oxford University Press. Perry, J., 1975. Personal Identity. Berkeley: University of California Press. Shoemaker, S. 1984. ‘Personal Identity: a Materialist’s Account’ in Shoemaker, S. and Swinburne, S. eds., Personal Identity: Great Debates in Philosophy. Oxford: Blackwell: 57–132.

384

35 AGENCY, NARRATIVE, AND MORTALITY Roman Altshuler

According to reductionism à la Parfit (1984), personal identity consists of psychological states bound together over time by psychological continuity. Similarly, causalism in the philosophy of action holds that an action is simply an event, typically a bodily movement, caused by mental states such as beliefs and desires. Narrative accounts offer a response to this reductionism in both domains. The basic idea is that narratives can incorporate a wide range of heterogeneous states, processes, and events as well as shape them. Most narrativists espouse a kind of holism on which the building blocks of lives and actions are structured by those lives and actions themselves.

1 Narrative Since the function of narrative is to unite heterogeneous elements—motives, actions, circumstances—what makes it a fruitful philosophical tool is ironically the slipperiness of the concept. Is the narrative connection primarily emotional (Velleman, 2003) or causal (Carroll, 2007)? While narrative is more than a simple list of events, is the difference one of kind or degree (Currie, 2010)? I will largely eschew these concerns and focus on a key point of agreement: ‘the most salient feature of narrative form in general is that the individual incidents and episodes in a narrative take their meaning from the broader context of the story in which they occur’ (Schechtman, 1996, p. 96). Since the meaning of the whole depends on how the narrative turns out—whether a story is one of triumph or failure will depend on whether the Hobbits destroy the ring—it follows that the meaning of earlier events is fixed by later events. Some narrativists present narrative as constitutive of personal identity, envisioning an entire life as embodying a single overarching narrative (Dennett, 1992; MacIntyre, 2007; Rudd, 2007). Recent theorists have often switched tactics, in part following Christine Korsgaard’s response to Parfit, to emphasize the role of agency in self-constitution. While Parfit holds that persons can be described metaphysically in impersonal terms as involving causal relations between their psychological states, Korsgaard defends a first-personal and practical conception. Although we have conflicting motives, we have only one body; thus, we cannot simply act on our motives, but must select which ones to act on, on the basis of reasons we commit to. In order to lead a life, that is, to act on motives such that we don’t DOI: 10.4324/9780429202131-43385

Roman Altshuler

constantly trip over ourselves, we must be able to commit to future goals. ‘You normally think you lead one continuing life because you are one person, but according to this argument the truth is the reverse. You are one continuing person because you have one life to lead’ (Korsgaard, 1989, p. 113). Our identity is thus not a metaphysical matter, but a practical one, and what unifies our agency is our set of agential commitments or practical identity, ‘a description under which you value yourself, a description under which you find your life to be worth living and your actions to be worth undertaking’ (Korsgaard, 1996, p. 101). Marya Schechtman (1996), meanwhile, has argued for a distinction between questions of reidentification and questions of characterization. The former have to do with identifying a person as the same being at different times. The latter deal with such questions as survival, moral responsibility, and self-directed concern. She views narrative as constitutive of the latter rather than the former sort of identity. Discussions of narrative have consequently largely—though not entirely—moved to a focus on its role in agency and practical identity. This is a salutary shift. Narrative conceptions of personal identity, insofar as they aim to oppose reductionism, run into a number of problems, which can be summarized under the headings of the accretion objection and the fragmentation objection (Altshuler, 2015a). According to the former, on a reductionist view, the future becomes a mere causal upshot of the past. If we think of the self as constituted by causal chains of psychological states, then the future is nothing more than a series of effects, leaving our agency—which narrative was meant to bring in—out of the picture. The suggestion of narrative identity theories is that narrative introduces something over and above psychological states, namely, the whole in light of which the elements must be interpreted. When thinking of self-constitution, this translates roughly to the idea that my sense of who I am and what I must do to be that sort of person shapes how I interpret and react to my past. Thus, it is my agency, rather than simply the force of my accreted past, that shapes who I am. But it isn’t clear how narrative helps with this: it is, after all, still my past that shapes my future. I may interpret and react to my past on the basis of my conception of my future, but isn’t that conception itself simply a product of the past? The fragmentation objection has several variants. One is the claim that self-narratives are prone to self-deception (Lippitt, 2007; Strawson, 2004). While narrativists can respond that self-deception is not a problem for narrative as such—we just need better narratives (Rudd, 2007)—the underlying problem is that narratives necessarily involve self-deception, precisely because, if narrative is to have a unity, it must exclude all features of a life that don’t fit. Consider Harry Frankfurt’s (1971, 1992) theory of identification. Frankfurt argues that we are responsible for our actions when we identify with the first-order desires that lead to those actions, and thus we aim to establish a unified will by identifying with some desires and excluding others as foreign. But, as David Velleman (2002) notes, this looks like a recipe for disaster: the desires we refuse to identify with continue to operate on us, but now outside of our ability to reign them in. Narrative accounts of personal identity seem to fall prey to the same problem. A better narrative won’t help since any unified narrative, by virtue of its unity, will necessarily exclude some elements of our identity, which nevertheless continue to shape it. The two objections reinforce each other. In responding to the accretion objection, narrativists argue that narrative shapes our identities, over and above the causal features of our lives, and it is the unity of narrative that gives it its significance (Malpas, 1998). But this response runs into the fragmentation objection: any attempt to impose such a unity will of necessity leave out features of our identity and thus cannot be constitutive of that identity. If the narrativist responds to the fragmentation objection by insisting that we need better 386

Agency, narrative, and mortality

narratives capable of incorporating all the elements of our identities, we run into the accretion objection: attempts to incorporate all such elements will destroy the narrative’s unity, so that once again our past becomes an accretion of events that do not cohere with each other, propelling us into the future through sheer causal force. The strongest response seems to be to abandon the idea that narrative constitutes our metaphysical identities, and instead emphasize its role in practical identity. Although all elements of my past play some role in my identity, I must choose some, form a coherent and unified narrative from them, and rely on it to shape my agency. This refocuses us on the extent of narrative’s participation in our agential self-constitution.

2 Agency Perhaps the most straightforward role for narrative in agency is developed by Peter Goldie in relation to Michael Bratman’s planning theory. Bratman argues that our agency contains three core features: ‘We are reflective about our motivation. We form prior plans and policies that organize our activity over time. And we see ourselves as agents who persist over time’ (Bratman, 2000, p. 35). Goldie aims to show that ‘there is something very important that makes narrative thinking a natural way of doing these things’ (Goldie, 2009, p. 101). This ‘something very important’ is the multiplicity of perspectives narrative allows. In telling narratives we display not only the emotions of their characters, but also the emotions we take to be appropriate to the telling; the audience, in turn, may have different emotions altogether. After a night out, full of confidence, a 23-year-old me attempts jumping a fence and rips his favorite jeans. In remembering my confidence retrospectively, I remember the episode as shameful and regret my bravado; in response to such regret, I envision a narrative in which I walk around instead, and I experience satisfaction at this image. Goldie’s point is that it is the emotional structure of narrative that allows for this interaction of emotions: the confidence I feel in the narrative of my past, the shamefulness with which I present it to myself in the narrative, the regret I experience in envisioning it, the calm I attribute to myself in the future, and the satisfaction I feel in relation to it. This emotional structure is key to planning, but it also makes reflectiveness unnecessary. If reflectiveness allows us to endorse policies that guide our future actions, Goldie argues that the emotional force that drives us to adopt narratives for future agency can dispense with such policies: through narrative’s emotional power, we can inculcate virtuous character traits, making certain actions unthinkable, and reflective policies designed to block them otiose. Meanwhile, by creating these links between past and future agency, narrative allows us to grasp ourselves as beings that exist over time. In this account, narrative is only a tool, though without narrative such agential capacities would require more thought and likely be vastly reduced. But narratives also provide context for exercises of agency, as shown by Velleman’s account of scenarios, which specify how to carry out various kinds of social moves. In visiting a post office, for example, I may wait in line, fill out a form, and walk up to the window, and I may improvise the details of these moves. I cannot, however, present my lunch or a lecture on Hegel, at least not if I want to send a piece of mail. ‘The scenario doesn’t supersede the ordinary psychological factors in explaining the agents’ behavior… indeed, it incorporates them’ (Velleman, 2009, p. 74). My desire to send a piece of mail is certainly part of the explanation for my waiting in line and approaching the teller’s window with a filled-out form; but it is the scenario, together with that desire, that explains my action. Velleman doesn’t claim that scenarios have a narrative 387

Roman Altshuler

form, but it’s reasonable to treat them as such: the desired outcome structures the antecedent steps, prescribing how the characters within the story are to move. And while desires are part of the explanation, most desires—especially ones like a desire to send a piece of mail— themselves have narrative explanations. Narratives may thus tap into our motivational structure while providing a background understanding of the social world within which motivation unfolds. Narratives also provide a background understanding of our own histories, values, capacities, and aspirations, and guide our agency via that combination. As Alasdair MacIntyre noted, ‘we identify a particular action only by invoking two kinds of context, implicitly if not explicitly. We place the agent’s intentions… in causal and temporal order with reference to their role in his or her history; and we also place them with reference to their role in the history of the setting or settings to which they belong’ (MacIntyre, 2007, p. 208). In MacIntyre’s view, we cannot understand human actions without narratives, because actions can be explained only via intentions, and we cannot understand those intentions without narratives that place them within the contexts of a culture and an individual life: whether a man gardening is doing so to please his wife or to prepare for winter will depend on two different sorts of narratives—one of spousal relations, and one of a ‘household-cum-garden setting’ (MacIntyre, 2007, p. 206). This picture is also key to Schecthman’s narrative self-constitution view, according to which ‘a person’s identity is created by a self-conception that is narrative in form’ such that ‘constituting an identity requires that an individual conceive of his life as having the form and the logic of a story’ (Schechtman, 1996, p. 96). We make sense of our past and future via narratives, and we use those narratives to guide our actions, thus constituting our selves. For example, a person who has lived through the Great Depression and one who was raised in times of plenty are likely to treat their finances very differently. MacIntyre held that in explaining an action, we need to place it in the context of an entire life, since it is from the whole that the part draws its significance. Schechtman adapts this view—dropping MacIntyre’s insistence that life must be oriented toward the good—that every action must derive its significance from an implicit narrative sense of a whole life (Schechtman, 2011). I may want to start my morning with coffee without a further thought, for example, but if questioned why, I can mention my poor sleep habits, a desire to finish some writing, and a recollection of several months in which I cut out coffee with unfortunate results for my research. Each of these explanations, in turn, has ties to other aspects of my life—aspirations, career ambitions, daily habits, a coffee-less experience abroad—each of which will have further such links. Not everyone believes that a narrative explanation of this kind, ‘an explanation of a mental state or action [that] connects that mental state/action to the subject’s sense of himself, of where he has been and where he is going (or trying to go)’ (Schroer and Schroer, 2014, p. 457), must tie into a sense of a whole life. Lumsden and Ulatowski, for example, adopt Schecthman’s view with the caveat that such connections take place only within individual ‘narrative threads’, allowing us to act in very different ways in the context of, say, a first date and a job interview, such that ‘within the person there are different stories running in the two situations, where each story provides not only a distinct history of events but also a distinct set of themes, vocabularies, descriptions, values, and associated emotional color’ (Lumsden and Ulatowski, 2017, p. 320). On their view, narrative self-constitution applies within each role we play, but switching between such narratives is largely an unconscious process, not reliant on some meta-narrative. This disunity of self thesis, however, may find difficulty explaining how concepts we utilize in our practical deliberation are narratively 388

Agency, narrative, and mortality

inseparable from elements across these different narrative threads (Malpas, 1998). While I may want a beer, both the desire for a beer and what would count as an object of this desire are, for me, shaped by a wide range of other desires, past experiences, and future aspirations (yes, aspirations, as indeed each beer must be logged into an app!). Thus, while it is probably true that there is no meta-narrative of a whole life in the same sense in which we might think of an action or pattern of actions having narrative form, it seems unlikely that the narrative threads governing different patterns of action could coexist without extensive narrative links to each other, and thus without some narrative sense that allows us to mediate between them. Indeed, insofar as these distinct threads can provide divergent perspectives on the same content, they allow for revision of our practical identities going forward (Atkins, 2008, p. 76). Narrative may also help structure action internally. Narratives make agents’ motives intelligible, but action itself may have a narrative form, since actions are shaped by their aim as narratives are by their conclusion (Carr, 1991, pp. 61–62). Furthermore, the connections necessary for practical reasoning that concludes in action seem to also have a narrative form. Beliefs and pro-attitudes do not typically fit so tightly together that an action necessarily follows from them; rather, they share themes with each other and with the proposed action, and these themes fit together in the form of a narrative (Bevir, 2000). Velleman proposes two further roles for narrative. First, via their emotional import, narratives can provide a level of self-understanding running parallel with causal-psychological self-understanding. On Velleman’s view, self-understanding is a constitutive aim of agency: in order to act we must act on attitudes that make us coherent to ourselves and act on them because they make us coherent to ourselves (Velleman, 1992). Insofar as narrative provides self-understanding, it can guide our agency. Although Velleman states that narrative self-understanding may be ‘supplementary’ and ‘optional’ (Velleman, 2009, p. 203), he elsewhere gives it a central role in our self-governance: a narrative self-understanding can provide an agent with a rationale of how to act next while preserving coherence (Velleman, 2006, p. 219).1 For Velleman, narrative can also explain why the value of a life seems to depend not simply on the sum of momentary well-being in that life, but on its shape. A life that begins with failure and slowly leads to success seems more desirable, especially if the success is a result of learning from failure, than a life that goes the other way, even if the latter has the same well-being overall. Velleman’s suggestion is that this difference is due to the more satisfying narrative the first sort of life affords (Velleman, 2000). John Fischer develops this idea, noting the distinctive feature of narratives, ‘that later events can alter the “meaning” or “significance” of earlier events. In this sense narratives can have “loops”… It is not that we can change the physico-causal past; but we can sometimes change its meaning and thus its contribution to the value of our lives overall’ (Fischer, 2009, p. 147). On this picture, every action has the potential to change the significance of past events, and the value of acting freely lies in the ability to redeem the past. Horace Pippin took up painting in order to strengthen his arm, wounded in the First World War. The resulting display of talent and vision can serve to cast the wound in a more positive light. If narratives play a crucial role in explaining an agent’s motivation and the connection between motivation and action, as well as between an individual action and the wider project or life in which it is embedded, and we accept that ‘loops’ are central to narratives, narrative may give us far more power over our past than simply the ability to change its emotional significance (Altshuler, 2015b). While some of our motives are simple—an itch causes me to scratch it—others are far more complex, and thus require complex narratives. And since narratives function by creating unity out of heterogeneous events, they specify motives not 389

Roman Altshuler

simply by attempting to recapture the various physiological, emotional, and conative events that occurred in an agent prior to action, but by picking out the relevant ones. Thus, if we want to provide a causal-psychological account of any given complex action, we cannot do so without narratives, and the motives we arrive at are at least partially constituted by those very narratives. In understanding ourselves and others, we do so through narratives that explain the meaning of actions by specifying what motives could make sense of both the action and the various psychological, physiological, and social events that preceded it. But since the meaning of events in a narrative depends on later events, the precise motives that led to an action will depend partially on the action they in fact lead to, and the actions will derive their meaning partially from their consequences and the projects of which they are a part, which, in turn, will depend on the life or narrative thread within which they occur. Part of what it means to be an agent, then, is to have the ability, through one’s actions, to determine, and perhaps reconfigure, one’s own motivational and psychological past.

3 Mortality The concept of death that emerges from a narrative view cannot be a biological one. Imagine an author constructing a narrative. She painstakingly crafts her characters, putting them in complex situations and giving them multifaceted goals. In one scene, they are relaxing on the eve of battle, but the author suffers a devastating loss and never returns to her story. The narrative ceases without a conclusion. We might say that it stops, but it does not end. The biological concept of death seems like the wrong sense of ending here. At least, it seems to have little meaning for narrative theorists, except in the sense that it represents a point at which our narratives are entirely taken over by others (Sartre, 2012, pp. 695–696). If death is going to be meaningful from a narrative perspective, it has to mean more than cessation. Heidegger makes some helpful distinctions here: first, death is not a stopping in the sense in which a road or rain stops. Second, it is not simply a ‘perishing’ or the cessation of biological processes. Nor is it a ‘demise’ or the sense in which we know we will one day stop living and leave behind bodies mourned by others (Heidegger, 1996, pp. 244, 247).2 In distinguishing death from all these biological, cultural, and psychological senses, Heidegger’s aim is to bring out death’s significance within the context of any human life rather than provide an analysis of it. For the narrative theorist, such significance could view death as a telos, the summation of life. Relatedly, death may be seen as the prospective conclusion in light of which the meanings of our lives take shape. 3 A narrativist account might thus suggest that death is necessary for life to have a meaning. After all, if the elements of a life derive their meaning from interconnections, and thus no part of a life can be understood apart from the whole, there cannot be meaning unless the boundaries are in view, and one of those boundaries is set by death. Without endings, Fischer (2009, p. 157) suggests before launching a rebuttal, our lives would lack narrative structure and thus the meaning it affords. This view that an immortal life would be meaningless finds its classic defense in Bernard Williams (1973). Williams’s argument famously proceeds along the following lines: there are two options for immortality. One possibility is that an immortal’s character goes on changing, just as it does in the course of an ordinary life, such that over time the person becomes unrecognizable to her past self. Another possibility is that one’s character remains constant and thus an immortal life would eventually result in interminable boredom. The reason immortality

390

Agency, narrative, and mortality

would grow boring, for Williams, seems to be that we would run out of categorical desires— desires that are not conditional on whether or not we remain alive. Such categorical desires are closely tied to our projects and commitments (Altshuler, 2013), and one might think that without them, nothing could make our lives meaningful to us. A defense of Williams suggests two possibilities: first, an endless life would be shapeless, and thus impossible to evaluate for its desirability (Altshuler, 2015c; Burley, 2009); second, the very values by means of which we evaluate lives are themselves constitutively dependent on mortal lives (Malpas, 1998; Nussbaum, 1989; Scheffler, 2013). Williams’s critics respond that immortal lives need not be shapeless: they can gain a shape from their projects, which create their own (typically overlapping) narratives (Chappell, 2009; Ferrero, 2015; Fischer, 2009). These local narratives, in turn, can allow for the finitude necessary to sustain value. Such responses would seem a natural fit for Lumsden and Ulatowski’s version of narrative disunity. But while these narratives can guide action in particular domains and support values internal to them, it is less clear that they can equally well support values that seem to have a global place within life. Much of the debate proceeds as if the narrative functions of shaping and guiding agency were intrinsic to agency as such, and thus could be adapted to different kinds of agents. But the ubiquity of narrative may be an adaptation to our mortality rather than an accompaniment. Consider the suggestions developed by Velleman and Fischer, that a life’s value depends on its narrative shape, and that the value of acting freely lies in our ability to redeem our past. It is relatively clear why, given our limited life-spans, we might employ narrative forms of thinking in order to maximize those lives’ values. But in an endless life, redeeming the past is a pointless task, and once an overall narrative is no longer possible, it is less obvious that local narratives can maintain their cohesion. Part of Galen Strawson’s critique is that any view that plays up the ubiquity of narrative in life will be either too strong to be believable or too weak to be informative (Strawson, 2004). It will either define narrative such that practically no one qualifies as possessing one or makes narrativism trivially true—if simply acting on a desire requires a narrative, the thesis that narrative permeates human life becomes vapid. Without some overall guidance from a global narrative, the role of narrative might degrade to this trivial condition. Why use a complex cognitive mechanism to temporally organize one’s life when such temporal organization is unnecessary? One response is that opportunities for action will present themselves to immortals as well as mortals. But whether immortals have reason to care about such opportunities remains an open question. Perhaps we derive meaning from narratives within our mortal lives, but it is only mortal lives that make room for such narratives.

Related topics Agency and personal identity; planning agency; Diachronic agency; Agency and morality; Agency and the emotions; Agency and identification.

Notes 1 Against Velleman’s claim of a supplementary role for narrative, see Altshuler (2015b) and Ward (2019). 2 Page numbers refer to the standard German edition. 3 For discussion of the mortality problem for narrativism, see Chapter 5 of Davenport (2012). Behrendt (2016) argues that mortality poses no special problem for narrative accounts.

391

Roman Altshuler

Further reading Atkins, K., Mackenzie, C. (Eds.), 2008. Practical Identity and Narrative Agency. Routledge, New York. Key work in the transition from a focus on metaphysical to practical identity, with a helpful introduction. Hutto, D. (Ed.), 2007. Narrative and Understanding Persons. Cambridge University Press, Cambridge. A collection of important essays by both prominent critics and proponents of narrative identity. Ricoeur, P., 1992. Oneself as Another. University of Chicago Press, Chicago. Along with Ricoeur’s earlier work on narrative, this is a founding source for work on narrative identity. Stokes, P., 2015. The Naked Self. Oxford University Press, Oxford. Critically examines how Kierkegaard can shed light on narrative and identity.

References Altshuler, R., 2015a. Teleology, Narrative, and Death, in: Stokes, P., Lippitt, J. (Eds.), Narrative, Identity and the Kierkegaardian Self. Edinburgh University Press, Edingburgh, pp. 29–45. Altshuler, R., 2015b. Free Will, Narrative, and Retroactive Self-Constitution. Phenomenology and the Cognitive Sciences 14, 867–883. Altshuler, R., 2015c. Immortality, Identity, and Desirability, in: Cholbi, M. (Ed.), Immortality and the Philosophy of Death. Rowman & Littlefield, Lantham, MD, pp. 191–203. Altshuler, R., 2013. Practical Necessity and the Constitution of Character, in: Herrera, C.D., Perry, A. (Eds.), The Moral Philosophy of Bernard Williams. Cambridge Scholars Publishing, Newcastle, pp. 40–53. Atkins, K., 2008. Narrative Identity and Moral Identity: A Practical Perspective, 1st ed. Routledge, New York. Behrendt, K., 2016. Learning to Be Dead: The Narrative Problem of Mortality, in: Cholbi, M. (Ed.), Immortality and the Philosophy of Death. Rowman & Littlefield, Lantham, MD, pp. 157–172. Bevir, M., 2000. Historical Explanation, Folk Psychology, and Narrative. Philosophical Explorations 3, 152–168. Bratman, M., 2000. Reflection, Planning, and Temporally Extended Agency. The Philosophical Review 109, 35–61. Burley, M., 2009. Immortality and Meaning: Reflections on the Makropulos Debate. Philosophy 84, 529. Carr, D., 1991. Time, Narrative, and History. Indiana University Press, Bloomington, IN. Carroll, N., 2007. Narrative Closure. Philosophical Studies 135, 1–15. Chappell, T., 2009. Infinity Goes Up On Trial: Must Immortality be Meaningless? European Journal of Philosophy 17, 30–44. Currie, G., 2010. Narratives and Narrators: A Philosophy of Stories. Oxford University Press, Oxford. Davenport, J.J., 2012. Narrative Identity, Autonomy, and Mortality: From Frankfurt and MacIntyre to Kierkegaard, 1 ed. Routledge, New York. Dennett, D.C., 1992. The Self as a Center of Narrative Gravity, in: Kessel, F.S., Cole, P.M., Johnson, D.L. (Eds.), Self and Consciousness: Multiple Perspectives. Lawrence Erlbaum, Hillsdale, NJ, pp. 103–115. Ferrero, L., 2015. Agency, Scarcity, and Mortality. Journal of Ethics 19, 349–378. Fischer, J.M., 2009. Free Will, Death and Immortality: The Role of Narrative, Chap. 9 in Our Stories: Essays on Life, Death, and Free Will. Oxford University Press, New York, pp. 145–164. Frankfurt, H., 1971. Freedom of the Will and the Concept of a Person. The Journal of Philosophy 68 (1), 5–20. Frankfurt, H., 1992. The Faintest Passion. Proceedings and Addresses of the American Philosophical Association 66, 5–16. Goldie, P., 2009. Narrative Thinking, Emotion, and Planning. The Journal of Aesthetics and Art Criticism 67, 97–106. Heidegger, M., 1996. Being and Time. SUNY Press, Albany. Korsgaard, C., 1996. The Sources of Normativity. Cambridge University Press, Cambridge. Korsgaard, C., 1989. Personal Identity and the Unity of Agency: A Kantian Response to Parfit. Philosophy & Public Affairs 18, 101–132. Lippitt, J., 2007. Getting the Story Straight: Kierkegaard, MacIntyre and Some Problems with Narrative. Inquiry 50, 34–69.

392

Agency, narrative, and mortality Lumsden, D. and Ulatowski, J., 2017. One Self per Customer? From Disunified Agency to Disunified Self. The Southern Journal of Philosophy 55, 314–335. MacIntyre, A.C., 2007. After Virtue: A Study in Moral Theory. University of Notre Dame Press, Notre Dame, IN. Malpas, J., 1998. Death and the Unity of a Life, in: Solomon, R.C. and Malpas, J. (Eds.), Death and Philosophy. Routledge, New York, pp. 120–134. Nussbaum, M.C., 1989. Mortal Immortals: Lucretius on Death and the Voice of Nature. Philosophy and Phenomenological Research 50, 303–351. Parfit, D., 1984. Reasons and Persons. Oxford University Press, Oxford. Rudd, A., 2007. In Defence of Narrative. European Journal of Philosophy 17, 60–75. Sartre, J.-P., 2012. Being and Nothingness: An Essay on Phenomenological Ontology. Washington Square Press, New York. Schechtman, M., 2011. The Narrative Self, in: Gallagher, S. (Ed.), The Oxford Handbook of the Self. Oxford University Press, Oxford. Schechtman, M., 1996. The Constitution of Selves. Cornell University Press, Ithaca, NY. Scheffler, S., 2013. Death and the Afterlife. Oxford University Press, Oxford. Schroer, J.W. and Schroer, R., 2014. Getting the Story Right: A Reductionist Narrative Account of Personal Identity. Philosophical Studies 171, 445–469. Strawson, G., 2004. Against Narrativity. Ratio XVII, 428–452. Velleman, J.D., 2009. How We Get Along. Cambridge University Press, New York. Velleman, J.D., 2006. The Self as Narrator, Chap. 8 in Self to Self. Cambridge University Press, Cambridge, pp. 203–223. Velleman, J.D., 2003. Narrative Explanation. The Philosophical Review 112, 1–25. Velleman, J.D., 2002. Identification and Identity, in: Buss, S. and Overton, L. (eds.), Contours of Agency: Essays on Themes from Harry Frankfurt. MIT Press, Cambridge, MA, pp. 91–123. Velleman, J.D., 2000. Well-Being and Time, in: The Possibility of Practical Reason. Oxford University Press, New York, pp. 56–84. Velleman, J.D., 1992. What Happens When Someone Acts? Mind 101, 461–481. Ward, D., 2019. Moving Stories: Agency, Emotion and Practical Rationality, in: Candiotto, L. (Ed.), The Value of Emotions for Knowledge. Springer International Publishing, New York, pp. 145–176. Williams, B., 1973. The Makropulos Case: Reflections on the Tedium of Immortality, in: Problems of the Self. Cambridge University Press, New York, pp. 82–100.

393

PART 8

Agency, reasoning, and normativity Introduction to Part 8 Luca Ferrero

There seems to be an intimate connection between agency and normativity. The conduct of agents is expected to respond to normative pressures, and it is explained, understood, and evaluated in normative terms. By normative pressures, I mean such notions as permissions, prohibitions, obligations, requirements, demands, prescriptions, norms, principles, and laws; by normative terms, I mean such notions as reasons, rationality, correctness, rightness, and goodness. The existence of the connection between agency and normativity can also be made apparent by considering entities that are not agents or manifestations or expressions of agency. Non-agential entities are neither responsive to, nor moved by, or explained, or understood in terms of, or held to any normative pressures, standards, or expectations. How the laws of nature ‘govern’ physical phenomena, after all, is not the same way in which the laws of morality or rationality govern and guide the behavior of agents. But what about assessments of functional goodness that are ordinarily applied to nonagents, such as parts of organisms or entire artifacts (or parts thereof )? Evaluations are indeed applied to these non-agential entities but it seems that agency still looms in the background. After all, these functional attributions are either about parts of organisms (which by their nature possess at least some minimal or proto-agency) or about artifacts (which are such only because of the agency of their designers, creators, users, or evaluators). In any event, as we move beyond the evaluations made possible by the attribution of functions, the connection between agency and normativity seems uncontroversial. And different kinds of agency might map onto different kinds of subjection and responsiveness to normative pressures. An important step in this progression is the emergence of genuine aims or ends and the correlated notion of agency as goal-directedness or purposiveness. By the time we get to our full-blooded agency, normativity is present in full swing. We are the paradigmatic examples of beings with all the following features. We are beings (i) who are expected to respond to reasons, assessments, evaluations, and normative pressures; (ii) whose behavior is explained, understood, and evaluated in terms of this responsiveness; (iii) who hold fellow creatures responsible, by holding them to similar expectations and normative pressures; (iv) who explain, try to make sense of, and evaluate fellow creatures on the basis of similar expectations and pressures; (v) who are expected to at least try to explicitly articulate the contents of these expectations and pressures, and to explicitly justify their conduct in light of these articulations—by what is sometimes called the ‘asking and giving’ DOI: 10.4324/9780429202131-44395

Luca Ferrero

of reasons; (vi) who have some understanding and appreciation of normativity as such (i.e., who explicitly make demands in normative terms when engaging in explicit theoretical and practical reasoning, who use the appropriate modal and deontic terms both in describing and prescribing the conduct of their fellow creatures and of themselves, who insist in offering or asking for reasons or justifications in their interactions with each other); and, finally, (vii) who might, occasionally, try to be self-reflective about the nature of normative pressures and the source of and the justification for the contents and authority of these pressures. Normativity pervades full-blooded agency through and through. It should come as no surprise, therefore, that philosophers have been trying to leverage the very nature of agency to try to get a better understanding of the nature, content, and force of normative pressures and demands. This is what is known as ‘constitutivism.’ How this project is supposed to work and whether it has any hope of succeeding are something that the readers will need to ascertain for themselves after looking at the chapters in this section (especially ‘Agency and normativity,’ ‘The aim of agency,’ and ‘Agency and morality.’) All I need to remark here is that the attempt to pursue such a project in the context of the philosophical reflection on the nature of agency has at least some intuitive plausibility. Finally, the connection between normativity and agency is not limited to the pressures that come from the largest or more general domains of assessment and justification, that is, the domains of rationality and morality. A similar connection with agency is also presupposed in other, more specific domains, enterprises, and practices including games (which are discussed in an earlier chapter on ‘Agency and games’), the law, and aesthetics. The latter case, which is discussed in detail in the closing chapter, is especially interesting in this respect. The recognition of the existence of a genuine kind of aesthetic agency depends on the acknowledgment of an intimate connection between agency and normativity in our practices of rationally assessing, reactively responding to, and holding people responsible for a variety of attitudes like aesthetic appreciation, even if these attitudes are different from intentional actions in that they neither are nor should be formed ‘at will.’

Agency and reasons In ‘Agency and Reasons,’ Maria Alvarez offers a survey of the different kinds of reasons that are invoked in connection with rational agency in three different contexts: assessment, deliberation, and explanation. These are, respectively, normative, motivating, and explanatory reasons. Normative reasons are considerations that favor an action or decision, as judged by a well-informed, impartial observer. A normative reason might just be ‘pro-tanto’; it might favor an action but not necessarily justify it, since it might be outweighed or defeated by other normative reasons. Normative reasons are usually considered to be facts. For so-called ‘perspectivists,’ however, what gives these facts justificatory force partly depends on the agent’s beliefs or knowledge of these facts. For the ‘objectivists,’ instead, justification is due solely to the facts, although whether the justification offered by an agent is praiseworthy or blameworthy might depend on features of the agent’s epistemic situation and psychology. Motivating reasons are considerations that the agent takes to favor her action, and in light of which she acts. These are the reasons that figure as premises in her practical reasoning. This is a technical use of the term ‘reason.’ Motivating reasons are not to be confused with ‘motivating factors’ (such as intentions, desires, and emotions) which play a role in moving the agent but are not considerations that the agent necessarily takes to favor her action. 396

Agency, reasoning, and normativity

Although they are sometimes bundled together with motivating reasons, Alvarez considers explanatory reasons as a separate kind. These are the considerations that explain the agent’s conduct in the mode of rationalization—they explain the action by making it intelligible. Similar to the distinction between normative and motivating reasons (not every justifying consideration is one that motivates the agent—and vice versa), not every explanatory consideration is among the reasons that motivates the agent. Using Alvarez’s example, Othello’s motivating reason to kill Desdemona is the (putative) fact that Desdemona has been unfaithful (which is what Othello takes as a premise in his reasoning) but this is different from what explains his conduct, which is the (actual) fact that he believes (although falsely) that she has been unfaithful. ‘Psychologism’ is the standard view that distinguishes between normative/justifying reasons and motivating reasons in ontological terms: facts justify, whereas only mental states motivate. However, several ‘non-psychologists’ views have been recently developed, centered on the idea that motivating reasons are premises of practical deliberation, which under normal conditions are not about the agent’s own psychological states. The main problems for non-psychologism are so-called error cases, when the agent is motivated to act by a false consideration (like Othello). Two possible solutions have been offered to these cases: (a) the agent is acting for a consideration that is false but that the agent believes in, or (b) the agent is acting on an ‘apparent reason.’ Another issue for non-psychologism is whether acting for a motivating reason requires that the agent be in some kind of epistemic relation to that reason. As Alvarez argues, psychologism might nonetheless be plausible for explanatory reasons since, in order for a reason to be able to rationalize an action, that reason must be part of the agent’s psychology. This is especially true in error cases: given that all explanations are factive, one cannot offer a true explanation in terms of something false.

Agency and practical reasoning In ‘Agency and practical reasoning,’ Jules Salomone-Sehr and Jennifer M. Morton discuss how practical reasoning is the one form of coming to act that imprints our actions with the distinctive marks of rational full-blooded agency. This can be seen by contrasting practical reasoning with acting out of habits or impulses. Salomone-Sehr and Morton argue that any adequate account of practical reasoning must meet four criteria: (i) mentality: practical deliberation is by and large a conscious mental process (rather than unconscious or sub-personal); (ii) practical reasoning is answerable to a range of evaluative standards, for example as rational, effective, and conclusive (contrast with such mental processes as free association or daydreaming); (iii) practical reasoning is primarily concerned with settling what to do (which leaves open whether it culminates in a belief, an intention, or an action); (iv) reasoning is something that the reasoner does—it is attributable to her—rather than something that merely occurs to her. Salomone-Sehr and Morton then consider the influential account of John Broome. According to Broome, we are subject to both synchronic and diachronic requirements to make our attitudes consistent, coherent, and to stand in the right relations with each other. We can satisfy these requirements through automatic processes but reasoning helps us make adjustments to our attitudes as a result of a ‘rule-guided activity’ that operates on the content of the attitudes. (This is to be contrasted with accounts of reasoning in which we reason from the attitudes themselves, accounts that Broome deems to be more cognitively sophisticated than ordinary reasoning appears to be.) 397

 Luca Ferrero

The second half of the chapter is devoted to three challenges to Broome’s account. First, there is the question of the role that explicit self-attributions of attitudes might play as premises in practical reasoning. Second, Salomone-Sehr and Morton discuss the nature of the conclusion of practical reasoning. Against Jonathan Dancy’s proposal that practical reasoning does not conclude in an intention but in action, they raise two worries: that action does not satisfy the mentality criterion of reasoning and that action could not be the conclusion of practical reasoning directed at a distal action. Then they consider the normative status of good reasoning and whether assessments of rationality might come in degree rather than being one-off. Lastly, they look at the issue of whether we should care about having a tidy psychic economy by avoiding inconsistencies between our attitudes.

Agency and normativity In ‘Agency and normativity,’ Kenneth Walden considers whether it is possible to establish the normative authority of certain principles by showing that they are constitutive of agency, or at least of certain kinds of agency. The starting point for establishing a connection between normativity and agency is to consider that, when we deliberate, we ask normative questions (about what we ought to do or what reasons we have to do so-and-so) that presuppose distinctive forms of agency. These forms of agency have certain constitutive features, which can engender normative authority for the deliberator via three possible lines of argument. First, as long as one takes up the constitutive aim of a certain agency as one of the premises of one’s deliberation, one has reason to do what is necessary to accomplish this aim. Second, the constitutive features of a kind of agency come with inherent evaluative standards: the more fully one possesses these features, the better one is as that kind of agent. In presupposing that agency in one’s deliberation, one is subject to these evaluative standards and thus one has reason to be a good instance of that kind, rather than a defective one. And questions about what reasons one has can be deferred to the ideal version of that kind of agent. Third, in presupposing a kind of agency in one’s deliberations, one also presupposes to have internalized a psychological schema that allows one to exercise that agency. And this schema affects the reasons one has by providing the background framework that structures one’s practical reasoning (e.g., the kinds of inferences that are licensed or the premises that are taken as given) rather than one’s aims or ideals. The derivation of authority from agency is usually done in terms of the constitutive conditions of generic agency, or agency tout court, which promises to derive unconditional normative conclusions. But similar arguments can also be offered about the normative implications of narrower, more contingent, and more local kinds of agency. The reasons that can be derived from the latter kinds of agency are conditional on presupposing the narrower agency. But this contingency does not make them necessarily unimportant: for certain kinds of agency, it might be very hard for a deliberator to give up asking the deliberative questions from within that narrower agency (for instance, powerful social and economic forces might make it impossible to leave that agency). Walden argues for ‘deliberative externalism’: the objects and presuppositions of our deliberations are dependent on features of the world beyond the beliefs and intentions of the deliberator. So we might have far less control over what our deliberations require than we think. If one takes seriously that there are different kinds of agency keyed to different physical and social systems, difficult questions arise about how to individuate the specific kinds of agency and to determine which kinds there are. 398

 Agency, reasoning, and normativity

As for standard constitutivist arguments based on generic agency, they purport to derive unconditional normative authority given that generic agency is ‘inescapable,’ that is, it is presupposed when entering any deliberation at all. The arguments can be formulated in terms of constitutive aims of agency (for instance, self-understanding or will-to-power) or in terms of evaluative standards that attach to ideals of agency. Two objections arise. First, one might deny that (generic) agency is normatively special. According to the so-called ‘shmagency’ challenge, one might imagine a subject how is perfectly content to be a shmagent—a nonagent who is very similar to agents but lacks or doesn’t care about the constitutive features of agency, and thus is not subject to their normative implications. As Walden shows, there are plausible avenues for the constitutivists to respond to this challenge. Second, there is the worry that the constitutive features of generic agency are too thin to provide the basis for any interesting normative implications. Walden sketches a possible reply: it is possible that agency by itself does not give us any special aim or standards but we still need to secure a kind of convergence to avoid too much heterogeneity in our behavior. If so, we might be able to derive a different kind of normative principle from such agency, “not a principle to adhere to standards reflecting the fixed nature of agency,” but as Walden concludes “a principle requiring our cooperation in creating and sustaining agency—whatever that ends up being.”

The aim of agency The aims of particular actions are explanatory not just of the actual unfolding of the actions but also of the normative standards that apply to them. One can appeal to these aims in the evaluation, justification, and motivation of actions. For instance, in playing chess, one is under the demand to checkmate one’s opponent when given the (known) opportunity to do so. This demand can be explained by the constitutive aim of chess playing (winning by checkmating, that is). What if agency itself had a constitutive aim? Could this aim be used to explain the content and authority of the normative standards of agency? As discussed by Kathryn Lindeman in ‘The aim of agency,’ this is what ‘aim theorists’ want to accomplish. There is a variety of aim theories, which can be characterized along four dimensions. First, the particular aim that they identify. Second, the scope of their explanatory ambitions (from the most ambitious, which purport to ground practical rationality and morality, to the more modest, which purport to ground only the instrumental principle—the requirement to take the recognized means to one’s ends). Third, the role played by the aim in guiding action. For some, it is an end that all actions must be understood as having because of their function or purpose; for others, it is a sub-agential motive or drive whose guidance makes a certain behavior into genuine action. Finally, the aim might be either formal (such as aiming at the good as the formal object of choice) or substantive (e.g., self-constitution). Even assuming that there could be some agreement in identifying what the aim of agency is, aim theories face several challenges. To begin with, there are three concerns about the explanation of the content of normative standards. First, there seems to be an inverse relationship between an aim’s plausibility and its explanatory power; second, the opacity of the aim of agency might make the standards practically inert (including the risk that they might be self-effacing); third, there is the worry that the aim might generate highly implausible standards. Even if the content of plausible standards might be derived from the aim of agency, there are still concerns about the explanation of the authority of these standards. Lindeman helpfully characterizes them as follows. First, by itself, having an aim cannot provide a reason to satisfy or achieve that aim (the shmagency challenge). Second, the ‘why bother 399

 Luca Ferrero

objection’—having a reason to satisfy an aim does not entail having a reason to satisfy every standard explained by that aim. For instance, if the aim is a scalar one, it is unclear why it provides the agent with a reason to maximally satisfy it; additionally, for any aim, it seems that one could always achieve it badly or luckily. Third, the ‘presupposition objection:’ the explanatory project might smuggle in normative presuppositions (including the endorsement of the instrumental principle, which was supposed to be the minimal norm that could be grounded by any aim theory).

Agency and morality What can we learn about the nature of morally good action from the nature of agency? This is the topic of Christine M. Korsgaard’s chapter on ‘Agency and morality.’ Many moral philosophers have been unclear about what they mean by ‘action,’ especially whether the end that one is pursuing is included in the content of the action or not. According to the empiricist/consequentialist views, the end is given to the agent, whose only choice is about the act that she takes to bring about that end. Hence, for the consequentialist an action is right if the end that it promotes is a good one. By contrast, in the deontological views (which for Korsgaard comprise both Kant and Aristotle), agents choose their ends, which are part of what they choose when they chose an action. For these views, the rightness of an action is a form of goodness, applied not to their consequences but to the form of the action itself. For the consequentialist, what is right or wrong in acting is what Korsgaard calls the ‘act,’ the movement through which one promotes the (given) end. Its goodness is instrumental to the goodness of the end. For the deontologist, instead, what has moral value is the action has a whole: the act-for-the-sake-of-the-end. The goodness is intrinsic to the action. An action is good on account of its nature as action. What is the nature of action? What makes an action well-formed, and why does an action that is well-formed count as a morally good action? According to so-called ‘constitutivist’ accounts, normative standards arise from the function or nature of the object to which they are applied, that is, from its constitutive, internal standards. The function of action is to make its agent efficacious for a given end. This function gives rise to two standards. First, the act must relate to the end so as to make the agent the kind of thing that tends to achieve that end. Second, the act must be related to the whole action to make sure that the agent herself is non-accidentally efficacious rather than by virtue of some alien force working through her. The efficacy that makes an action good is not just the efficacy of the act. It is the efficacy of the agent as an agent. The action as a whole must be attributable to the agent, and hence, it has to originate in a choice that represents the activity of the agent as such. For Kant, the agent is the source of the action when the action originates in the agent’s autonomy, that is, when the agent acts on principles that she recognizes as universal laws. For him, autonomy involves acting on moral principles, and hence it is only in meeting the standards of action that one can act morally. It is only in acting morally that an agent can make herself genuinely efficacious through her actions. What one brings about in acting morally is nothing other than the goodness of action as such. It is the very nature of action that it is to be brought about, as the well-formed action, for its own sake. Put in Aristotelian terms, as Korsgaard concludes: “morally good action is a good thing for its own sake because there is a way in which our lives consist of rational, chosen, action, and morally good action is the excellence of rational chosen action. Doing what is right is good simply because it is a way of living well.” 400

 Agency, reasoning, and normativity

Agency in the law According to Anglophone criminal law, criminal liability and punishment are only appropriate for agents. Someone can only be punished for an act or an omission, that is, for a willed bodily motion (or the absence of such an act) by someone who is able to perform it. In either case, the law explicitly operates with a common sense view of agency—where the agent is a creature capable of acting and omitting. But, as Gideon Yaffe argues in ‘Agency in the law,’ there is also a somewhat hidden conception of agency in the law: the peculiar and elusive doctrine of ‘proximate causation.’ This is a conception of agency that applies to particular events, according to which a person can be more or less the agent of a particular harm; the harm can be more or less one’s doing. From this distinction, it follows that different treatments are appropriate to the parties involved. There is an ordinary idea of being the agent-of a certain event which is used to assess moral responsibility. This idea has both an intrinsic and a comparative aspect. In the intrinsic sense, one is the agent of an event just in case one’s action is one of the predictable causes of that event. This condition is quite easy to satisfy since even a minimal causal contribution counts. But as such it is not very useful for attribution of responsibility. The comparative aspect is what matters, since it helps determine who is ‘most the agent of ’ a harm and bears the burden of responsibility. The law mirrors this aspect of ordinary morality and thus needs a theory of comparative agency-of. This is what is partially encoded in the legal doctrine of ‘proximate causation.’ Under the law, a person is liable for a particular occurrence only if she ‘proximately caused’ it. As Yaffe argues, the legal criteria for being the agent-of an event are buried in the criteria for proximate causation. There are three central tenets to the doctrine of proximate causation: cause-in-fact (one is an agent-of if and only if the event would not have occurred had the defendant not performed some act or omission), reasonable foreseeability (to rule out freakish or unpredictable turn of events), and the absence of voluntary intervention of another agent (informally put, the interventions of other agents can break chains of proximate causation). The latter tenet is the one that bears on the comparative sense of agency-of (which matters for the determination of whom has the obligation to remedy the harm). The law isolates three factors that contribute to making a person more an agent of some events than others: first, the degree of commitment to the occurrence of the event; second, temporal proximity; third, the more one is trying to do something wrongful when causing the harmful event, the more one is the agent of an event that one does not intend. The last factor, however, is simply irrelevant to comparative agency-of. And the legal rule encodes no clear theory of how to weigh the three factors. These would be serious flaws if this were a philosophical doctrine of agency-of. But there is no such problem for the legal doctrine once we realize, as Yaffe argues in closing, that the point of the legal doctrine is to guide practices of assigning liability that might go beyond the mere determination of who is most the agent-of the relevant event.

Aesthetic agency Engagements with aesthetic value pervade human life. Are these engagements a matter of agency? In ‘Aesthetic agency,’ Keren Gorodeisky defends the idea that there is a genuine kind of aesthetic agency. This claim goes against a traditional view, which she calls the ‘practical approach,’ that identifies agency with the capacity for intentional action and voluntary control. This is no surprise, given that aesthetics has traditionally focused on appreciation rather 401

 Luca Ferrero

than action. But Gorodeisky argues that aesthetic appreciation counts as a manifestation of rational agency once we take what she calls the ‘authority approach.’ This is an account that does justice to the attributions of various conative and affective attitudes (including beliefs, emotions, and desires) to our agency even if these attitudes are not-voluntary. This is evinced by our practices of rationally assessing, reactively responding, and holding people responsible for these attitudes even if they are not formed ‘at will.’ The authority approach has the merit of accounting for the derivative character of aesthetic actions in relation to appreciation. Actions are of the utmost importance in the aesthetic domain, which would be unthinkable and unrecognizable without the creative acts of artists and the actions of curators, editors, landscapers, and so forth. But Gorodeisky argues that we cannot understand what makes these actions aesthetics independently of their relation to aesthetic appreciation. The ‘aesthetic’ as such cannot be characterized by any set of properties or objects independently of a reference to appreciation. Genuine aesthetic agency is first and foremost exercised in acts of the rational-affective capacity for appreciation, which is different from the will. The aesthetic is that which merits a particular kind of appreciation. Gorodeisky conjectures that the identification of agency with the voluntary might be grounded in the correct thought that agency is required for the responsibility that goes with the evaluation of the appropriateness of certain attitudes. But the attitudes in question are not restricted to intentional actions and exercises of the will. This responsibility is not causal responsibility but the one that we take over the attitudes that are expressions of our rational agency: the attitudes that we can avow or disavow when subject to rational reflection or criticism. Aesthetic appreciation (the cognitive kind of feeling that presents its object as meriting appreciation and involves an awareness of itself as merited by its object) is one of these attitudes. This appreciation is an exercise of rational agency insofar as it is constituted by the consciousness of its own appropriateness. As Gorodeisky writes, “emotions, desires, and acts of aesthetic appreciation can themselves be direct modes of experiencing the world and what it merits: they are affective and conative presentations of the world as meriting themselves.” The self-awareness of the appropriateness of these attitudes is the central mark of rational agency. This is not to deny that aesthetic appreciation is receptive. But according to the authority approach, this receptivity is one of the modes in which we, as human beings, exercise our rational agency. Still, this does negate the fact that acts of aesthetic appreciation are reason-responsive and active: unlike sensations, they are regularly assessed, explained, and criticized in terms of this responsiveness.

402

36 AGENCY, REASONS AND RATIONALITY Maria Alvarez

1 Introduction: reasons and actions On 24 September 2019, the UK’s Supreme Court, the final court of appeal for civil cases, ruled that the decision by the prime minister (PM) ‘to prorogue, or to advise Her Majesty the Queen to prorogue Parliament’ for five weeks was unlawful because prorogation had ‘the effect of frustrating or preventing, without reasonable justification, the ability of Parliament to carry out its constitutional functions’ ([2019] UKSC 41; my italics). The Court also noted that it saw no need to consider whether the PM’s motive in proroguing Parliament was itself unlawful. The distinction between the lawfulness of the decision to prorogue and that of the PM’s motive in deciding to prorogue Parliament can be put in terms of reasons of different kinds. The lawfulness of the first concerns whether there were reasons that justified the decision to prorogue Parliament, while the second is about the reason(s) that motivated the PM to do so. Philosophers call reasons of the first kind ‘normative’ or ‘justifying’ reasons: they are reasons that, very roughly, favour and may justify an action or decision, as judged by a well-informed, impartial observer – for example, a law court. Reasons of the second kind are called ‘motivating’ reasons because, again roughly, they guide and motivate agents in their deliberations and decisions about what to do. This distinction raises a host of questions. First, what exactly is a reason? And are there just two or more types of reasons? Further, what characterises each type? And how, if at all, are the different types of reason connected? Philosophers have developed different accounts of reasons and, accordingly, different answers to these questions. This chapter provides an overview of the various accounts and responses, highlighting connections and disagreements among them. The focus of the chapter is on reasons for actions and decisions – on so-called ‘practical reasons’ – by contrast with reasons for believing, wanting, feeling emotions and any other phenomenon where we are responsive to reasons. Some of the chapter’s arguments and conclusions may well apply to reasons in those domains too, but no attempt will be made to ascertain whether that is so or to identify peculiar difficulties that pertain to those other domains.

DOI: 10.4324/9780429202131-45403

Maria Alvarez

2 Reasons: their nature and kinds The capacity to reason is central to our higher mental capacities. In the practical domain, we reason more or less explicitly about what to do and how to do it. Normally, we aim to act in light of the reasons that guide such deliberations and, when we do, we cite those reasons in our attempts to make ourselves intelligible to others and to justify our actions. We also engage in second-order deliberation: we examine and assess the practical reasoning that underlies our decisions and behaviour. And attempts to evaluate behaviour, whether from a first-, second-, or third-person perspective depend, first, in identifying and assessing any reasons that may recommend or counsel against such behaviour, along different evaluative dimensions: for example, rationality, morality, virtue or legality. And second, we evaluate behaviour by assessing the conduct but also the reasons that motivated the agent to act as she did: she did the right thing, but did she do it for the right reasons? Or, vice versa, she may have done the wrong thing but for good reasons. As noted above, the contemporary literature uses the labels ‘normative’ and ‘motivating reasons’ for the two kinds just described (see, e.g. Raz 1975, 1999; Dancy 2000; Parfit 1997). Until relatively recently, this distinction was typically understood in ontological terms so that reasons of each type were thought to be different kinds of thing. Normative reasons were conceived of as facts and were, therefore, regarded as mind-independent – at least in this sense: whether something is a fact does not seem to depend on whether anyone has ever considered, let alone knows, that thing. By contrast, motivating reasons were regarded as mental states of agents’ and, as such, as mind-dependent entities (see Audi 2001; Mele 2003). In recent years, however, this ontological classification has come under close scrutiny. As we examine each kind of reason, we shall encounter the corresponding ontological debates. Before moving on, a word about the taxonomy of reasons. Although the orthodoxy is that there are two main kinds of reasons, normative and motivating-explanatory, it is plausible to argue that this twofold division is simplistic and that a three-partite division may be preferable. For there are at least three distinct questions about the relation between reasons and actions. There are questions about whether there are reasons that favour or justify someone’s action; questions about what reason(s) motivated or guided an agent in acting; and also questions about what reasons explain someone’s action. Given this, the account that follows proceeds by dividing reasons for action initially into the two familiar categories: normative and motivating-explanatory. It will then present the case for treating motivating and explanatory reasons separately.

3 Reasons that justify: normative reasons When does a reason justify an action? Let us return to the Supreme Court’s judgements on the lawfulness of the decision to prorogue Parliament for five weeks. The Court deemed that a decision to prorogue would be unlawful if it was done without justification, given the severe effects of prorogation on Parliamentary business. Since it judged that no adequate justification for proroguing had been put before the Court, it ruled the decision to be unlawful. But note that some reason for deciding to prorogue had been presented to the Court, namely, that prorogation would be ‘desirable’ in order to prepare the Queen’s Speech for a new session of Parliament. Still, the Court ruled that such a reason did not justify proroguing: it didn’t make it the right thing to do. First, because, normally, preparation for the Queen’s Speech requires no more than four to six days, so that reason justified a much shorter recess. And, second, because there was a weightier reason against proroguing for five weeks: the 404

Agency, reasons and rationality

need for parliamentary scrutiny of important legislation. So, on balance, the Court ruled that the reasons of the case did not justify a lengthy prorogation – with the result that the latter was declared unlawful. This illustrates the point that a normative reason may speak in favour of an action without justifying it, for example, if it is outweighed or in some other way defeated by other normative reasons that speak against the relevant action. Here, it helps to introduce the concept of a ‘pro-tanto’ reason, which may favour an action but not succeed in justifying it because it is defeated in one way or another by other reasons. But how should we understand the idea that reasons (speak in) favour (of ) actions? That is, in virtue of what precisely can reasons justify actions? One way to cash out the metaphor is in terms of goodness or value: on this view, ‘reasons are facts in virtue of which… actions are good in some respect and to some degree’ (Raz 1999, 23). Another is to ground the normativity of reasons on rationality (e.g. Korsgaard 1996; Smith 1994). A different proposal echoes Hume, claiming that the normativity of reasons is based on their relation to our desires. Roughly, whether there is a normative reason for someone to do something depends ultimately on their desires and motivations (see, e.g. Williams 1979, 1989). Each view faces challenges. For example, the neo-Humean view does well in explaining how normative reasons can motivate but less well on securing a sense of the objectivity of normative claims. By contrast, the goodness view explains the latter but is often seen incapable of bridging the gap between reason and motivation. Rationality views are faced with the problem of characterising the relation between ideally rational and real agents. The neo-Humean view just outlined, that is, that whether a normative reason applies to an agent depends on their motivations, is sometimes classified as ‘internalist.’ A different question is whether normative reasons apply to agents depending on their epistemic status. To see the issues, consider Othello’s plight in Shakespeare’s work of the same name. In the play, Othello kills Desdemona in the belief that she has been unfaithful to him. The tragedy is that she has not: Desdemona remains faithful to the end. Did Othello have a normative reason to kill Desdemona? On the one hand it seems clear that Othello had no normative reason to kill her since the reason he thinks he has – that she is unfaithful – is no reason at all. On the other hand, it might seem that Othello does have a normative reason, since he believes that Desdemona is unfaithful. And that belief appears to give him a reason to do what he does, at least from his perspective. It is not the point here to invoke the distinction between normative and motivating reasons, and say that Othello has no normative reason to kill Desdemona but has a motivating reason. The question is whether he as a normative reason. ‘Objectivists’ claim that whether an agent has an (objective) normative reason to act depends solely on the facts and not on the agent’s beliefs or knowledge of relevant facts. ‘Perspectivists’ disagree. Both objectivists and perspectivists may agree that all normative reasons are facts (or truths); their disagreement is about whether a fact’s being a normative reason for an agent to act depends on her epistemic status. Perspectivists often defend their position by reference to considerations of rationality. Agents who don’t know all the relevant facts or have mistaken beliefs about the situation often do what is reasonable or rational for them to do – given their perspective. If, as seems plausible, you act rationally when you act for reasons that make it rational for you to act as you do, perspectivism must be right: agents who act in error or ignorance act rationally if they act on the normative reasons they think they have; and normative reasons they are not aware of, or are not in a position to become aware of, are not normative reasons that apply to them. Similar arguments are articulated in relation to justification (though often questions 405

Maria Alvarez

about rationality and justification are run together). Accordingly, it is said that there are cases where an agent may well be justified in doing something even though there are conclusive reasons against doing it. And such an agent would be justified precisely because, since he is not (in a position to be) aware of those reasons, the reasons don’t bear on assessments about the justification of what he does. For example, the fact that there are valuable antiquities hidden in a cave may be a conclusive normative reason not to blow up the cave. But you might nonetheless be justified in blowing it up, the perspectivist may argue, if you have a justified belief that the cave is empty and have some good reason to blow it up – say, that it is necessary in order to build a much-needed hospital. It seems, then, that considerations about the justification of action support perspectivism because they show that whether you act rationally or are justified in what you do by the reasons that apply depends partly on your epistemic perspective. An objectivist can concede that an agent who acts guided by a false belief may act rationally, but deny that one acts rationally only if one acts in accordance with the normative reasons that apply. Instead, the objectivist may say, acting rationally only requires acting in a way that is consistent with one’s beliefs, so long as these are themselves rational (see, e.g. Parfit 2001; Kolodny 2005). In a similar vein, the objectivist can deny that the actions of agents who act in ignorance or guided by a mistaken belief are justified by the normative reasons that apply to her given her epistemic state. Whether the action is justified, the objectivist will say, depends purely on whether the facts make it the right thing to do, and not on the agent’s epistemic situation. So, in the example above, blowing the cave is not justified in the sense of being the right thing to do because there is a decisive normative reason against doing it: it will result in the destruction of invaluable objects. And that is so regardless of what the agent knows or believes. A different question, the objectivist will say, is whether an agent who does something that is not objectively justified but does it because of his false beliefs or ignorance is himself justified and/or blameworthy for so acting. If ignorance about the cave is not culpable, the agent may be justified and not blameworthy even though his action is not justified, because the justification of the agent may be relative to his epistemic status but the justification of the action is not. This example helps to bring out the point that the justification of agents, as distinct from the justification of their actions, depends largely on the reasons that motivate them to act, to which we now turn.

4 Motivating reasons, rationality and the explanation of action I said in the introduction that, when the UK Supreme Court declared the decision of the PM to prorogue Parliament unlawful without enquiring into the lawfulness of the PM’s motive, the distinction could be articulated in terms of two different questions about reasons: one is whether there was a justifying normative reason for the decision, and the other whether the PM’s motivating reason was a good reason. The semi-technical term ‘motivating reason’ is generally used in the literature to refer to a reason that the agent takes to favour her action, and in light of which she acts. These are reasons that figure as premises in the practical reasoning, if any, that leads to action. Because the concept is somewhat technical, further clarification is needed. First, the term so used does not apply to such things as someone’s goal or intention in acting, states of desiring, or emotions because they are not considerations that the agent takes to favour acting. For example, what the UK’s PM desired (to dissolve Parliament), his further goal (to get Brexit done), his state of desiring those things and his motive (say, love for his 406

Agency, reasons and rationality

country, or ambition) are all things that motivated him to decide to prorogue but they are not his motivating reasons as the term is currently used. Second, talk of an agent’s motivating reason, or of ‘the agent’s reason’ always involves some simplification because, first, an agent may be motivated to act by more than one reason; second, because facts are reasons only in combination with other facts (other reasons); and third, because the deliberation that leads me to act may include a fact that counts against doing something (a ‘con’-reason); if I then do that thing, I do not act for the ‘con’-reason, but the latter still plays a role in my deliberation. I mentioned above that the distinction between normative and motivating reasons was, for many years, more or less explicitly assumed to correspond to an ontological distinction. Until the turn of the 21st century, the assumption was that motivating reasons were psychological entities, in particular, mental states of agents, such as Othello’s believing that Desdemona is unfaithful to him. This view of the ontology of reasons is called ‘psychologism.’ Donald Davidson’s 1963 paper, ‘Actions, Reasons, and Causes’ is often cited as the locus classicus of psychologism. In that paper, he characterises a reason as follows: C1. R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent toward actions with a certain property, and a belief of the agent that A, under the description d, has that property. (1963, 687) On Davidson’s view, a primary reason is a combination of two mental states: a pro-attitude or desire, and a belief. When an agent acts for a reason, he acts motivated by an end that he desires (an end towards which he has a ‘pro-attitude’) and guided by a belief about how to achieve that end. It is, therefore, possible to explain why someone acted by citing what someone wanted and what they believed. This sort of consideration led to widespread acceptance of the view that the reasons that explain actions are mental states and, since the latter were not distinguished from motivating reasons, it also led to the view that motivating reasons too are mental states. That consensus began to dissolve at the turn of the century and psychologism came under sustained attack and various ‘non-psychologist’ positions were developed. Before examining that issue, we need to return to remarks about the classification of reasons I made in Section 2. I noted there that, although reasons are traditionally divided into two kinds – normative and motivating – explanatory – there may be a case for a threefold distinction: normative, motivating and explanatory reasons. There are, after all, three distinct questions about reasons: whether a reason favours an action (normative); whether a reason motivates an agent (motivating); and whether a reason explains an agent’s action (explanatory). This way of classifying reasons is explicitly accepted and/or defended by various authors (Baier 1958; Alvarez 2010), and it is hinted at, using different terminology, by others (Smith 1994; Darwall 2003). The distinction between the last two types may seem unnecessary: isn’t the reason that motivates someone also a reason that explains their action? Perhaps, but then a reason that motivates may also justify: if Val opened the window because the room was unbearably stuffy, the stuffiness of the room both motivates and justifies her action. But this doesn’t obliterate the distinction between motivating and normative reasons. Similarly, a reason that motivates may also explain but keeping them apart is helpful is helpful. Moreover, just as not every reason that justifies an action is the reason that motivated the agent (and vice-versa), similarly not every reason that explains why someone acted is among the reasons that motivated the agent. Let’s go back to Othello. In the play, Othello is motivated to kill Desdemona 407

Maria Alvarez

by the (putative) fact that she has been unfaithful. But what explains Othello’s action is not the (putative) fact that she has been unfaithful. Rather, what explains it is that, blinded by the jealousy that Iago so cunningly inflames, Othello comes to believe that Desdemona has been unfaithful and that her death is the only possible redress for such offense. So, the reason that motivates Othello to kill – the (putative) fact that Desdemona has been unfaithful – is different from the one that explains why he does it: the (actual) fact that Othello believes that she has. Given this, we need to separate also the question whether motivating reasons are mental states, and whether explanatory reasons are. I start with the first: whether psychologism is the right view about motivating reasons. An argument for non-psychologism for motivating reasons focuses on the premises of practical deliberation. For example, as Othello considers what to do, the premises of his reasoning do not include considerations about whether he believes this or that but rather considerations about what Desdemona has or has not done. The things that Othello considers, then, are not his mental states but rather facts, or alleged facts, about the world around him, in particular about Desdemona. And, in general, although considerations to the effect that one believes this or that may figure in one’s premises, this is rare, and much more often it is considerations about the world, about the value or goodness of things and people around us, the means of achieving those things and so forth that figure as premises in deliberation. This and other arguments have weakened the hold of psychologism. Non-psychologism is not free from difficulties, however. In particular, some think that the position fails in ‘error cases:’ where an agent is motivated to act by a false consideration. The problem is what, according to non-psychologism, is the reason that motivates agents: is it the content of her false belief, the fact that she holds that belief, or something else? The first position says that, in error cases, agents act for a reason that is a falsehood the agent believes (not her believing something false, which would bring us back to psychologism, but the false belief: the content; see, e.g. Dancy 2000, 2014). A difficulty here is that stating these alleged reasons often leads to paradox or infelicitous claims. Many think that a claim such as ‘Ellie’s reason for stepping on your toes is that you are stepping on her toes, although you are not stepping on her toes’ is paradoxical (Unger 1975, 208) and that, therefore, this response to the error cases – that a reason can be a falsehood – is problematic. An alternative response is that, in error cases, an agent acts on a consideration that he treats as a reason and in light of which he acts but which is in fact not a reason. So, in these cases an agent acts for an ‘apparent reason.’ Parfit, for example, characterises apparent reasons as follows: ‘We have some apparent reason when we have some belief whose truth would give us that reason’ (2001, 25; see also Alvarez 2010; Williamson 2017). On this view, agents who act on false beliefs are motivated by something, a false belief. These agents treat that belief as a reason and are guided by it in acting. Nonetheless, that false belief is not a motivating reason because it is not a fact; it is merely an apparent fact and, hence, merely an apparent reason. The issue behind these two views, beyond what might seem a terminological dispute, is whether the notion of a reason we apply in different contexts (of assessment, deliberation, or explanation) is a unified notion. If it is, the choice between the alternative non-psychological views just outlined will depend largely on what features are taken to be central to that notion, for example, whether truth or factivity are such features. Among non-psychologists, there is another issue that has aroused controversy concerning motivating reasons. Most, if not all, accounts of acting for a motivating reason require as a condition that the agent be in some kind of epistemic relation to the reason that motivates her. A widespread view, perhaps derived from Davidson’s conception of a primary reason, is 408

Agency, reasons and rationality

that this epistemic relation is one of belief: for an agent to act for the reason that p, the agent must merely believe that p. But it has been argued that, in order for a fact to be a reason that guides your actions, you need knowledge of the relevant fact (see, e.g. Unger 1975; Hyman 1999, 2015; Hornsby 2007; Williamson 2017). Without such knowledge, the claim goes, you cannot act in light of or guided by the relevant fact and so the fact cannot be your reason. These debates about motivating reasons focus primarily on what sort of thing motivating reasons are and what it takes for an agent to act for a reason. We now turn to when and how reasons explain actions. A person’s action may be explained in a variety of ways. For instance, we may say that the PM prorogued Parliament in order to avoid further legislation, or because he is reckless, or because it was the only way to get Brexit done. These statements explain because, given certain background assumptions, they show the action to be rational: they make it intelligible by citing a goal, a character trait and a normative reason, respectively. But only the third explanation gives a reason why he did it that was his reason for doing it: a reason that, from his perspective, spoke in favour of proroguing and that, hence, rationalises his decision. In Davidson’s words: A reason rationalizes an action only if it leads us to see something the agent saw, or thought he saw, in his action – some feature, consequence, or aspect of the action the agent wanted, desired, prized, held dear, thought dutiful, beneficial, obligatory, or agreeable. (Davidson 1963, 685) Psychologism is a plausible account of the reasons that explain actions because, for a reason to be able to rationalise your action, that reason must be part of your psychology: a fact that is merely ‘out there’ cannot explain why you did anything. Your believing or knowing that fact, by contrast, can explain why you acted. True, explanations of actions often don’t mention mental states but, it may be argued, that is because those explanations are elliptical and, when fully spelled out, their explanans (the part of the explanation that does the explaining) contain facts about what the agent knew or believed. Whether we accept that rationalisations in general must involve mental states or not, it seems that we must accept that they do in ‘error cases’ because all explanations are factive: a true explanation cannot have a falsehood as its explanans (but see Dancy 2000). To illustrate, the reason that explains why Andy punched James cannot be that James punched Andy first if James in fact did not punch Andy at all. Here we see again why it is helpful to distinguish motivating and explanatory reasons, for the reason that explains why Andy punched James may well be the fact that Andy thinks James punched him first; but what motivated Andy is not that fact but rather the (mistaken) consideration that James punched him first.

5 Conclusion The chapter has focused on practical reasons but limitations of space make omission of important debates inevitable. Before closing, however, I shall mention an issue that has gained some prominence recently. The preceding discussion relied on the assumption that our reasons for acting are in principle accessible to us as agents. It is true that we often act unthinkingly but even then we are typically capable of explaining our actions by citing the reasons, if any, that, perhaps implicitly, guided our behaviour. This does not mean that we are infallible when it comes to identifying our reasons: lack of self-knowledge, self-deception 409

Maria Alvarez

and ‘Freudian’ motivations may lead us astray. Nonetheless, we have a kind of authority over our  reasons. Recently, however, work in experimental psychology that purports to show the prevalence of epistemic biases has led some to challenge the idea that we can know our ‘real’ reasons for acting, and to argue that we in fact largely ‘confabulate’ when explaining and attempting to justify our behaviour (see Hirstein 2009). If right, this would appear to threaten that authority over our reasons, as well as the explanatory power of the ordinary explanations of action that cite the agent’s reasons for acting. The plausibility of these conclusions depends to a large extent on whether the notion of ‘the agent’s real reason’ at issue here is the same as the notion of a motivating reason studied by philosophers – an issue that I cannot explore here. I hope, however, that this survey suffices to show the extent to which our understanding of ourselves as rational agents is shaped and complicated by our understanding of practical reasons.

Related topics Rational agency; Agency and normativity; Agency and practical reasoning; Agency and free will.

Further reading Alvarez, M., 2010, Kinds of Reasons: An Essay on the Philosophy of Action, Oxford: Oxford University Press. A detailed study of various kinds of reasons and their role in justifying, motivating and explaining action. Dancy, J., 2000, Practical Reality, Oxford: Clarendon Press. A spirited defence of non-psychologism about reasons based on the connection between the reasons why we do things and the reasons why we should. Davidson, D., 1963, ‘Actions, Reasons, and Causes’, The Journal of Philosophy, 60 (23): 685–700. The locus classicus for the claim that the reasons that explain an action are mental states of agents that also cause the action Parfit, D., 1997, ‘Reasons and Motivation’, Proceedings of the Aristotelian Society, Supplementary Volume, 71: 99–129. A helpful introduction to the distinction between normative and motivating reasons, and to some issues concerning their relation to each other. Raz, J., 1999, Engaging Reason: On the Theory of Value and Action, Oxford: Oxford University Press. An examination of central themes about reason, thought and action, such as the nature of normativity, reason, the will and value.

References Alvarez, M., 2010, Kinds of Reasons: An Essay on the Philosophy of Action, Oxford: Oxford University Press. Audi, R., 2001, The Architecture of Reason, Oxford: Oxford University Press. Baier, K., 1958, The Moral Point of View, Ithaca, NY: Cornell University Press. Dancy, J., 2000, Practical Reality, Oxford: Clarendon Press. ———, 2014, ‘On Knowing One’s Own Reasons’, in Littlejohn and Turri (eds.), 2014: 81–96. Darwall, S., 2003, ‘Desires, Reasons and Causes’, Philosophy and Phenomenological Research, 67: 435–443. Davidson, D., 1963, ‘Actions, Reasons, and Causes’, The Journal of Philosophy, 60 (23): 685–700; repr. in his 1980, Essays on Actions and Events, Oxford: Clarendon Press: 3–21. Hirstein, B., 2009, ‘Confabulation’, in The Oxford Companion to Consciousness, T. Bayne, A. Cleeremans, and P. Wilken (eds.), Oxford: Oxford University Press: 174–177. Hornsby, J., 2007, ‘Knowledge, Belief, and Reasons for Acting’, in Explaining the Mental, C. Penco, M. Beaney, M. Vignolo (eds.), Newcastle: Cambridge Scholars Publishing: 88–105. Hyman, J., 1999, ‘How Knowledge Works’, Philosophical Quarterly, 49 (197): 433–451. ———, 2015, Action, Knowledge, and Will, Oxford: Oxford University Press. Kolodny, N., 2005, ‘Why Be Rational?’ Mind, 114 (455): 509–563.

410

Agency, reasons and rationality Korsgaard, C., 1996, The Sources of Normativity, Cambridge: Cambridge University Press. Littlejohn and J. Turri (eds.), 2014, Epistemic Norms, New Essays on Action, Belief and Assertion, Oxford: Oxford University Press. Mele, A., 2003, Motivation and Agency, Oxford: Oxford University Press. Parfit, D., 1997, ‘Reasons and Motivation’, Proceedings of the Aristotelian Society, Supplementary Volume, 71: 99–129. ———, 2001, ‘Rationality and Reasons’, in Exploring Practical Philosophy: From Action to Values, D. Egonsson, J. Josefsson, B. Petersson and T. Rønnow-Rasmussen (eds.), Ashgate: 19–39. Raz, J., 1975, Practical Reasoning and Norms, London: Hutchinson & Co., reprinted, Oxford University Press, 1990 and 1999. ——— 1999, Engaging Reason: On the Theory of Value and Action, Oxford: Oxford University Press. Smith, M., 1994, The Moral Problem. Oxford: Blackwell. The Supreme Court, [2019] UKSC 41. JUDGMENT. R (on the application of Miller) (Appellant) v The Prime Minister (Respondent)Cherry and others (Respondents) v Advocate General for Scotland (Appellant) (Scotland). Unger, P., 1975, Ignorance. A Case for Scepticism, Oxford: Clarendon Press. Williams, B.A.O, 1979, ‘Internal and External Reasons’, reprinted in his 1981, Moral Luck, Cambridge: Cambridge University Press: 101–113. ———, 1989, ‘Internal Reasons and the Obscurity of Blame’, in reprinted in his 1995, Making Sense of Humanity, Cambridge: Cambridge University Press: 35–45. Williamson, T., 2017, ‘Acting on Knowledge’, in J.A. Carter, E. Gordon, and B. Jarvis (eds.), Knowledge-First, Oxford: Oxford University Press: 163–181.

411

37 AGENCY AND PRACTICAL REASONING Jules Salomone-Sehr and Jennifer M. Morton

Unlike other ways of coming to act, for example as a result of habit or impulse, reasoning imprints our actions with the distinctive mark of rational full-blooded agency. Most mornings, many of us wake up and make coffee without giving it much thought. Coffee making is responsive to reasons we have (caffeine helps us kick-start our day), but it is merely habitual. Alternatively, we could deliberate about whether we want to start our day with a steaming cup of coffee or a healthy green juice, consider the reasons we have, and then make coffee. In this alternative scenario, we do something other than simply respond habitually to the reasons we have. But what does this bit of deliberative thinking add to the agent’s arsenal? In order to answer this question, we need to think more carefully about what practical deliberation consists in.

1 Four adequacy criteria In this section, we lay out four basic criteria—mentality, evaluation, practicality, and attributability—that an account of practical reasoning ought to satisfy in order to capture essential features of the phenomenon. First, practical reasoning is by and large a conscious mental process—call this the mentality criterion. More specifically, any episode of practical reasoning involves a temporally extended sequence of mental attitudes of which the reasoner is aware. Not every aspect of practical reasoning need be wholly mental. The mentality criterion does not stipulate away views on which practical deliberation culminates in action rather than intention, that is, in something that is at least partly physical (Dancy 2018). Similarly, practical reasoners need not be conscious of every constituent of a deliberative episode: I might know I am applying a rule of practical reasoning without having in full view a statement of the rule. The mentality criterion does set a limit to the amount of non-mental and/or sub-personal constituents of practical reasoning that there might be. Obviously, processes that take up no psychological space whatsoever (like the nervous process initiated by a stimulus that results in a reflex action) cannot count as deliberative. More important, cognitive yet sub-personal processes (like the routine production of natural language sentences that satisfy grammar rules) cannot count as deliberative either.

412

DOI: 10.4324/9780429202131-46

Agency and practical reasoning

Interestingly, the mentality criterion doesn’t prevent practical deliberation from being distributed intra- and interpersonally. A deliberative process need not occur in succession. It can be interrupted (voluntarily or involuntarily), resumed, and led to completion minutes or hours after it started. And if we are to allow for shared practical reasoning, then it must be the case that the mental process in question can take place in the minds of different practical reasoners. Turning to the second criterion, practical reasoning is a process that can go better or worse. Our deliberations can be rational, effective, cautious, impulsive, prudent, inconclusive, and so on. In sum, practical reasoning is answerable to a range of evaluative standards— call this the evaluation criterion. An implication of this is that episodes of thought that do not lend themselves to appraisal (like free association and daydreaming) cannot be counted as deliberative. The evaluation criterion remains silent on the standards of good reasoning. It also leaves open the question whether we have reason to engage in good reasoning and whether we deserve criticism for reasoning badly. Third, practical reasoning is primarily concerned with settling what to do—call this the practicality criterion. The practicality criterion does not require that there be a clear-cut distinction between practical and theoretical reasoning, though an account that takes these both to be exercises of the same faculty owes us an explanation of why reasoning leads to action in some cases and not others. Nor does this criterion prevent episodes of deliberation from including belief or episodes of theoretical reasoning. Premises of practical reasoning surely include—hopefully true—beliefs about the context of decision, for otherwise the subsequent reasoning would get off track. And, on some views, episodes of practical reasoning culminate in a belief about what one ought to do (Davidson 2001b, 99). Still, the practicality criterion sets limits on the role that attitudes that aim to represent the world (rather than change it) can play. In particular, an episode of reasoning whose conclusion has no practical import whatsoever cannot count as practical. Finally, practical reasoning is something that a reasoner does—call this the attributability criterion. As opposed to drifting in thought and other non-agential cognitive processes, practical reasoning does not just happen to us, but is an activity that we engage in, and that we can be evaluated for according to how well we satisfy the relevant standards. This is not to say that all the steps of an episode of practical reasoning are things that the reasoner does directly. To illustrate, suppose that in reasoning about how to go to Coney Island this Saturday, an agent suddenly remembers that the F train is disrupted on weekends. Though being disposed to remember facts relevant to going to Coney Island is something she sets out to do, having this particular memory about the defective F train is not something she does directly.1 Still, the attributability criterion sets limits on the extent to which practical reasoning might involve unthinking and automatic processes that the agent does not direct herself. In particular, it requires that the arriving at the decision that concludes an episode of practical reasoning be something that the agent does. Adapting Nomy Arpaly and Timothy Schroeder’s example, Maria might often make decisions by entering ‘a state of meditative tranquillity’ (Arpaly and Schroeder 2012, 211–212) during which she often resolves practical problems that concern her. But this process does not count as deliberative insofar as the decisions Maria arrives at merely occur to her. Bearing these four criteria in mind is useful, if only to not be caught off guard when a purported account of reasoning is in fact an account of responding to and acting for reasons, or an account of rationality.2 As we will see, many accounts do satisfy these criteria, but do so in vastly different ways.

413

Jules Salomone-Sehr and Jennifer M. Morton

2 John Broome’s account John Broome’s theory of reasoning—both practical and theoretical—is the most systematic account defended in the literature (2013). He claims that reasoning is the rule-guided activity by which we bring our mental attitudes to satisfy requirements of rationality. In the practical case, these requirements primarily apply to our intentions. As we will see, Broome’s account easily satisfies the four criteria we laid out in the previous section. Like morality, prudence, and the law, rationality is the source of certain distinctive requirements. Broome takes such requirements to bear on our mental attitudes. On his view, we are rational just when our attitudes are consistent, coherent, and stand in the right sorts of relation with each other. Put differently, rationality is a kind of psychic tidiness (Kolodny 2008a). To illustrate, rationality requires that neither our beliefs nor our intentions contradict one another (Broome 2013, 156). It also requires means-end coherence, roughly, that if you intend some end and believe that some means is necessary to its achievement, then you intend to take that means (159).3 More controversially, Broome thinks that rationality requires enkrasia; that is, if you believe you ought to perform some action φ, then you intend to φ (171). These three requirements are synchronic in the sense that they place constraints on attitudes that a reasoner has at a given time. Other requirements are diachronic. That is, they place demands on the attitudes we have over time. To give an example, rationality prohibits me from dropping the belief that I ought to φ based on my having no intention to φ. Every move that is not barred by a prohibition is permitted. For instance, it is permitted to form the intention to take a necessary means to some end based on the intention to achieve that end, and the belief that that end requires the taking of that means. These observations clarify how reasoning is supposed to help us satisfy rational requirements. Perhaps your attitudes are in synchronic conflict. In that case, reasoning might help you to resolve that conflict by dropping and/or forming new attitudes in accordance with rational permissions. Alternatively, your attitudes might already satisfy synchronic requirements, but a project of yours that you would like to bring to completion, or perhaps some unexpected circumstance might call for some adjustments in what you believe and intend. Reasoning might help you make these adjustments in such a way that you neither break any diachronic requirements nor end up with a collection of synchronically conflicting attitudes. However, in many cases we satisfy requirements of rationality through automatic processes that do not require much thinking from us, if any at all.4 And sometimes we face practical conundrums that are best solved by a good night’s sleep. What distinguishes these ways of satisfying rational requirements from reasoning? The answer for Broome is that ‘reasoning is a rule-governed operation on the contents of your conscious attitudes’ (2013, 234). This summary requires parsing. First, in contrast with automatic processes, Broome takes reasoning to be a conscious mental operation. His account thereby straightforwardly satisfies the mentality criterion. Second, he takes rule-following to be essential to reasoning. More specifically, an episode of thought counts as deliberative so long as we form and revise consciously attitudes in light of rules. The account, thus, satisfies the evaluation criterion. For on his account, reasoning can go better or worse depending on whether we correctly follow correct rules, that is, rules that do not run afoul of rational prohibitions. Additionally, as Broome takes rule-following to be something that we do, his account would seem to satisfy the attributability criterion. Finally, he believes that practical deliberation results in the formation of intentions. Thereby, Broome’s account also satisfies the practicality criterion. 414

Agency and practical reasoning

Broome’s influential position has been subject to several challenges. In what remains we focus on a few that highlight current, fruitful areas of inquiry for philosophers working on practical reasoning.

3 What do we reason from? Broome analyses reasoning as a rule-governed operation on the contents of our attitudes, not on the attitudes themselves—though the result of this operation is a revision of our attitudes. In doing so, he rejects a higher-order picture of reasoning on which reasoners first become aware of which attitudes they have, then realize that in order to satisfy rational requirements, they ought to revise their attitudes, and then proceed to do so accordingly. Broome rejects this picture because it requires more cognitive sophistication and self-consciousness than he takes everyday reasoning to exhibit. On Broome’s view, we typically do not deliberate on our attitudes qua attitudes in bringing them in conformity with rational requirements. Instead, we reason at the first-order level, that is, from the contents of our attitudes.5 Broome’s view does not just bear on what deliberation consists in. It also takes a stance on the kind of things that we can deliberate from—call these the inputs of reasoning. And insofar as he holds that reasoning operates on the contents of our attitudes, he would seem, rather oddly, to bar explicitly autobiographical considerations (e.g., ‘I want to have nice food with you’, ‘I intend to become a professional philosopher’) from serving as inputs of practical deliberation. For if they served as such inputs, reasoners would reason from their attitudes (e.g., their desire for sharing food, their intention to become professional philosophers) rather than simply the contents of those attitudes. But this type of reasoning has little place in Broome’s first-order account.6 This is a rather surprising conclusion given how pervasive autobiographical premises seem to be. After all, an agent’s desire to share adventurous food with friends seems as plausible a premise of practical deliberation as any; and in fact, a premise of that sort has led many an agent to look for restaurants that serve exciting food. Additionally, it is not entirely clear how to make sense of the idea that we reason from the contents of desires and intentions. Interestingly, Broome does not give examples of practical but theoretical reasoning when discussing the respective merits of first- and higher-order accounts of reasoning. Using his example (2013, p. 216), suppose you believe that it is raining, and that if it is raining, then the snow will melt. When reasoning toward your belief that the snow will melt using modus ponens, it is pretty clear that you attend to the contents of your premise-beliefs (that is, to the snow), and not to your believing such contents. But the practical case seems different: when looking for a lovely restaurant for an evening out, don’t agents sometimes attend quite directly to their desires? Nonetheless, one might have genuine worries about autobiographical premises. When we engage in practical deliberation, we take, albeit implicitly, our practical conclusions to be supported by the considerations we have reasoned from. As a result, if autobiographical considerations were good candidates for the role of premises of reasoning, it would follow that reasoners would take them to favour the conclusions they drew. But what if autobiographical considerations turned out devoid of reason-giving force? T. M. Scanlon (1998 43–44) has offered an argument in favour of a rather radical version of this point. On his view, desires rarely, if ever, supply reasons for action.7 For instance, he claims that an agent’s desire to buy a new computer has no reason-giving force over and above the value there is in owning a new computer. To elaborate, this agent has no reason to buy a new computer just for the sake of satisfying a desire; rather, whatever reasons she has for this purchase lie in the defects of her ageing laptop, the costs of fixing the old relative to buying a new one, and so on. In sum, 415

Jules Salomone-Sehr and Jennifer M. Morton

desires qua desires do not make a difference to the reasons that a reasoner has. If Scanlon were right, then considerations of the sort ‘I want x’ or ‘I’d like to φ’ would not be apt inputs for practical reasoning. Autobiographical premises face another worry. If we grant that an intention or desire can serve as good premises from which to reason, then such autobiographical premises seem to open the floodgates of what philosophers refer to as bootstrapping. To illustrate, suppose Sam intends to kill his boss, and suppose that that consideration is one of his premises in an episode of practical reasoning aimed at identifying the most effective and least risky means to satisfy his murderous desire. Under the assumption that deliberating well involves taking one’s intention for an end to give one reason to intend the means, then deliberating well would seem to involve concluding that Sam has reason to use an undetectable lethal poison—a conclusion that conflicts with our intuition that intending to kill one’s boss cannot bootstrap a reason to poison her (Bratman 1981, 1987, 24–27). One response to all these problems is to conclude that reasoners do not deliberate from their desires and intentions, but from the desirability of certain actions or states of affairs (Dancy 2018; Scanlon 1998). On views of this kind, I decide to go see a movie because good reviews recommend it. But one advantage of autobiographical premises over views of this sort lies in their straightforward explanation of the motivational pull of the conclusions we arrive at through practical deliberation: roughly, the picture is one on which the motivation expressed in an autobiographical premise—usually, desire or intention—gets transmitted to the conclusion of the deliberation. The agent is motivated to buy a movie ticket because she sees this as a means to something she is already motivated to do—see a good movie by her favourite director. In effect, these views satisfy the practicality and attributability criteria by taking reasoning to be an activity that is driven by the agent’s motivational attitudes. By contrast, views on which the valuable features of certain states of affairs serve as inputs of practical reasoning will have to argue that the motivational attraction of the conclusions of practical reasoning need not trace back to antecedent motivational states but rather to the goodness of the reviews.8 On these views, theoretical and practical reasoning are not as far from each other as it might initially seem, but this feature leads these accounts to face an additional hurdle in satisfying the practicality criterion.

4 What do we reason towards? As said earlier, Broome believes that practical reasoning concludes with the formation of an intention. This premise, which many have accepted, has also come under scrutiny recently, particularly by Jonathan Dancy. Dancy defends the Aristotelian thought that practical reasoning concludes in action (see also Ford 2018). At first, this position might seem like a nonstarter. After all, we often do not succeed in acting through no fault of our own. Our limbs fail to cooperate or the phone rings just as we are about to head out the door. In these cases, we seem to have done all the deliberating we could have done without yet having acted. But Dancy denies that a theory of practical reasoning should offer necessary and sufficient conditions for practical reasoning, suggesting that a better approach is to offer a theory that can account for paradigmatic cases (the focal cases, as he terms them). He points out that in these focal cases there is little space between deliberating about what to do and action: we consider the factors at play and off we go. Even when presented in its best light, Dancy’s account faces a couple of challenges. First, it is in relative tension with the mentality criterion, that is, the requirement that practical reasoning is a psychological process. In contrast with beliefs and intentions, actions are not 416

Agency and practical reasoning

wholly mental entities.9 As a result, should they be a part of the reasoning process, they would introduce a jarring physical constituent in a by and large mental process. In response, Dancy observes that actions, especially intentional ones, are not merely physical: insofar as they are shaped by the understanding of the agent, they have a certain sort of mentality. And this mentality in turn explains how actions might count as the output of deliberation despite their physical aspects. But Dancy’s view about the mentality of actions is a substantive, controversial view in action theory—for a helpful contrast, see Donald Davidson’s view that many actions (in fact, all the actions he calls primitive) are no more than bodily movements (Davidson 2001a, 49). Moreover, practical reasoning is very often devoted to thinking about what we will do in the future. We deliberate about where to go on vacation or what college to attend months in advance of taking any steps towards those ends (Bratman 1987). A view on which practical reasoning leads straightforwardly to action, like Dancy’s, seems ill-suited to account for such cases. Yet, these cases do not seem less paradigmatic, but rather an important exercise of practical reasoning for agents for whom planning is a critical capacity. We depend on each other’s capacity to plan ahead and hold each other accountable for failing to deliberate ahead of time. Good practical reasoning seems to concern both reasoning that results in action and reasoning that results in attitudes that stop short of action. Intentions have seemed, to many philosophers, natural intermediate attitudes involved in both exercises of reasoning. Hence the traditional view that when deliberating, we primarily reason towards an intention. However, this traditional view has its shortcomings. On this view, the gap between intention and action remains either mysterious or simply a matter of causal contingency (Velleman 2000). Yet, it seems that there is a distinctive kind of failure of agency when one reaches a conclusion about what to do and yet fails to move one’s body in order to carry out that intention (putting aside cases in which we are prevented from executing our intentions by outside forces). Sometimes, though capable of acting, we remain sitting on the sofa and the time for action passes us by. On the traditional account, this isn’t a failure of reasoning, but what kind of failure is it then?

5 What is good reasoning? Why care? We end with a discussion of good reasoning and its normative status. On Broome’s view, we reason well just in case we correctly follow correct rules, that is, rules licensed by rational permissions. We observed earlier that this view satisfies the evaluation criterion: unlike daydreaming, reasoning can go better or worse depending on whether the rules I follow are correct, and whether I follow them well. However, some have taken issue with the onoff evaluations of one’s rationality that Broome’s view induces (Fogal 2020; McHugh and Way 2018). On Broome’s requirement-based view, either you satisfy the rational requirements, thereby counting as rational, or you do not, thereby counting as irrational. But pace Broome, rationality might come in degrees. Consider again the enkratic requirement against believing you ought to φ while failing to intend to φ. Adapting John Brunero’s (2013) example, suppose you have the deeply entrenched normative belief that you ought to become a nurse. Furthermore, though entrenched (perhaps your parents urged you to embark on that career), this normative belief conflicts with a lot of your other beliefs and desires (especially your desire to become a history teacher). Yet, you cannot quite get yourself to intend to become a nurse. Though your attitudes do not bear the right sort of relations to one another, you seem more rational than someone who has the same belief, but does not experience a conflict with his other attitudes and fails to form the intention to become a nurse out of weakness of will. 417

Jules Salomone-Sehr and Jennifer M. Morton

Examples like these give rise to the question of why an agent ought to care about having a tidy psychic economy. After all, embarking on the career favoured by the weightiest reasons might be all that matters, and perhaps becoming a history teacher is truly what you should do—and what you will do despite your messy psychic economy. Some think that psychic tidiness is worth caring about. Michael Bratman (2009, 2012) has argued that agents whose intentions are inconsistent or incoherent have no place where they stand with respect to a given practical issue. Similarly, Christine Korsgaard (2009) suggests that this kind of internal inconsistency undermines the constitution of one’s agency. Yet, an agent’s self-governance or self-constitution does not seem at stake in all cases of attitudinal inconsistency. DeVonn intends to have an espresso with breakfast and also intends to skip their caffeine fix this morning. This inconsistency seems pretty inconsequential. Thus, it certainly does not threaten their agency in any fundamental way, even though they are to some extent practically irrational. Additionally, the quest for psychic tidiness might sometimes stand in the way of doing what we have the weightiest reasons to do, as in the aforementioned conflict between career paths. This has led philosophers like Niko Kolodny (2005, 2008a, 2008b) and Joseph Raz (2005) to adopt a myth theory about the norms of practical reasoning. For Kolodny, the norms of practical reasoning are diagnostic insofar as their violation tells us that we have gone astray somewhere. But he thinks that inconsistency or incoherence are not inherently what we should worry about. The goal of reasoning is the same goal as that of acting—responding to the reasons we have. Practical reasoning carries no special normativity over and above that of reasons. Myth theorists invite piecemeal approaches to reasoning on which practical deliberation is good when it tracks the reasons we have, quite independently from whether it satisfies or runs afoul of rules of reasoning. But these approaches face difficulties too. After all, some patterns of reasoning seem good even when they sometimes lead us to practical responses that no reason favours. Suppose you intend to eat a jawbreaker, and believe that the round thing within arm’s reach is a jawbreaker, when in fact it is just a marble. Surely you have no good reason to eat a marble, but the piece of means-ends coherent reasoning that led you to intend to eat one seems as good as any. Practical reasoning is a useful tool in the agent’s arsenal. We often turn to it when decisions are difficult, important, or complex. Of course, sometimes agents just act out of whims or impulse. Doing so can be more appropriate in some cases than deliberating (Arpaly and Schroeder 2012). But to deny that practical reasoning has distinct normative standards is to fail to recognize its unique role in our agential lives. In particular, denying the normative standards of practical reasoning leads us to ignore the central role that our evaluation of each other’s rationality plays in our social lives, including in the evaluation of our behavioural responses to social policy. We hold agents accountable for failing to take advantage of certain resources (for example, medical-savings accounts or food stamps) when we criticize their actions as irrational. Yet, we have some evidence (Mullainathan and Shafir 2013) that agents who are deliberating in contexts of severe scarcity might be reasoning differently from agents in contexts of moderate scarcity. One of us has argued (Morton 2010, 2017) based on this empirical research that theories of rationality have not adequately considered the possibility that standards of deliberation might be different for agents in contexts of poverty. In fact, given all the interesting research done by cognitive scientists and behavioural economists on reasoning, it is important philosophers take into account what this research shows in offering a theory of practical reasoning that does justice to the role reasoning concretely plays in our lives. 418

Agency and practical reasoning

Related topics Rational agency; Planning agency; Agency and reasons; Agency and normativity.

Notes 1 See Arpaly and Schroeder for a similar example (2012, 213–214). 2 To illustrate, as usefully noticed by Broome (2013, 208), Gilbert Harman’s Change in View: Principles of Reasoning does not so much offer an account of reasoning as an account of rationality. 3 According to Broome, these requirements are wide-scope. For instance, one can satisfy meansend coherence by either giving up the intended end or by taking the intended means. This has given rise to a literature on wide-scope vs. narrow-scope, see Broome (2007), Schroeder (2004) and Way (2010). 4 Broome’s view is that good reasoning is of no help when we are in the grip of two contradictory beliefs. According to him, only automatic processes can resolve this specific sort of psychic untidiness (2013, 279–280). 5 For more details on this, see Broome’s discussion of ‘marked contents’ in (2013, 251–252). 6 Broome does not exclude higher-order reasoning altogether but gives it a rather limited role: on occasions, it might influence the direction of first-order reasoning (2013, 245–246). 7 Scanlon concedes that unlike a reasoner’s desires, their intentions can make a difference to the reasons they have. 8 For a view along these lines, see Joseph Raz’s discussion of his facilitative principle (2005). 9 Perhaps with the exception of mental actions.

Further reading Arpaly, N., and Schroeder, T. (2012). Deliberation and Acting for Reasons. Philosophical Review, 121(2), 209–239. They argue against the widespread claim that reasoning is crucial to responding to reasons. Instead, they show that reasoning is but one—relatively powerful—tool that allows us to track the reasons we have. Bratman, M. (1987). Intention, Plans, and Practical Reason. Cambridge: Cambridge University Press. A central book, not just for the novel and groundbreaking account of intentions, but also for the claim that intentions are not just outputs, but also distinctive inputs of practical reasoning. Chapter 3 is central. Broome, J. (2013). Rationality through Reasoning. Wiley-Blackwell. This book offers the most systematic account of reasoning to date. Readers should prioritize chapters 9, 10, 13, and 14. Korsgaard, C. M. (1997). The Normativity of Instrumental Reason, in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason, Oxford: Oxford University Press, pp. 215–254. Korsgaard lays out a challenge to instrumentalism that animated much of the subsequent debate. Raz, J. (2005). The Myth of Instrumental Rationality. Journal of Ethics & Social Philosophy, 1(1), 1–28. A key paper in laying out the rejection of distinctive normativity of rational requirements.

References Arpaly, N., and Schroeder, T. (2012). Deliberation and Acting for Reasons. Philosophical Review, 121(2): 209–239. Bratman, M. (1981). Intention and Means-End. Philosophical Review, 90(2): 252–265. Bratman, M. (1987). Intention, Plans, and Practical Reason. Cambridge: Cambridge University Press. Bratman, M. (2009). Intention, Practical Rationality, and Self-governance. Ethics, 119(3): 411–443. Bratman, M. (2012). Time, Rationality, and Self-Governance. Philosophical Issues, 22(1): 73–88. Broome, J. (2007). Wide or Narrow Scope. Mind, 116: 359–370. Broome, J. (2013). Rationality through Reasoning. Hoboken, NJ: Wiley-Blackwell. Brunero, J. (2013). Rational Akrasia. Organon F: Medzinárodný Časopis Pre Analytickú Filozofiu, 20(4): 546–566.

419

Jules Salomone-Sehr and Jennifer M. Morton Dancy, J. (2018). Practical Shape: A Theory of Practical Reasoning. Oxford: Oxford University Press. Davidson, D. (2001a). Agency. In Essays on Actions and Events. Oxford: Clarendon Press, pp. 43–62. Davidson, D. (2001b). Intending. In Essays on Actions and Events. Oxford: Clarendon Press, pp. 83–102. Fogal, D. (2020). Rational Requirements and the Primacy of Pressure. Mind, 129(516): 1033–1070. Ford, A. (2018). The Province of Human Agency. Noûs, 52(3): 697–672. Harman, G. (1986). Change in View: Principles of Reasoning. Cambridge, MA: MIT Press. Kolodny, N. (2005). Why Be Rational? Mind, 114(455): 509–563. Kolodny, N. (2008a). The Myth of Practical Consistency. European Journal of Philosophy, 16(3): 366–402. Kolodny, N. (2008b). Why Be Disposed to Be Coherent? Ethics, 118(3): 437–463. Korsgaard, C. M. (1997). The Normativity of Instrumental Reason, in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason, Oxford: Oxford University Press, pp. 215–254. Korsgaard, C. M. (2009). Self-Constitution: Agency, Identity, and Integrity. New York: Oxford University Press. McHugh, C., and Way, J. (2018). What Is Good Reasoning? Philosophy and Phenomenological Research, 96(1): 153–174. Morton, J. M. (2010). Toward an Ecological Theory of The Norms of Practical Deliberation. European Journal of Philosophy, 19(4): 561–584. Morton, J. M. (2017). Reasoning under Scarcity. Australasian Journal of Philosophy, 95(3): 543–559. Mullainathan, S., and Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. New York: Henry Holt and Company. Raz, J. (2005). The Myth of Instrumental Rationality. Journal of Ethics & Social Philosophy, 1(1): 1–28. Scanlon, T. (1998). What We Owe to Each Other. Cambridge, MA: Belknap Press of Harvard University Press. Schroeder, M. (2004). The Scope of Instrumental Reason. Philosophical Perspectives, 18: 337–364. Velleman, J. D. (2000). What Happens When Someone Acts? The Possibility of Practical Reason. Oxford: Oxford University Press, pp. 123–143. Way, J. (2010). Defending the Wide-scope Approach to Instrumental Reason. Philosophical Studies, 147(2): 213–233.

420

38 AGENCY AND NORMATIVITY Kenneth Walden

1 From agency to reasons Deliberation involves distinctive questions. “Do I have any reason to stop by the store on my way home?” “Should I order the men to come to starboard?” “Do we as a society have an obligation to expropriate the wealth of billionaires?” These questions have two things in common. First, they employ normative vocabulary, words like “reason,” “should,” and “obligation”. Second, they presuppose that their subject can do something—stop by the store, order sailors steer the ship, expropriate the wealth of billionaires. We can tell that this is a presupposition because the questions can be rendered senseless by their denial. If you tell me the store burned down, my question begins to sound unfounded. The capacity to do things is agency. So these questions presuppose the agency of their subjects. More precisely, they presuppose different kinds of agency. The first presupposes something like the physical, spatiotemporal agency required to move the mass of my body from one place to another. The second presupposes a social agency associated with occupying an office in a hierarchical structure. And the third presupposes the collective agency of a society or a state. Some might balk at the idea that there are distinctive kinds of agency apart from the maximally generic capacity for action. But I think this is backward. I have a better grasp of the narrower notions. I can understand the agency characteristic of certain stations and sorts of people—the navy captain, the mid-level bureaucrat, the peasant in medieval France—by thinking about the distinctive actions performed by these individuals, the forces that enable, resist, guide, and otherwise affect those actions, and the specific capacities and attunements that are called upon on in their performance. Insofar as I understand agency tout court, it’s actually by starting with these more concrete specimens and abstracting.1 These forms of agency come with distinctive features. When I deliberate about whether to bring the ship to starboard, I presuppose what we could call “captain’s agency”: a power to do things derived from my occupying a particular node in the social structure that constitutes a ship’s crew. This kind of agency obviously consists in something more than the ability to order sailors to bring the ship to starboard. It entails other powers, like the power to review subordinates’ orders. It arguably includes characteristic aims, like the aim of commanding

DOI: 10.4324/9780429202131-47421

Kenneth Walden

sailors. And it involves responsibilities, like adjudicating complaints from those subordinates and communicating with the admiralty. Exercising captain’s agency requires attention to these responsibilities because they are part of the office of captain, and captain’s agency just is the agency characteristic of a captain. Of course, ignoring one dispatch from the admiralty does not rob one of the ability to give orders. But continued neglect of these features may render a person an ex-captain—either causally (mutiny) or constitutively (a transition to Head Pirate). Let’s call these things constitutive features of captain’s agency. We have seen, then, that in deliberation we ask normative questions, that these questions presuppose distinctive forms of agency, and that those forms of agency are characterized by certain constitutive features. Can we infer anything interesting from this? One thing we might be tempted to say is that this arrangement somehow endows the relevant constitutive features with normative authority for the deliberating agent. There are a few ways we might try to establish this claim. First argument. Suppose that amongst the constitutive features of captain’s agency is the aim to command sailors. This means that any time I enter into deliberations that presuppose captain’s agency, I am also presupposing that I have this aim. So, for the purposes of those deliberations, we can take my aiming to command sailors as a received premise in my deliberations. Within the context of these deliberations, then, I will have a host of reasons for doing what is necessary to accomplish this aim. Finally, since playing this role in our deliberations suffices for something to really be a reason for action, I do in fact have these reasons. Second argument. Artifact concepts come with evaluative standards related to their constitutive features. A car is something that goes, so a broken-down car is a defective car. A cup holds liquid, so a leaky cup is a defective cup. (These kinds are goodness-fixing in Judith Jarvis Thomson’s (2008) sense.) Captain’s agency also comes with inherent evaluative standards related to its constitutive features. The more fully one possesses these constitutive features, the better a captain’s agent one is. Thus, successfully adjudicating complaints, answering missives from the admiralty, commanding sailors, and so forth make me better qua captain’s agent. In presupposing my captain’s agency, I am accepting that I am subject to these evaluative standards. This acceptance can then affect what reasons I have in one of at least two ways. (i) Accepting the evaluative standards of captain’s agency means that I have a reason to be a good instance of this kind rather than a defective one. Consequently, I have a reason to adhere to any requirements that are constitutive of captain’s agency. (ii) Questions about our reasons are simply deferred to the ideal of captain’s agency. That is, whether I have a reason to do something as a captain turns on whether a certain ideal captain’s agent version of myself would want it. Insofar ideal captain’s agency involves having distinctive desires, my reasons will reflect that ideal. Third argument. I have so far mentioned what we might call the external conditions on having captain’s agency: how I must be related to the world and other persons to occupy the place in the social structure that gives me the powers of captaincy. But there are also internal conditions. Imagine that my mind is inserted into the body of Captain Hardy aboard the HMS Victory. If this transition is seamless, I will immediately satisfy all the external conditions of captain’s agency. But I will obviously not really possess captain’s agency, since I will have no idea how to exercise it. Some of my deficit will consist in simple propositional knowledge. But some of it will consist in more complicated pieces of mental hardware. It’s one thing to know that there is a norm about who should discipline the petty officers; it’s another thing to have one’s behavior guided effortlessly and unconsciously by this norm. It’s one thing to recognize your third lieutenant as the man who just died; it’s another to see this as a reason to 422

Agency and normativity

immediately promote someone. To exercise my captain’s agency, I need special psychological equipment that allows me to interface with the social structure that I inhabit. Following Sally Haslanger (2016), we can call this equipment a psychological schema. In presupposing captain’s agency in my deliberations, I am also presupposing my having internalized the schema that allows me to exercise that agency. This matters because psychological schemata affect what reasons we have. They do this by structuring what I have elsewhere called the “background framework” that we employ in practical reasoning (Walden 2018a). Consider an analogy. What I should believe about a scientific question depends on the evidence I have, but it’s not just the evidence that delivers a conclusion. Evidence underdetermines theory choice. It also depends on the battery of principles and methods that I use to form conclusions on the basis of this evidence. By the same token, the reasons a person has will depend not just on their values and commitments, but also on how their practical reasoning is structured: on what kinds of inferences are licensed, what premises are taken as given, and so forth. To use one of the examples from before, if I have internalized the right sort of schema, I can infer directly from the death of the third lieutenant to my having a reason to promote someone immediately. Given my adoption of the captain’s schema, the death of the third lieutenant practically necessitates a promotion. On this picture, the constitutive features of agency affect our reasons not by dictating particular aims or ideals, but by affecting the background framework that structures our practical reasoning and thereby determining what reasons we have given particular aims and ideals.

2 Deliberative externalism The above arguments depend on two connections: between a certain normative question and captain’s agency, and between captain’s agency and its allegedly constitutive features. We might worry about slippage at both junctures. Must I presuppose captain’s agency when I wonder whether to order the men to turn to starboard? Consider the capacity impostor captain’s agency. It is like captain’s agency, but it has fewer constitutive responsibilities, powers, and aims since it only involves pretending to be captain. Couldn’t I presuppose impostor captain’s agency in my deliberations about whether I should order the men to starboard and in so doing avoid the responsibilities of actual captains? These worries are justified. It’s not true that a given normative question must presuppose one and only one kind of agency. Different individuals in different circumstances may presuppose somewhat different powers, depending on those circumstances. But this point is not fatal to the argument. Even if I am not presupposing captain’s agency in asking this question, I am presupposing something similar, like impostor captain’s agency. And this may also have interesting constitutive features that beget reasons via one of the pathways described above. What is actually being presupposed, and so what the commitments of the deliberator are will be an empirical question. But a skeptic could reply with what looks like a trump card. Look, they say, when I ask normative questions, I am not presupposing captain’s agency or even impostor captain’s agency, or indeed any other kind of agency that comes with substantive demands. I am presupposing a bespoke agency: a kind of agency that gives me all the powers I want and none of the responsibilities I don’t, a kind of agency that has the aims I openly accept and none of the ones I don’t. This escape plan won’t work. The reason why brings out an important implicit premise in the arguments from above. What kind of agency I presuppose is no more up to me than the meaning of the words I use. I cannot stipulate that by “arthritis” I mean a disease of the 423

Kenneth Walden

bones or that when I’m talking about “that shiny gold stuff” I do not mean to be talking about iron disulfide. Nor can I stipulate that I am presupposing this bespoke kind of agency that involves the power to order around sailors but absolutely no responsibilities. The first of these claims depends on the doctrine of semantic externalism: what I mean by my words is to a certain extent beholden to the world and other people (Putnam 1975 and Burge 1979). The latter expresses an analogous thesis that we could call deliberative externalism: the objects and presuppositions of our deliberations are also dependent on features of the world beyond the beliefs and intentions of the deliberator. Whenever an agent asks normative questions, there is a fact of the matter about the kind of agency they are presupposing, one that is fixed by their behavior, their beliefs and desires, the contours of situation they find themselves in, their particular capacities, and what others recognize them as doing. It is not up to them to say willy-nilly that they are actually presupposing something else. On this way of understanding the argument, quite a bit hangs on which kinds of agency there are, how finely they are individuated, and how much they have in common. If there are five different kinds of agency that someone in my situation could realistically presuppose when deliberating and they all involve responsibilities for adjudicating subordinates’ complaints, then we can reasonably conclude that they are presupposing those responsibilities. If there are 72 kinds and half of them don’t involve this kind of responsibility, then we have to look more closely at the case and decide which kind of agency is being presupposed. These issues may sound slightly fanciful, but once we take seriously the idea that there are different kinds of agency keyed to different physical and social systems, they become perfectly comprehensible—if difficult—empirical questions.

3 Generic agency Captain’s agency is a niche kind of agency, and only certain kinds of deliberation presuppose it. This makes the reasons we can derive from its presupposition conditional. Stop doing the things that presuppose captain’s agency and those reasons evaporate. But we can also imagine a completely generic agency. This generic agency would be the capacity that any creature calls on whenever they do anything whatsoever. Whereas the demands inherent in captain’s agency can be shrugged off by ceasing to engage in captain deliberations, those of generic agency would be inescapable in the sense that I must presuppose them when I enter any deliberation at all. So if one of the arguments canvassed above succeeds, then we may be able to derive claims with unconditional normative authority by identifying the constitutive features of this generic agency. Before discussing attempts to do this, I want to give a short brief against the temptation to brush off the normative claims associated with non-generic agency as unimportant. Suppose that there is such a thing as medieval French peasant’s agency—a kind of agency that underwrites the distinctive activities of French peasants. Such agency would be a presupposition of a great many of the deliberations that medieval French peasants undertake, and so the constitutive features of that agency will shape the reasons that medieval French peasants have. These effects are conditional since one can, in principle, give up asking the relevant deliberative questions. But it’s really hard to do this. There are powerful social and economic forces that keep peasants locked in their station. And even if one could leave the social structure that this kind of agency is associated with, the psychological conversion would not be instantaneous. So while medieval French peasant’s agency is not relevant to the heady project of discovering the foundations of normativity as such, it is relevant to the equally important

424

Agency and normativity

project of understanding the ways in which the conditions of agency shape the broader normative landscape. Back to generic agency. I will first summarize some arguments concerning generic agency roughly analogous to the First argument above. In that argument, we hypothesized that the aim of commanding sailors was constitutive of captain’s agency and argued that this fact grounded concomitant reasons. Is any aim constitutive of generic agency in this way? Is there anything such that insofar as we are deliberating at all, we are aiming at it? David Velleman (2009) says that self-understanding is such an aim. When we act, Velleman says, we are trying to make sense of what we are doing, and so whatever else we may be aiming at as agents, we are also aiming at making ourselves intelligible. Motivating this claim is the observation that such understanding is very often what distinguishes action par excellence from mere movement. Taking a sip of tea and chewing on sandpaper are both behaviors that could be caused by my desires. But only one of them makes sense in light of my withering thirst. And this is why sipping tea is an action while chewing on sandpaper must be something considerably less. If self-understanding is a constitutive aim of generic agency, then every deliberation presupposes this aim. So we can take the aim as given in those deliberations. This means that we have a reason to do those things that will help bring about the aim and avoid those that won’t. Velleman goes on to argue that this aim is well-served by forms of interaction that have certain pro-social features. The aim encourages the cultivation of coherent and shared values, inspires the adoption of a stock of scripted social scenarios, tends to inculcate a regard for one’s partners in these scenarios, and fosters the development of a mental faculty very much like conscience. Velleman doesn’t think that the constitutive aim of agency mandates a particular moral system, but he does think that it nudges us toward our moral way of life, and the reasons we have in virtue of this aim rationalize that way of life. Paul Katsafanas (2013) proposes a theory with a similar structure but a different constitutive aim. He begins by saying that actions are generated by our drives. He then observes a familiar dynamic. A drive produces a desire—for a professional recognition, say—we are motivated by that desire, we satisfy it, and the desire is snuffed out. But after a time, the drive produces another desire of the same genus. What explains this phenomenon, he says, is an underlying drive that aims at encountering and overcoming resistance. Pairing this fact with the central role of drives in our agency yields a constitutive aim. Following Nietzsche, Katsafanas calls this “will to power”. On his view, then, it is constitutive of agency that we aim at organizing our activities so that they provide meaningful and significant challenges that we must exert ourselves in overcoming. Reasons to pursue such activities follow in train. There are also examples that more closely resemble the Second argument from above. One is from Christine Korsgaard. Insofar as I am to act, there must be an I who acts: a unified, independent entity that is the agent. And insofar as I am to act, this entity must be efficacious: it must be capable of effecting change. These facts constitute an “ideal” of agency: The ideal of agency is the ideal of inserting yourself into the causal order, in such a way as to make a genuine difference in the world. Autonomy, in Kant’s sense of not being determined by an alien cause, and efficacy, in the sense of making a difference in the world that is genuinely your own, are just the two faces of that ideal, one looking behind, and the other looking forward. (2009: 89)

425

Kenneth Walden

This ideal sets a standard of evaluation for our actions. Actions are exercises of agency, and the ideal of agency involves autonomy and efficacy, so actions that are not efficacious or autonomous are defective as actions in the same way that a leaky cup is defective as a cup. Korsgaard goes on to associate these faces of the ideal of agency with two principles, both from Kant. Efficacious action, she says, depends on following the hypothetical imperative: take the means to your ends. Autonomous action depends on following the universal law formulation of the categorical imperative: act only on maxims that can be willed to be universal laws. This makes these principles constitutive requirements of agency. Thus, an action that does not conform to these principles is defective as an action. Michael Smith (2011 and 2015) defends a view on which my reasons are fixed by the desires of an idealized version of myself. The idealization involves my agency. To figure out what reasons I have we should consult the desires of the agentially best version of myself. More specifically, we should consult the desires of my counterpart in the nearest possible world where I have and exercise maximal capacities of agency. Obviously, ideal agents, so understood, will have varied desires. But perhaps there are some desires we can be certain that all such agents have just in virtue of their being ideal. This is Smith’s strategy. He begins with a spare, Humean conception that identifies agency with capacities to come to know about the world and realize one’s desires. He notes that such an agent is vulnerable to self-frustration, if, for example, they have a desire to believe that p whether or not it’s true. But an ideal agent wouldn’t be vulnerable in this way, so we should doubt that they would have such a desire. More generally, Smith says that we have good cause to attribute “coherence-inducing desires” to our ideal agent. These desires, he argues, will include desires to help and not interfere with both oneself and other persons. From this and his account of the relationship between our reasons and the desires of our ideal counterparts, it follows that every agent has reasons to help and not interfere with their own projects and those of others.

4 Two objections The arguments just summarized claim to establish the normative authority of certain principles by showing that they are constitutive of agency. And because they are constitutive of generic agency, it is claimed, this authority is unconditional. Call these constitutivist arguments. In closing I want to briefly deal with two objections to such arguments. The first denies that agency is normatively special in the way constitutivists claim. If we want to show that we have certain reasons, David Enoch says, it is not sufficient to show that some aims or motives or capacities are constitutive of agency. Rather, it is also necessary to show that the “game” of agency is one we have reason to play, that we have reasons to be agents …. And this, of course, is a paradigmatically normative judgment. (2006: 181) Enoch dramatizes this point by having us imagine someone who wonders, upon hearing constitutivist arguments, “perhaps you’re right about what is constitutive of agency, but why should I be an agent in the first place?” This person may instead be perfectly content to be a shmagent—a non-agent who is very similar to agents but lacks those features of agency that constitutivists seize on in their arguments. So our question is, “why be an agent instead of a shmagent?” If we answered this question by citing some wholly exogenous reason—that 426

Agency and normativity

agency is more fun—constitutivism would lose much of its foundational appeal. So we need a more subtle approach.2 My reply begins with a question. I think this objection is ambiguous between a few different concerns. We can classify them by asking the following question: Do the normative questions asked in the mode of deliberation presuppose the agency of the subject deliberating? There are two ways Enoch’s skeptic could respond: (1) yes, but this has no normative consequences: it doesn’t entail anything about what reasons we have or what we ought to do, and (2) not necessarily: it could presuppose something else. Option (1) ramifies into two further possibilities depending on whether our skeptic accepts a platitude about reasoning—whatever reasons are, they must bear some important connection to the systems that generate and regulate our actions. (1a) rejects the platitude and says that reasons have no special connection to the systems that regulate action. This is a strange view that requires substantial elaboration to be even minimally viable. (1b) accepts the platitude but says that reasons do not regulate action through deliberation. Instead they regulate it through some other, probably less active and self-conscious process—something more like a well-developed sensitivity. One might hold this view because they think deliberation is epiphenomenal (Wegner 2002) or as part of more general doubts about the intellectualism of deliberation-centric views of rationality (Wiggins 1975). I cannot adequately reckon with this possibility here. I mention it mainly because it offers an important but often overlooked route of resistance to the arguments sketched above (compare Silverstein 2016). Option (2) denies that deliberation presupposes agency. If this is the skeptic’s position, we might wonder, then what does it presuppose? Deliberation is not an activity of shouting questions into the void. It’s about something. So if it doesn’t presuppose agency, it must presuppose something similar, like shmagency. Is this at all plausible? I don’t think so, for reasons relating to deliberative externalism. If a person talks about “reasons” and what he “ought” to do, if he carries out his deliberations in the normal way, and if he is in regular contact with other persons, it is very unlikely that his deliberations could presuppose shmagency (if “shmagency” denotes something different from our ordinary concept of agency). That he doesn’t care to be an agent and is happy being a shmagent doesn’t matter to the question of whether he has failed according to a standard that he has (implicitly) committed himself to. It doesn’t matter for the same reason that it doesn’t matter whether he would prefer to avoid the subject of joints whenever he mentions arthritis. This is one sense in which agency is inescapable. If a person is integrated into our actual practices of deliberation, reasoning, action, evaluation, interpretation, and accountability, then agency is the only thing he can realistically claim to be presupposing because those practices presuppose agency, not its kooky cousin shmagency. If he falls short of the demands of agency, then he also falls short of the internal standards of his own deliberation. This does not rule out a more far-fetched possibility: that there are aliens whose practices revolve around shmagency, shmaction, shmreasons, and so forth. These aliens may very well presuppose shmagency when they shmeliberate. This way of escaping the constitutivist argument is bolder. It doesn’t just stipulate that one is interested in a single state designated by a neologism. It involves an entirely different conceptual scheme that cannot be translated into our own. The second objection claims that the constitutive nature of generic agency is, pace the constitutivist arguments, too thin a reed on which to build any remotely interesting normative conclusions. There are so many ways that agents can do things and such a multiplicity of different kinds of agency that we should doubt that the generic power of agency that unites all these things has any substantive nature at all (Setiya 2007; Tiffany 2011; Millgram 2016). 427

Kenneth Walden

This skepticism naturally engenders charges of equivocation. Constitutivists begin by asserting the inescapability of a truly generic kind of agency, but then adduce constitutive features of a particular kind of agency. Much of the debate over this objection must take place in the trenches: constitutivists need to show that the notion of agency they are relying on is properly generic rather than just one species of a larger genus. There is, however, one slightly different reply. Suppose that agency has no substantive constitutive nature for vaguely existentialist reasons: the nature of agency is not fixed in advance of our actions but constituted by those actions. If this is true, then agency cannot furnish agents with any special aims or standards. But there may still be a way to ground the authority of certain normative principles in the nature of agency. For even if agency is whatever we make of it, this means that we have to make something of it. We have to achieve a kind of convergence. If everyone follows wholly disparate principles of action, then our behavior will be too heterogeneous a mass to constitute a kind. In this instance there wouldn’t be such a thing as agency. Insofar as we presuppose our agency in deliberation, we have a reason to avoid this kind of heterogeneity. And this suggests a different kind of normative principle that could be derived from agency. Not a principle to adhere to standards reflecting the fixed nature of agency, but a principle requiring our cooperation in creating and sustaining agency—whatever that ends up being (Walden 2012; Walden 2018b).

Related topics Minimal agency; Diachronic agency; Agency and practical reasoning; Agency and morality; Agency and reasons; Pathologies of agency; Bounded agency; The aim of agency.

Notes 1 For more on pluralism about action, see Millgram (2010, 2016), and Lavin (2017). There is a large literature on the connections between social structure and agency. Giddens (1979) is a classic. 2 There is a rich literature on this objection that I can hardly do justice to in this summary. See, amongst others, Ferrero (2009), Enoch (2011), Silverstein (2012, 2015), Rosati (2016), and Ferrero (2018).

Further reading Katsafanas, Paul. 2018. “Constitutivism about Practical Reasons,” in Daniel Star (ed.), Oxford Handbook of Reasons and Normativity (New York: Oxford University Press). Millgram, Elijah. 2016. “Practical Reason and the Structure of Actions,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2016 ed) . Paakkunainen, Hille. 2017. “Normativity and Agency,” in Tristram McPherson and David Plunkett (eds), The Routledge Handbook of Metaethics (London: Routledge). Each of these offers an excellent overview of similar terrain from a slightly different point of view.

References Burge, Tyler. 1979. “Individualism and the Mental,” Midwest Studies in Philosophy 4: 73–121. Enoch, David. 2006. “Agency, Shmagency,” Philosophical Review 115(2): 169–198. Enoch, David. 2011. “Shmagency Revisited,” in Michael Brady (ed.), New Waves in Metaethics (London: Palgrave-Macmillan). Ferrero, Luca. 2009. “Constitutivism and the Inescapability of Agency,” in Russ Shafer-Landau (ed.), Oxford Studies in Metaethics 4: 303–333.

428

Agency and normativity Ferrero, Luca. 2018. “Inescapability Revisited,” Manuscrito 41: 4. Giddens, Anthony. 1979. “Structure, Agency,” in Central Problems in Social Theory (Berkeley, CA: University of California Press). Haslanger, Sally. 2016. “What Is a (Social) Structural Explanation?,” Philosophical Studies 173(1): 113–130. Katsafanas, Paul. 2013. Agency and the Foundations of Ethics (New York: Oxford University Press). Korsgaard, Christine M. 2009. Self-Constitution (New York: Oxford University Press). Lavin, Douglas. 2017. “Forms of Rational Agency,” Royal Institute of Philosophy Supplement 80: 171–193. Millgram, Elijah. 2010. “Pluralism about Action,” in Timothy O’Connor and Constantine Sandis (eds), A Companion to the Philosophy of Action (Malden, MA: Wiley-Blackwell). Millgram, Eljiah. 2015. The Great Endarkenment (New York: Oxford University Press). Putnam, Hilary. 1975. “The Meaning of Meaning,” in his Philosophical Papers, Vol. II: Mind, Language, and Reality (New York: Cambridge University Press). Rosati, Connie. 2016. “Agents and ‘Shmagents’,” Oxford Studies in Metaethics 11: 182–113. Setiya, Kieran. 2007. Reasons without Rationalism (Princeton: Princeton University Press). Silverstein, Matthew. 2012. “Inescapability and Normativity,” Journal of Ethics and Social Philosophy 6(3): 1–27. Silverstein, Matthew. 2015. “The Shmagency Question,” Philosophical Studies 172(5): 1127–1142. Silverstein, Matthew. 2016. “Teleology and Normativity,” Oxford Studies in Metaethics 11: 214–240. Smith, Michael. 2011. “A Constitutivist Theory of Reasons: Its Promise and Parts,” Law, Ethics, and Philosophy 1: 9–30. Smith, Michael. 2015. “The Magic of Constitutivism,” American Philosophical Quarterly 52(2): 187–200. Thomson, Judith Jarvis. 2008. Normativity (La Salle, IL: Open Court). Tiffany, Evan. 2011. “Why Be an Agent?” Australasian Journal of Philosophy 90(2): 233–233. Velleman, J. David. 2009. How We Get Along (New York: Cambridge University Press). Walden, Kenneth. 2012. “Laws of Nature, Laws of Freedom, and the Social Construction of Normativity,” Oxford Studies in Metaethics 7: 37–79. Walden, Kenneth. 2018a. “Practical Reason Not as Such,” Journal of Ethics and Social Philosophy 13(2): 1–32. Walden, Kenneth. 2018b. “Nature, Agency, and the Nature of Agency,” Philosophical Inquiries 6(2): 51–72. Wegner, Daniel. 2002. The Illusion of Conscious Will (Cambridge: MIT Press). Wiggins, David. 1975. “Deliberation and Practical Reason,” in his Needs, Values, Truth (New York: Oxford University Press).

429

39 THE AIM OF AGENCY Kathryn Lindeman

1 Introduction When an agent acts intentionally, she does so with some particular aim or end in mind, say, preparing dinner. Some claim that in addition to having particular aims, actions and agents also have some more general aim as a constitutive feature. Agent’s particular aims are often taken to explain the reasons and normative standards governing the means they take to achieving them. If agency itself has an aim, perhaps this aim could similarly provide the grounds for an explanation of more general normative results. I’ll call philosophers who pursue this explanatory project Aim Theorists. The Aim Theorist’s project faces two immediate challenges. First, Aim Theorists must explain how agents as such have a constitutive aim. The explanation of how individual actions have aims relies on the intentional goals or purposes agents have in performing those actions. But Aim Theorists need another account for how agency itself could have a constitutive aim, given that actions and agents admit of too much diversity to imagine that all agents share some particular intentional aim or goal in all of their actions. Second, Aim Theorists must account for the normative authority of the reasons and standards they propose to explain by appeal to an aim of agency. These two demands generate the tension in the Aim Theorist’s explanatory project: the standards and reasons arising from agents’ particular aims derive their authority from the internal connection to the agent’s own intentions and goals, rather than an aim of agency itself. Aim Theorists all agree that appealing to an aim of agency or action can do important explanatory work, but they differ in many respects. They differ in their explanatory ambitions, that is in the extent of the normative explanation they appeal to aims to provide. They disagree about whether to understand the aim of agency as an end that all exercises of agency share or as a unifying drive or motive that guides and structures any of agency’s exercises. They disagree about whether the aim can be given a substantial description or whether it must be understood as a formal feature of agency. Finally, there is substantial disagreement about what the aim of agency is.

430

DOI: 10.4324/9780429202131-48

The aim of agency

2 The aim of agency and the aim of belief Explanations of normativity in other domains also appeal to constitutive aims, most notably in epistemology (see Côté-Bouchard 2016; Flowerree 2018 for recent critical discussion). Despite the commonality, the two literatures have two important differences. First, there is much more disagreement about the nature of the aim of action than about the aim of belief. Despite some notable resistance (see, e.g., McHugh 2011; Whiting 2012, 2014; Littlejohn 2013; Nolfi 2018), the consensus is that belief constitutively aims at truth. There is no similar consensus view about the aim of agency, either on the question of whether there is such an aim or, if there is, what it might be. Second, the aim of belief is taken to be accessible to believers in a way that the aim of action or agency is not accessible to agents. Shah (2003) has compellingly argued that because deliberation about what to believe collapses into deliberation about what is true, the aim of belief is ‘transparent’ in the sense that it is either impossible or incoherent to intentionally attempt to believe against it. Whatever the aim of agency might be, it does not seem impossible or incoherent to intentionally attempt to act against it, so the aim of agency seems to be opaque to agents (Silverstein 2017). It is worth noting that there is a prominent account of the aim of belief that is not a version of Aim Theory. Wedgwood (2002: 267) has noted that because beliefs are not “little archers armed with little bows and arrows,” they cannot have intentional aims. This leads Wedgwood to endorse normativism, according to which the aim of truth for belief is a metaphor to indicate that belief is the mental state with truth as its correctness condition. Normativism explains aim-talk as expressing commitments to normative facts, typically as conceptual truths, rather than literal reports that could be used to explain normative facts. Normativism thus requires an independent explanation of normative facts.

3 Explanatory ambitions of Aim Theories All Aim Theorists take the aim of action or agency to provide resources to explain at least some normative standards and their authority. The most ambitious strategy promises to explain the entire content of practical rationality and an account of its normative authority (Korsgaard 2009). Other, more restrained accounts only seek to explain the normative authority and content of a restricted set of standards, requiring additional resources to provide further constraints (e.g., Katsafanas 2013). Still more minimal accounts propose only to explain the authority of rational requirements like the instrumental principle (the requirement to take the recognized means to one’s ends). Among Aim Theorists who seek to explain substantial normative commitments, the dominant explanatory strategy relies on subjectivist commitments about reasons (see, e.g., Schroeder 2007; Markovits 2014). Katsafanas (2013: 39–41) is more explicit than many Aim Theorists about his reliance on the position that aims are themselves reason-providing. Participating in an activity with a constitutive aim generates reasons to take the means to achieve the aim and thus to satisfy the success conditions. On Katsafanas’s Nietzschean account, the aim of agency is exercising the will to power by encountering and overcoming resistances to the relevant activity type. What characterizes an activity as an action is being structured by the second-order aim of encountering and overcoming resistance to it (Katsafanas 2013: 178). So, whenever we engage in any activity or in practical deliberation

431

Kathryn Lindeman

about how to engage in that activity, we have reason to engage in challenging versions of it (Katsafanas 2013: 184ff ). For instance, I cannot engage in an activity with a first-order aim, like cooking dinner, without encountering and overcoming resistances to cooking: for example, I have to clear my schedule to cook, I have to pay attention and not injure myself while chopping. Encountering and overcoming these resistances isn’t the end of my action, having cooked a meal is, but I cannot cook the meal without encountering and overcoming meal-making resistances. Encountering and overcoming resistance to cooking dinner, then, is a success condition, and this, in turn, explains correctness conditions on how I go about cooking as those that increase the odds of encountering and overcoming resistance to my cooking. One resistance to my cooking dinner is the fact that it must be served at a certain time. If I chop too slowly, my family will eventually give up and order pizza. If I chop too quickly, I might slice my hand, requiring a trip to the ER. Chopping quickly and carefully is thereby the correct way to go about cooking dinner. Granting certain assumptions about when reasons transfer between ends and means, the constitutive aims also explain my reasons to satisfy the correctness conditions on cooking. Significantly, the reasons that the aim of agency produces are universal because, unlike the aim of cooking dinner, the aim of agency is present in each action. It is what is sometimes called inescapable (see Ferrero 2009; Katsafanas 2013: 61). This strategy thus promises to explain the content and normativity of at least a subset of the demands of practical reason. At the other end of the explanatory ambition scale are those like Dreier (1997, 2001), who endorse Instrumentalist Aim Theories. Dreier takes the teleological directedness of action to at least explain the normative authority of the instrumental rule-telling agents to take the acknowledged means to their desired ends, what Dreier calls M/E: M/E If you desire to ψ, and believe that by ϕ-ing you will ψ, then you ought to ϕ. Dreier (1997: 96) argues that the normativity of M/E can be established, not by demonstrating the reasons all agents have to accept it but by seeing the role that it plays in making it possible to give reasons to an agent. Being disposed to recognize and follow M/E is a precondition on anything being a reason for you, or put another way, a precondition on being an agent. M/E thus must be recognized as having a ground-level normative status, explained by the fact that any agent must satisfy it in order to exercise their agency at all (see also Korsgaard 1997 and Smith 2009: 9). Such Instrumentalist Aim Theories are barely Aim Theories. Insofar as they are, it is because they take the disposition to satisfy the instrumental principle to be partially constitutive of exercises of agency and as itself an aim that grounds a sort of transcendental argument for its own justification.

4 Teleological versus efficient Aim Theories All Aim Theorists must account for the sense in which agency has an aim. As Rosati (2016: 5) cautions “talk of an aim of agency may give the misleading impression that it is an agent’s very own aim—what an agent consciously aims at—rather than something sub-agential that makes for agential functioning, thereby making agency possible.” Though some Aim Theorists write as if agents are conscious of or endorse the aim of agency, none think that agents have their constitutive aim because of their intentional conscious aims. Aim Theorists account for agents’ having aims in two ways. Some take agency to have an aim in a functional or teleological sense; the aim of agency is an end that all action must be understood as seeking to achieve because of the function or purpose of action. Others take 432

The aim of agency

agency to have an aim in a governing or efficient sense; the aim of agency is a motive or drive whose guidance makes behavior into agentive activity. Smith’s (2012) Humeanish account exemplifies a teleological Aim Theory that promises to derive reasons to satisfy moral requirements. According to Smith, the aim of agency is intrinsic desire-realization, which Smith (2012: 329–330) claims explains universal categorical reasons to help and not hinder the exercise of all rational capacities. Roughly, he argues that not only are representational and motivational capacities necessary for pursuing the aim of desire realization, but the desires to help and not hinder their exercise are required to ensure they are not exercised at cross purposes. So, these desires are needed to guarantee the ideal exercise of agentive capacities. On Smith (1994)’s earlier account, agents’ reasons are a function of the desires of their idealized counterparts. Given that every idealized counterpart would have the desires required to ensure ideal exercise of its capacities, every agent has a reason to help and not hinder the rational capacities of all rational agents (see Bukoski 2016 and Lindeman 2019 for discussion). Efficient Aim Theorists, instead, focus on sub-agential commitments, dispositions, or principles that, they argue, must guide agency for it to have any ends at all. On such sub-agential accounts, the aim of agency is not an end that all actions aim at achieving, but an aim that must be understood as guiding or shaping agential activity in its pursuits. A good example is Velleman’s account of agency as aiming at self-understanding. Velleman begins not with the metaphysics of agency, but with its phenomenology: a central difference between action and mere behavior is that, in acting, agents have practical, non-observational self-knowledge. To account for this self-knowledge, Velleman argues, we must understand action as governed by an intellectual drive with two central roles. First, this drive ensures that the agent only acts for reasons that make sense to her. Second, it disposes her against engaging in behavior that she cannot reconcile to her self-conception. In doing so it influences “which desired objects we choose to pursue, how we harmonize them with one another, organize our effects towards them” and to do so in ways that allow “the agent to understand what she herself is up to in so acting” (Velleman 2009: 28, 133). Moreover, it leads us to choose actions that make more rather than less sense, and in the limit case, it prevents us from acting if we could not make sense of someone like me acting in that way. This intellectual drive for self-knowledge must be operative in any agentive capacity, and its guidance accounts for how agency aims at self-knowledge in all of its exercises. It is not what he calls an end-in-view, or something that we act in order to bring about. Nonetheless, it shapes what we have reason to do and, Velleman (2015: 50ff ) argues, favors social coordination because of distinctive ways in which self-understanding requires sociality.

5 Formal versus substantial Aim Theories Understood in the most formal sense, the aim of agency can be described as “figuring out what to do,” that is, the correct thing to do. However, “a mode of reasoning whose goal was specified solely as ‘figuring out what to do’ would be like a search whose object was specified solely as ‘figuring out where to look,’ or a question whose object was specified solely as ‘figuring out how to reply’” (Velleman 2000: 177). Many Aim Theories identify more substantive aims, where this can be understood as a specification of what it is to achieve the formal aim (e.g., Gauthier 1997). The most formal Aim Theories understand agency as the capacity to bring about action through deliberation and choice, and identify agency’s aim as good or choiceworthy action. Tenenbaum (2007) defends such a version of the Guise of the Good based in action’s 433

Kathryn Lindeman

connection to desire. In order for desire to play its role as “an attitude whose object would have positive weight in deliberation” (Tenenbaum 2007: 38), it must present its object as good in some way. In acting, for example, realizing a desire, Tenenbaum argues, an agent makes an unconditional judgment of the desired end as good from their evaluative perspective. Agency thus has the aim of the good in the same way, Tenenbaum argues, as theoretical deliberation has the aim of the truth. In acting, we aim to bring about good actions, in believing we aim to bring about true beliefs. The deliberative faculty is thus unified and its exercises are distinguished by their different formal aims. Such formal accounts are often taken to have difficulties establishing normative results; however, Tenenbaum (2007: 200ff ) argues that his account can establish formal constraints like consistency and coherence to determine permissible conceptions of the good, as well as what he calls “deontological goods,” which must appear good from any permissible conception of the good. Smith’s account, discussed above, is another fairly formal account that identifies choice-worthiness with the realization of intrinsic desires. Smith is able to explain normative constraints beyond those needed to merely satisfy an agent’s contingent occurrent desires. In part because satisfying desires requires not merely acting to realize them, but also deliberating about how to go about realizing them. For instance, I often do not know how to best cook dinner and must deliberate about when and how to do so. So satisfying even everyday desires often involves forming derivative desires, like the desire to chop carefully and quickly. In the language of desire-realization, I exercise my agency not only when I realize my desires, but also when I deliberate about how to realize them and form derivative desires. So, regardless of the desires I begin with, there will be consistency and coherence standards on how I form new desires. In addition, as discussed above, Smith also claims to be able to establish some intrinsic desires required by the end of intrinsic desire realization. Perhaps the most prominent account on which agency has a substantive constitutive aim is due to Korsgaard (2009). She argues that agency aims at self-constitution. Her argument begins by distinguishing human action from the purposive activity of non-rational animals. In acting, an agent determines herself to be the cause of some end by choosing which inclination to act on, thereby endorsing and identifying with it. On Korsgaard’s account, to act intentionally, you must be guided by principles that allow you to make yourself the cause of your action: in acting, an agent decides who to be by giving laws to herself. The aim of self-constitution allows Korsgaard to claim there are better and worse laws to give yourself. Some laws will make you ineffective. Some will make you less unified as an agent. Korsgaard, for reasons similar to Dreier, argues that to ensure you are effective, you must ensure your principle is in accordance with the hypothetical imperative. But Korsgaard (2009: 72) argues that you must also be governed by the categorical imperative, because it is only by following it that your action is not defective in its aim of self-constitution: unlike any other principle, the categorical imperative “is the principle that truly unifies the soul, and unifies it in a way that makes it capable of effective action” (Korsgaard 2009: 175).

6 Can aims explain the content of normative standards? The most immediate objection to Aim Theorists is that they fail to explain how the aim they identify is, in fact, constitutive of agency. As the diversity of views canvased above demonstrates, Aim Theorists are not able to convince even (especially) each other on this matter. But even those who accept that there might be some sense in which agency has some aim are often unwilling to accept that this aim could explain the content of normative standards. One problem, as Morton (2010: 566) has put it, is that of “spelling out an aim that is 434

The aim of agency

weak enough to truly be uncontested, but strong enough so that agents could fail to satisfy it and, consequently, […] establish a substantial normative standard.” So, for example, Setiya (2007: 112) is willing to grant that action “aims at self-understanding because one’s intention in acting is a desire-like belief about the motivation of one’s action, which its execution would make true.” He notes, however, that this aim isn’t better achieved by any particular intention or other. So, at most, the instrumentalist Aim Theorist is vindicated. Another worry is related to the opacity of agency’s aim. In any particular action, an agent’s particular aim may be directly opposed to the constitutive aim of agency. Some even worry that the aim of agency might be self-effacing, precluding agents from taking it as their aim. Silverstein (2015: 1135) notes that such a worry might threaten Velleman’s account because “it does not always promote intelligibility to pursue intelligibility; sometimes what makes the most sense is to stop making sense.” This might seem to be no more a problem for Aim Theorists than the paradox of hedonism is a problem for hedonism. But Bukoski (2017: 2673) notes that such self-effacement might be incompatible with other claims Aim Theorists make about the aim of agency, such as Velleman’s (2009: 138) claim that the aim is self-validating. So, the potential for self-effacement will at least constrain Aim Theorists’ possible responses to other objections. Other critics worry that Aim Theorists end up committing themselves to highly unintuitive normative standards. Bukoski (2017), for instance, notes that because on Velleman’s account correct action is action that makes an agent most intelligible to herself, if an agent is most intelligible as lazy, his view would say that she ought to procrastinate. Velleman (2009: 31–33) accepts that the demand to make folk-psychological sense will sometimes require manifesting vices, though he claims that the demand to be more intelligible will push us toward more moral behavior (see, however, Walden 2018). Similar worries obviously apply to Katsafanas’s Nietzschean account, given the limited constraints provided by the demand to engage in challenging versions of actions. Finally, Bukoski also objects (2016: 137–140) that Smith’s account entails implausible normative standards, in contradiction to Smith’s claims (2013: 26) that the results of his account are coextensive with those of common-sense morality. There are additional worries that Aim Theorists offer implausible accounts of the reasons agents have to satisfy moral standards. Walden (2018: 78) argues that on Velleman’s account, “other persons will merit moral treatment only insofar as they participate in the scenarios that further our self-understanding.” This seems to generate the wrong sort of reasons: an agent’s reasons to respect others should be based in their features rather than hers. According to Walden (2018, developing 2012), all accounts on which the aim of agency is not socially oriented will face such problems. Schafer (2018) also raises concerns about the inward-directedness of Aim Theorists like Korsgaard and Velleman, though not because their reasons don’t involve other people but because they are not outwardly directed at the world in general.

7 How could Aim Theorists explain normativity? Many prominent objections are willing to grant that Aim Theories are able to explain the content of purported normative standards but not their authority. One worry, most forcefully expressed in the Shmagency objection (Enoch 2006, 2011, 2019), is that the internal connection between aims and standards can only explain reasons to satisfy those standards if agents also have reasons to have the relevant aim. Another objection, what I’ll follow Katsafanas (2013: 63) in calling the Why Bother Objection, is that even if the connection does explain reasons we have to satisfy some of the standards, it cannot explain the reasons 435

Kathryn Lindeman

we have to satisfy all of them all of the time. Finally, what I’ll call the Presupposition Objection, states that even if the explanation does work, it does not provide the explanation of normativity it promises, because it relies on a normative premise in the explanation. The Shmagency Objection holds that even if agency is granted to have a constitutive aim, such an aim cannot itself be the source of the normative force governing agents. Enoch appeals to a common objection to subjectivist accounts of reasons, that while it appears that having an aim gives me a reason to take the means, it actually does so only if I also have a reason to have the aim. Enoch claims that constitutive aims of activities only have normative authority for agents if the activities themselves are normatively significant, because we must have a reason to endorse the activity’s aim in order for the standards deriving from that aim to have normative authority. Aims that one is alienated from do not generate reasons to satisfy them, much less to take any particular means to do so. In fact, according to Kolodny (2007), it isn’t the having the aim that is the reason to take the means, it is the reason to have the aim itself, and so actually having the aim is unnecessary for the reasons to take the relevant means. If this is true, Aim Theorists must explain what reasons agents have to endorse their constitutive aim, or put another way, to be an agent. Many Aim Theorists have argued that such a demand is unwarranted (see Rosati 2016), but it is often taken to be a devastating blow to the Aim Theorist’s claim to be able to answer the skeptic about morality (see Korsgaard 2009 for responses appealing to the proposed inescapability of agency, see Ferrero 2009, 2018 and Katsafanas 2013 for related argument posing the issue as a dilemma showing that any aim of agency capable of establishing normative conclusions is unable to generate content, see Tiffany 2012). There are two forms of the Why Bother Objection. The first begins with the observation that most of the purported aims of agency are scalar, that is, they can be achieved to a greater or lesser extent. One can, for instance, be more or less unified as an agent on Korsgaard’s account or more or less intelligible on Velleman’s. But it is not clear on these accounts why such a scalar aim provides agents with reasons to maximally satisfy the aim. As Tenenbaum (2019: 168) puts it, “given that we can realize the constitutive aim of agency to various degrees, we might be able to pursue the constitutive aim of agency without pursuing it to its highest degree or in its full perfection.” This is a particular threat for accounts like Korsgaard’s which appeal to an existential threat of failing to sufficiently satisfy the aim of agency. Korsgaard argues that we can explain the reasons an agent has to satisfy her internal standards because failing to do so amounts to risk of the literal disintegration of her agency. But if this is what explains the normative status of satisfying the aim of agency, it seems to only explain reasons to sufficiently rather than ideally satisfy agency’s internal standards. There are general worries about this appeal to existential threat (see Lindeman 2017), but even if it succeeds, it does not seem to explain the normative status of all of agency’s normative standards. A second version of the Why Bother Objection appeals to the fact that, for any aim, it seems possible to achieve that aim badly or luckily. If I manage to cook dinner, I might do so only because a neighbor surprised me with a rotisserie chicken right as I managed to put the oven fire out. I succeed in my aim in cooking dinner, but not because I satisfied the correctness conditions for cooking dinner (I set the oven on fire). Typical aims all seem to be like this: they can be achieved despite a poor showing on the part of the agent, but the Aim Theorist seems committed to denying this is true of the aim of agency. If one could get away with achieving the aim of agency, even maximally, while violating action’s normative standards, this would undermine their explanatory project. It is, however, very difficult to see why it shouldn’t be possible, for instance, to luckily manage to do what would makes oneself intelligible. Katsafanas (2013: 63–67) develops one line of response, claiming that though 436

The aim of agency

it is possible to cook dinner (thus achieving your aim) without satisfying all its correctness conditions, it is not possible to achieve the aim of agency in this way because in aiming as agents, we aim at perfection. Even if this were plausible, Katsafanas’s explanation undermines his explanatory project, which is supposed to show that what all agents must aim at can be used to explain, rather than include, the content of practical standards (see Lindeman ms). The Presupposition Objection, most clearly developed by Silverstein (2016), objects that the Aim Theorists’ explanatory project smuggles in normative presuppositions, undermining their claims to explain normativity by appeal to exclusively non-normative premises. He writes (2016:233) that because of the overcoming resistance to it, Aim Theorists “cannot rely on some other norm for justification—even a norm as seemingly uncontroversial as the principle that we have reason to achieve our aims.” There have been attempts recently to provide non-normative resources to explain how aims provide authoritative reasons. Lindeman (2019) argues that Smith’s recent account avoids these issues. Silverstein (2016, 2017) and others sympathetic to Aim Theories have noted the need to identify an account of reasons not subjectively but as tied to the good functioning of agency (Paakkunainen 2019) or rationality (Schafer 2017, 2018).

Related topics Agency and normativity; Agency and morality; Agency, functions, and teleology.

Further reading Brunero, J. 2017. “Recent Work on Internal and External Reasons.” American Philosophical Quarterly 54(2): 99–118. This survey piece provides an overview of three arguments for Internalism about Reasons, including one based in appeals to an aim of agency. Ferrero, L. 2019. “The Simple Constitutivist Move.” Philosophical Explorations 22(2): 146–162. This article provides an explanation of what is called constitutivism and argues that what I’ve here called Aim Theories are an instance of constitutivist explanations. Pauer-Studer, H. 2014. “A Constitutive Account of Group Agency.” Erkenntnis 79(9): 1623–1639. This article develops an instance of an Aim Theory to explain the nature and normative standards of Group Agency. Wald, B. and Tenenbaum, S. 2018. “Reasons and Action Explanation” in The Oxford Handbook of Reasons and Normativity, Daniel Star (ed.). Oxford: Oxford University Press: 214–232. This entry in the Handbook of Reasons and Normativity appeals to an account of agency’s aim to develop an answer to the long-standing problem causal accounts of action have dealing with deviant causal chains. Wiland, E. 2012. “Constitutivism.” in Reasons. London: Continuum Press. This chapter in a graduate-level text on reasons provides an accessible explanation of the motivation for believing that action or agency has an aim, its role in explaining reasons, and some general objections to the explanatory project.

References Bukoski, M. 2016. “A Critique of Smith’s Constitutivism.” Ethics 127: 116–146. ———. 2017. “Self-validation and Internalism in Velleman’s Constitutivism.” Philosophical Studies 174: 2667–2686. Côté-Bouchard, C. 2016. “Can The Aim of Belief Ground Epistemic Normativity?” Philosophical Studies 173 (12): 3181–3198. Dreier, J. 1997. “Humean Doubts about the Practical Justification of Morality,” in Ethics and Practical reason, ed. Cullity and Gaut. Oxford: Clarendon Press, 1997. Reprinted with some modifications in Millgram, ed. 2001. Varieties of Practical Reasoning. Cambridge, MA: MIT Press.

437

Kathryn Lindeman Dreier, J. 2001. “Humean Doubts about Categorical Imperatives,” in Varieties of Practical Reasoning, ed. E. Millgram, 27–48. Cambridge, MA: MIT Press. Enoch, D. 2006. “Agency, Shmagency: Why Normativity Won’t Come From What Is Constitutive of Action.” The Philosophical Review 115: 169–198. ———. 2011. “Shmagency Revisited.” in New Waves in Metaethics, ed. M. Brady, 208–233. London: Palgrave Macmillan. ———. 2019. “Constitutivism: On Rabbits, Hats, and Holy Grails.” Hebrew University of Jerusalem Legal Research Paper 19-14. Available at SSRN: http://dx.doi.org/10.2139/ssrn.3382239 Ferrero, L. 2009. “Constitutivism and the Inescapability of Agency.” Oxford Studies in Metaethics 4: 303–333. ———. 2018. “Inescapability Revisited.” Manuscrito 41: 113–158. Flowerree, A. K. 2018. “Epistemic Shmagency?” in Metaepistemology: Realism & Antirealism, ed. C. Kyriacou and R. McKenna, 289–310. London: Palgrave Macmillan. Gauthier, D. 1997. “Rationality and the Rational Aim.” in Reading, ed. J. Dancy Parfit, 24–41. Oxford: Blackwell. Katsafanas, P. 2013. Agency and the Foundations of Ethics. Oxford: Oxford University Press. Kolodny, N. 2007. “How Does Coherence Matter?” Proceedings of the Aristotelian Society 107: 229–63. Korsgaard, C. M. 1997. “The Normativity of Instrumental Reason.” in Ethics and Practical Reason, ed. G. Cullity and B. Gaut, 215–254. Oxford: Clarendon Press. ———. 2009. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press. Lindeman, K. 2017. “Constitutivism Without Thresholds.” Journal of Ethics and Social Philosophy 12(3): 231–258 ———. 2019. “Functional Constitutivism’s Misunderstood Resources.” Ethics 130(1): 79–91. ———. ms. “Constitutive Aims and Normative Explanations.” Littlejohn, C. 2013. “Are Epistemic Reasons Ever Reasons to Promote?” Logos and Episteme 4(3): 353–360. Markovits, J. 2014. Moral Reason. Oxford: Oxford University Press. McHugh, C. 2011. “What Do We Aim at When We Believe?” Dialectica 65: 369–392. Morton, J. 2010. “Toward an Ecological Theory of the Norms of Practical Deliberation.” European Journal of Philosophy 19(4): 561–584. Nolfi, K. 2018. “Why Only Evidential Considerations can Justify Belief.” in Normativity: Epstemic and Practical, ed. McHuhg, Way, Whiting. Oxford: Oxford University Press. Paakkunainen, H. 2018. “Doing Away with the Shmagency Objection to Constitutivism.” Manuscrito 41: 431–480. Rosati, C. S. 2016. “Agents and ‘Shmagents:’ An Essay on Agency and Normativity.” in Oxford Studies in Metaethics. Vol. 11, ed. R. Shafer-Landau, 183–211. Oxford: Oxford University Press. Schafer, K. 2017. “Rationality as the Capacity for Understanding.” Noûs. Advance Online Publication. doi:10.1111/nous.12231. ———. 2018. “Constitutivism About Reasons: Autonomy and Understanding.” in The Many Moral Rationalisms, ed. Karen Jones and Francois Schroeter, 70–90. Oxford: Oxford University Press. Schroeder, M. 2007. Slaves of the Passions. Oxford: Oxford University Press. Setiya, K. 2007. Reasons Without Rationalism. Princeton, NJ: Princeton University Press. Shah, N. 2003. “How Truth Governs Belief.” Philosophical Review 112: 447–482. Silverstein, M. 2015. “The Shmagency Question.” Philosophical Studies 172 (5): 1127–1142. ———. 2016. “Teleology and Normativity.” Oxford Studies in Metaethics 11: 214–240. ———. 2017. “Agency and Normative Self-Governance.” Australasian Journal of Philosophy 95(3): 517–528. Smith, M. 1994. The Moral Problem. Malden, MA: Blackwell Publishing. ———. 2009. “Reasons Without Rationalism After All.” Analysis 69(3) 521–30. ———. 2012. “Agents and Patients” Proceedings of the Aristotelian Society 112: 309–31. ———. 2013. “A constitutivist theory of reasons: Its promise and parts.” Law, Ethics, and Philosophy 1: 9–30. ———. 2019. “Formalism and Constitutivism in Kantian Practical Philosophy.” Philosophical Explorations 22(2): 163–176, DOI: 10.1080/13869795.2019.1599053 Tiffany, E. 2012. “Why Be an Agent?” Australasian Journal of Philosophy 90: 223–233. Velleman, J.D. 2000. The Possibility of Practical Reason. Oxford: Clarendon Press. ———. 2009. How We Get Along. New York: Cambridge University Press.

438

The aim of agency ———. 2015. Foundations for Moral Relativism: Second Expanded Edition. Cambridge: Open Book Publishers. http://dx.doi.org/10.11647/OBP.0086 Walden, K. 2012. “Laws of Nature, Laws of Freedom, and the Social Construction of Normativity.” Oxford Studies in Metaethics 7: 37–79. ———. 2018. “Practical Reason Not as Such.” Journal of Ethics and Social Philosophy 13(2): 125–153. Wedgwood, R. 2002. “The Aim of Belief.” Philosophical Perspectives 16(16): 267–297. Whiting, D. 2012. “Does Belief Aim (Only) at the Truth?” Pacific Philosophical Quarterly 93(2): 279–300. ———. 2014. “Reasons for Belief, Reasons for Action, the Aim of Belief, and the Aim of Action.” in Epistemic Norms, ed. Littlejohn and Turri. Oxford: Oxford University Press.

439

40 AGENCY AND MORALITY Christine M. Korsgaard

1 Action and the structure of moral theory According to a story widely accepted in the 20th and 21st centuries, there are three main domains of value relevant to morality. Goodness in the general sense applies to the ends that are worth aiming at or realizing, the states of affairs we should bring about for their own sakes. Rightness applies in the first instance to actions, and by extension to policies when the actions which those policies dictate are right. And virtue applies to people and to their dispositions and characteristics—specifically to the characteristics of people that incline us to do what is right. According to a related and also widely accepted story, we can categorize different types of ethical theory according to which of these domains of value contains the considerations that determine what is right. ‘Deontologists’ believe that actions are right or wrong in themselves, intrinsically, and that our pursuit of the good must be constrained by the necessity of doing only what is right. ‘Consequentialists’ believe that actions are right when they tend to maximize the production of good consequences. ‘Virtue theorists’ believe that the right action must be identified as the one that the virtuous person would choose to do.1 This story presupposes that all of these schools of thought mean the same thing by ‘action.’ Its simplicity falls apart when we begin to question that. Suppose that Jack drops his keys on the sidewalk, and Jill scoops them up and runs after him, in order to ensure that Jack is in possession of his keys. How much of that description is included in the correct description of Jill’s action? According to a broadly empiricist account, favored by consequentialists, Jill has an end— ensuring that Jack is in possession of his keys—and in light of that end she chooses to do something, namely scoop up the keys and run after Jack. The end, on this account, is something that stands outside of the action and purports both to explain and justify it. Perhaps Jill desires the end because Jack is her friend and she cares about him, or perhaps she desires the end because she sympathizes readily with anyone who loses their keys, or perhaps she holds the end as a matter of principle, believing that lost property should always be returned to its owner when possible. Usually, in the empiricist story, the end is not chosen, but if it is, it is chosen as a means to or an instance of some other end that in turn is not chosen. So the end is given to the agent, and she chooses an action that she thinks will bring about the end.

440

DOI: 10.4324/9780429202131-49

Agency and morality

According to consequentialists, whether the action is right depends on whether the end that it promotes is a good one, or perhaps on whether it is the best one that she could have achieved in the circumstances. As John Stuart Mill says, ‘All action is for the sake of some end, and rules of action, it seems natural to suppose, must take their whole character and color from the end to which they are subservient’ (Mill 1979, p. 2). What I am somewhat tendentiously going to call the ‘deontological’ account—because it is Kant’s, and Kant is usually taken to be the primary exemplar of deontology—differs from the consequentialist account in two ways. According to the deontological account, our ends are not given to us, but rather are chosen, and in particular, they are part of what we are choosing, when we choose an action. In other words, the object of Jill’s choice, and her action, is ‘to run after Jack in order to ensure that he remains in possession of his keys.’ Jill’s end may have suggested itself to her by desire or sympathy or principle, but she did not have to pursue it. No causal force such as natural desire impelled her to pursue it, nor did she choose it independently of the action, although she may have been on the lookout for actions through which she might promote this sort of end. But the end is not in any way set for her when she decides what to do. Instead, she makes a calculation according to which, given that Jack needs his keys, and given the possible moral and prudential costs of running after him, running after him in order to ensure that he has his keys is a thing worth doing, a good thing to do. That is the second difference between the two accounts: on what I am calling the ‘deontological’ account, rightness is a form of goodness, a form of goodness that applies to actions, just as virtue is a form of goodness that applies to people. Or if it is not, our account of which actions are right is derived from an account of what makes an action morally good. I say this because we sometimes use the word ‘right’ for actions that involve the same outward act that morally good actions do, whether they are themselves morally good or not. We are using the term this way when we say that someone ‘does the right thing for the wrong reason,’ as we might if Jill, for example, ran after Jack in hopes of getting a reward for retrieving his keys. So here are the two possibilities we have canvassed so far (Table 40.1). Table 40.1 The Consequentialist Account: (Running after Jack) ↑ What Jill chose to do; her action

(in order to ensure that Jack is in possession of his keys) ↑ End that moved Jill to do it The Deontological Account:

(Running after Jack in order to ensure that Jack in in possession of his keys) ↑ What Jill chose to do; her action What moved Jill to do it: the goodness of the action

To make it a little easier to talk about this, from now on I am going to call things like ‘running after Jack’ an ‘act’ and things like ‘running-after-Jack-in-order-to-ensure-thatJack-is-in-possession-of-his keys’ an ‘action.’ Then we can say that according to the consequentialist account, what we choose are acts, while deontologists think what we choose are, in this technical sense, actions. 441

Christine M. Korsgaard

I’ve called the second possibility the ‘deontological’ account in spite of the fact that either a deontologist or a virtue ethicist might conceivably hold it. Kant and Aristotle are often supposed to represent deontology and virtue ethics, respectively, yet they are the two main exemplars of the kind of account I have just described. Both of them begin their accounts of what is right, in the sense of we ought to do, from an account of the moral goodness of actions. I’ve also called it the ‘deontological’ account in spite of the fact that it is not clear that everyone we would style a ‘deontologist’ holds it.2 The fact is that in the tradition of moral philosophy, many philosophers have simply been unclear about what they mean by ‘action,’ and in particular about whether the end is included in the content of the action or not. When we turn to Aristotle and Kant, however, we find accounts according to which the thing that the agent chooses, and the thing that has moral value, is what I have been calling an ‘action,’ that is, an act-for-the-sake-of-an-end. Kant believed that the sort of thing that an agent chooses, and therefore the sort of thing that is a candidate for being morally good or bad, is the sort of thing that we describe by giving what he called the agent’s ‘maxim.’ Characteristically, the maxim, and therefore the action, include both the act that the agent performs and the end for the sake of which he performs it. It has to include both of those, because the way that we test a maxim to determine whether it is in accordance with duty is by applying the categorical imperative, the moral principle that tells us to act only on those maxims which we can will to be universal laws (Kant, 1998, 4:421–424). The question raised by the categorical imperative is whether the agent making the choice could will that every agent who has that end should pursue it by those means at the same time as he himself wills to pursue that end by those means, without generating a contradiction in his will. If the act can be a successful method of achieving the end only if not everyone attempts to use it as a method for achieving that end, then the maxim and so the action are ruled out as wrong.3 So, for instance, Kant argued that if everyone borrowed (or rather tried to borrow) money on the strength of false promises, possible lenders would not accept promises as a basis for lending money, and the act of making a false promise would no longer work as a means of getting a loan (Kant, 1998, 4:422). An agent who willed that situation at the same time as he himself willed to get money by means of a false promise would produce a contradiction in his own will, because he would be willing a situation in which the method he proposes to use to achieve his own end would no longer work. On the other hand, if the maxim of the action could serve as a universal law, it has what Kant calls the ‘form’ of a law—an idea I will come back to—and the action it describes is therefore permissible and to that extent morally good. When an action is rejected as impermissible, the opposite action is a duty and is in a stronger sense morally good. The parallel in Aristotle’s ethics to the Kantian maxim is the logos—a description of the action that contains or expresses the reason why it is done. Aristotle is more careful than Kant to ensure that all of the aspects of the action that might be relevant to its goodness are included in the logos. In a number of passages in the Nicomachean Ethics, Aristotle tells us that the good person acts on the right reason or right principle—the orthos logos—specifying that this means that the good person does the right act at the right time, in the right way, and for the right end (Aristotle, 1984. NE 2.9 1109a25–30). It is clear from these formulations that Aristotle thinks that the moral goodness of the action somehow rests in the way these factors are related to each other. Since the relation between the parts of something constitutes what Aristotle calls its ‘form,’ a good action is one that is well-formed.4 Kant thinks that the way the parts of a maxim, namely the act and the end, have to be related to each other for an action to be well-formed is ‘so that the maxim can serve as a universal law.’ Aristotle is not specific about what exactly makes an action well-formed, but since Aristotle also 442

Agency and morality

describes the first principles of our actions as ‘universals’ (Aristotle 1984 NE 6.8 1141b26; NE 6.81142a12–21; NE 7.31147a105), it is tempting to think that he shares with Kant the view that ‘universality’ of some sort is what gives an action the correct form, the form that makes it good or in Aristotle’s language ‘kalon’ (noble). The first difference between the consequentialist and what I am calling the deontological account, then, is that they identify different objects as the ‘action’ that may be right or wrong. For consequentialist, it is what I have called the act, the movement through which we promote the end, that is right or wrong. But on deontological account it is the whole action, the act-for-the-sake-of-the-end, that has moral value. The second difference concerns our motive for doing the right ‘action.’ On the consequentialist account, we know what moves Jill to do what that account identifies as her action, which on that account is the same as her act (running after Jack). It is her end: she is moved to run after Jack in order to ensure that Jack has his keys. On the deontological account, what is Jill’s motive for the choice of what that account identifies as her action (running-after-Jack-in-order-to-ensurethat-Jack-is-in-possession-of-his-keys)? The answer is that the agent sees the action as a whole as something that is good to do. Because it is good in virtue of the relations between its parts, we can even say, as deontologists are supposed to say, that the action is intrinsically good.5 These contrasts explain how John Stuart Mill can say that ‘all action is for the sake of some end,’ while Aristotle insists that ‘while making has an end other than itself, action cannot; for good action is itself is its end’ (Aristotle 1984 NE 6.5 1140b6–7). But they leave us with a question: what, according to Kant and Aristotle, makes an action morally good? What makes an action well-formed, and why does an action that is well-formed count as a morally good action?

2 Constitutive standards and the goodness of action When we say that something is good, we usually mean one of three things: that it is good of its kind, that it is good for some specific purpose, or that it is good as an end. If I say without further elaboration that a certain shoe is good, you will assume that it protects your feet from the environment while being comfortable and supportive, that it does not produce blisters, that it makes you less likely to slip on tenuous surfaces, things like that. These standards of goodness are derived from the function of a shoe, and may therefore be identified as ‘internal’ or ‘constitutive’ standards for shoes: normative standards that arise from the nature of the object to which they are applied. A good X is one that serves the function of an X well, and the things that I have mentioned above, protecting your feet and your footing and so on, are the functions of shoes. If I say that a certain shoe makes a good doorstop, on the other hand, you will assume that it is massive and heavy enough to hold a door open, though massiveness and heaviness are not ordinarily especially good properties in a shoe. This is an ‘external’ standard, one we apply to the shoe only in virtue of some special purpose we have in mind for it. In both of these cases we are normally talking about things that are good as means to other ends—to the tasks of protecting one’s feet, the usual purpose of shoes, or holding a door open, a specific purpose in the case at hand. Finally, if I say that happiness or virtue is good as an end, you will assume that these things are worth aiming at for their own sakes. Which of these things do we have in mind when we say that an action is morally good? At first glance, it might seem to make better sense to talk about the goodness of acts than it does to talk about the goodness of actions. Acts do seem to have a general function—namely to produce ends. An act, after all, is not just a movement, but a movement undertaken by an agent, and an agent who undertakes an act usually—some would say always—has an end 443

Christine M. Korsgaard

in view. Those who suppose that agents always act ‘under the guise of the good’ may even suppose that the agent always has some good end in view. If that is right, we might think that a good act is one that is effective, or even one that is effective for good. A little thought, however, shows us why this cannot be quite right. All acts effect something or other, because all acts are movements embedded in the causal network. If nothing else, the movements of my body displace molecules in the air around me, but that doesn’t make those movements ‘effective’ in the sense we need here. When we say that an act is effective, we mean that it tends to effect the end that the agent had it mind to produce through his movement. So it is not a mere act, but only an act within the context of a whole action, that can be said to be effective. So there is at least one standard for good actions: a good action is one whose act tends to achieve the end of that action, the one the agent has in view. In other words, the action must accord with what Kant calls the hypothetical imperative: the act is a means to the agent’s end (Kant 1998 4:414–415). You might wonder why I have said that the good action is one whose act ‘tends to’ achieve its end, rather than one that just does achieve its end. Isn’t success better than a mere tendency? It is not, because the kind of efficacy that we associate with agency is not exhausted by the idea of the agent actually bringing the end about. In fact, an agent who brought about his end only accidentally—say by a deviant causal pathway—would have failed as an agent. If I fire my gun wildly astray, but the bullet ricochets off a cast iron fence and happens to hit the target in exactly the spot that I intended, I have not made a good shot. Indeed, once I understand the situation, I would feel that I have been the beneficiary of a lucky accident, and that I didn’t really do anything at all. To be successful in action is not merely to do something that brings about your end. To be successful in action you have to use a means that makes you into the kind of thing that reliably achieves that sort of end. The efficacy that makes an action good is not just the efficacy of the act. It is the efficacy of the agent. This, in turn, has a further implication. In order for an action to be efficacious in the right way, the efficacy that it exhibits has to be the efficacy of the agent considered as an agent. This implies not only that the means must tend to bring about the end, but that the action as a whole has to be attributable to the agent as such. The source of the action has to be the agent herself. It is obvious that it cannot just be a force at work on the agent from the outside. If I knock Jack off a cliff by shoving Jill into him when she is standing behind him and he is on the edge, I make Jill’s body into the sort of thing that reliably knocks people off cliffs, but she has not made herself into that, and her agency is not involved. It is perhaps less obvious but nevertheless also true that it also cannot be just a force at work on the agent from the inside, unless we have a quite particular conception of what counts as ‘inside.’ If Jill shoves Jack off a cliff under the influence of a twitch or a spasm or while sleepwalking or under the influence of a posthypnotic suggestion, there is a sense in which she has done the shoving herself, but it is not the right sense. The action must originate in a choice that somehow represents the activity of the agent herself. Before we consider how that condition is met, let’s take stock. According to the deontological account, actions, acts-for-the-sake-of-ends, may be morally good or bad. This must mean they are good of their kind, good for some special purpose, or good in themselves. In order for actions to be good of their kind, they must have a function, which good actions succeed in serving or serve well. Surprising as it may seem, we have now identified a function for actions. An agent who undertakes an action undertakes to achieve some end in the world through her own efficacy. The function of an action, therefore, is to make its agent efficacious for a certain end. In order to do that, the action has to meet two standards. First, the act must be related to the end in such a way that it makes the agent efficacious, that is, it 444

Agency and morality

must make the agent into the kind of thing that tends to achieve that end. And, second, the choice must be related to the whole action in a way that ensures that it is the agent herself, and not some alien force working in her or on her or through her, that is in this way efficacious. Putting these two conditions together, she must make herself into the kind of thing that is efficacious for that end. An action that meets these two standards succeeds in making its agent efficacious, and therefore is good of its kind. It is good considered just as an action. We have already seen how to meet the first of these two standards: an act is related to its end in a way that makes the agent efficacious if the action as a whole conforms to what Kant called hypothetical imperative, the principle of taking the means to one’s ends. How do we guarantee that it meets the second standard, that it is the agent’s efficacy as an agent that is in this way established? Aristotle argued that the action must have its source in the agent’s character, his virtues and vices, in order for the action to be an expression of his own choice, and therefore of his own efficacy (Aristotle 1984 NE 3.2). That secures us part of the conclusion we want here: chosen actions are by their very nature morally good or bad. Moral goodness or badness is not some external standard that happens to apply to human actions, the way that the standards for serving as a good doorstop might happen to apply to a shoe. To put it another way, morality is not a special purpose we happen to have for actions. Actions have moral properties considered simply as actions. But Kant’s theory goes even further. According to Kant, an action has its source in the agent herself only if it has its source in the agent’s autonomy, that is, only when she acts on principles which she herself recognizes, and so legislates to herself, as universal laws. But if the agent is thinking clearly, the only principles that she can recognize as universal laws are the principles that really can be universal laws. That means that only the morally good agent is genuinely efficacious, that is, only she makes herself genuinely efficacious through her actions. Her actions make her efficacious because they achieve the end she has in mind, and because achieving that end by using that means is a principle of action that she herself endorses. To put it more intuitively, she does what she herself thinks that she and other agents should do. It follows that only the morally good action is good as an action, good of its kind. Autonomy, which for Kant involves acting on moral principles, is a constitutive standard for actions, one that they must meet in order to be good of their kind. It is a difficult question whether Aristotle also holds a view of this kind. It depends on whether his view allows us to see moral badness as a form of what Kant would have called ‘heteronomy,’ the enslavement of the agent to forces outside of her own will (Kant 1998 4:432–433). Aristotle’s identification of akrasia, or moral weakness, as a condition that is different from moral badness complicates this question (Aristotle 1984 NE VII). But A ristotle’s view raises another question we must deal with here. As I noted earlier, while Mill suggests that actions (by which he means what I have called ‘acts’) are good or bad as means to good ends, Aristotle says that good actions are ends in themselves. What are we to make of Aristotle’s claim? In Kant’s theory, a good action is one whose maxim can be a law, and a required action is one whose maxim must be a law, because the opposite maxim is impermissible. If the maxim of an action must be a law, then the action is one that it is necessary for the agent to do. In that sense, it is the nature of this action that it is something that is to be brought about for its own sake, and the things that are to be brought about for their own sakes are good. But I think a comparison that Aristotle often makes gives us a better way to put the point: the comparison between health and virtue. Healthy activity is a good thing for its own sake because there is a way in which our lives consist in physical activity, and healthy activity is 445

Christine M. Korsgaard

the excellence of physical activity. In the same way, morally good action is a good thing for its own sake because there is a way in which our lives consist of rational chosen action, and morally good action is the excellence of rational chosen action. Doing what is right is good simply because it is a way of living well.

Related topics Agency, functions, and teleology; Rational agency; Agency and the will; Agency and responsibility; Agency and autonomy; Agency and reasons; Agency and practical reasoning; Agency and normativity; The aim of agency.

Notes 1 ‘Must be identified’ deliberately leaves it open whether the identity in question is metaphysical or epistemological. Later I will suggest that, at least in its Aristotelian versions, it can only be epistemological. That means that Aristotelian virtue theory is not actually an alternative to consequentialism and deontology, which are views about the metaphysics of the right. 2 An alternative strategy is to deny that Kant is a deontologist. See Herman 1993. 3 For a defense of this interpretation of categorical imperative, see Korsgaard 1996. 4 If this is Aristotle’s view, then he does not hold that what makes an action right is that the virtuous person would choose it. What makes it right is that it is well-formed. But Aristotle does hold that the virtuous person is the only reliable judge of whether an action is well-formed. See note 1. 5 A caveat is in order. What makes the action good on these accounts is the relation between its parts, and in that sense its goodness is intrinsic, but we cannot ascertain whether that relation holds simply by examining the action itself, as we might expect to do if the goodness is intrinsic. For instance, Kant’s argument against false promising depends on the idea that if everyone made false promises, potential lenders would come to realize that fact. That obviously depends on causal relations that are not intrinsic to the action.

Further reading 1 The volumes below contain both classical accounts and contemporary essays on the three types of ethical theory. Darwall, Stephen, editor. Consequentialism. Oxford: Blackwell Publishers Ltd, 2003. Darwall, Stephen, editor. Deontology. Oxford: Blackwell Publishers Ltd, 2003. Darwall, Stephen, editor. Virtue Ethics. Oxford: Blackwell Publishers Ltd, 2003. 2 The book below defends the Kantian version of a constitutivist account of the moral goodness of action. Korsgaard, Christine M. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press, 2009. 3 The books below defend another version of constitutivism, one in which the nature of action provides the standards for the rationality of action. Velleman, J. David. Practical Reflection. Princeton: Princeton University Press, 1989. Reissued by the Center for the Study of Language and Information, 2007. Velleman, J. David. The Possibility of Practical Reason. 1st edition. Oxford: Oxford University Press, 2000. 2nd edition, Michigan Publishing Services, 2015. Velleman, J. David. How We Get Along. Cambridge: Cambridge University Press, 2009.

References Aristotle. (1984) The Nicomachean Ethics. Translated by W. D. Ross, revised by J. O. Urmson. in The Complete Works of Aristotle: The Revised Oxford Translation. Edited by Jonathan Barnes. Princeton, NJ: Princeton University Press. References to Aristotle’s works are given by the standard Bekker page, column, and line numbers.

446

Agency and morality Herman, Barbara. (1993) “Leaving Deontology Behind” Chap. 10 in The Practice of Moral Judgment. Cambridge, MA: Harvard University Press, pp. 208–240. Kant, Immanuel. (1998) Groundwork of the Metaphysics of Morals. Translated by Mary Gregor. Cambridge: Cambridge University Press. References to Kant’s works are given in the standard way by the page numbers in the relevant volume of Kants gesammelte Schriften, which appear in the margins of most translation. Korsgaard, Christine M. (1996) “Kant’s Formula of Universal Law” Chap. 4 in Creating the Kingdom of Ends. Cambridge: Cambridge University Press: 106–132. Mill, John Stuart. (1979) Utilitarianism. Edited by George Sher. Indianapolis: Hackett Publishing Company.

447

41 AGENCY IN THE LAW Gideon Yaffe

1 Introduction One might think that to inquire about the law’s conception of agency is to ask for the legal definition of the term “agent.” Maybe you can just look it up in some law book somewhere. And there is something to this. For instance, a person can be criminally punished only for an act or omission, and both terms are fairly well-defined. An act is a willed bodily motion, while an omission is the absence of an act that the person could have performed. You cannot be punished for what happens to you, or for being a certain way. Only exercises of agency are appropriate objects of criminal liability. It follows from this explicit legal doctrine that under the law, an agent is simply a creature that is capable of action or omission in these senses. If, and only if, you can will and guide your body in accordance with your will, are you an agent. This notion of agency does legal work outside of the criminal domain, as well. No court of law would hear a suit against a tornado for destroying a house. Tornadoes are not the kinds of entities that can be brought to court; they are not agents in the simple sense just defined. If we want to know more about the simple conception of agency that excludes tornadoes from court and limits criminal liability, we need to set the law aside and ask what it is to will something, or what is to possess an unexercised ability to do something. The law simply draws on the ordinary notions. Both of these are interesting questions about which progress can and has been made by philosophers. But neither are distinctively questions about the law. Is there no more than this to the idea of the law’s conception of agency? There is more. The law sometimes draws distinctions in liability between people who are equally capable of action or omission on the grounds that one, rather than the other, is more the agent of a particular event for which liability is being assigned. The law, for instance, sometimes assigns liability for a third party’s harm to one person rather than another on the grounds that the harm is more the one’s doing than the other’s even where both engaged in willed bodily motions, or omissions, that causally contributed to the harm. Our focus here will be on a deeply entrenched legal practice that involves making such discriminations between agents: the practice of classifying some who causally contribute to a harm as the “proximate cause” of the harm, to the exclusion of other causal contributors. Through its doctrine of proximate causation, the law draws a line between those who are more and less active with

448

DOI: 10.4324/9780429202131-50

Agency in the law

respect to the violation of a legally protected interest and then predicates important differences in government treatment of the parties involved on that classification. The goal here is to uncover a somewhat hidden conception of agency in the law, a conception distinct from, although consistent with, the law’s explicit view that agents are creatures that can move their bodies in accordance with their wills.

2 The agent-of The idea of being the agent of a particular event—the idea that the event is one’s doing—is familiar from ordinary contexts. It is an essential part of the assessment of moral responsibility both for harms and for benefits. It is the agent of your loss who owes you an apology; it is the agent of the benefit you accrued who is owed your gratitude. The ordinary notion of agent-of has both an intrinsic and a comparative aspect. Sometimes we fix our gaze only on a person and a harm and we ask whether the person is an agent of the harm. This is to consider agency-of intrinsically. We say, “This was your doing!” But sometimes we focus on more than one person and on a single harm and we ask whether one of those people is more the agent of the harm than the others. This is to consider agency-of comparatively. We say, “This was more your doing, than his!” Any account of what an action is, in contrast to events that are not anyone’s doing, naturally suggests a theory of agency-of: you are the agent of an event just in case it is caused, in a predictable way, by your action. We are not the agents of freak accidents that our actions cause. But anything else that we cause is something with respect to which we are an agent. The exclusion of freakish or unpredictable events from the class of things of which a person is an agent is motivated by a thin conception of control. On this thin conception, one can exercise control over an event by making a small contribution to a natural process that one was in position to realize might result in the event. When one instead makes a small contribution to a process that brings about the event in a way that one could not, even in principle, have anticipated, then one is not exercising this thin form of control over the event. This notion of agency-of is easy to satisfy. A very large number of events that people are not at all responsible for are things of which they are agents. If agency-of, then, is to be a useful concept, one that can carry moral or legal weight, it must be because of its comparative aspect. When there are multiple agents of a particular harm, it can sometimes seem appropriate for the person who is most the agent of the harm to bear all the burden of responsibility for it, whatever that burden involves. It often makes sense for the person who is most the agent of a harm to say, “I apologize on behalf of all of us.” Consider the following general rule: among the agents of a harm, the one who is most the agent of it should bear the burdens of responsibility for it (whatever they are). While this rule probably admits of exceptions, it is nonetheless a general rule of morality. When we need a spokesperson, as it were, for the group of people who are agents of another’s harm, ceteris paribus the one who is most the agent of the harm is the best choice for the job. Given this, it would not be surprising if the law should try to conform to this general rule. Often, the law assigns a single party the job of bearing a joint, societal burden to remediate a harm. Some of any citizen’s interests are legally protected; they are our joint responsibility. When such an interest is violated, the citizen has suffered a setback that the state should somehow remedy, if it can. The remedy can take many different forms. Perhaps the setback is to be remediated through a cash payment for the loss, or perhaps through collective 449

Gideon Yaffe

acknowledgment of the loss’s importance through punishment of someone. But whatever the remedial burden is, there is the question of who (if anyone) should bear it. Who should the state give the job of remediating a setback to a legally protected interest? The bearer of that burden is the person to whom liability is assigned, where the state will exercise coercive force against the bearer of liability to assure that the burden is in fact borne by that party. The state remediates the suffered setback by forcing one of its members, namely the person to whom liability is assigned, to remediate it. There is something fitting about assigning legal liability for a particular setback to a legally protected interest to a party only if the party is the agent of that setback. And there is something fitting, often, about assigning liability to that person who is more the agent of it than others. When the law does that, it mirrors an aspect of ordinary morality, and that is something the law strives, and ought to strive, to do. The result: when what makes moral sense is for liability to be assigned to the person who is most the agent-of a particular setback to a legally protected interest, the law needs a theory of comparative agency-of; it needs an account of the conditions under which one person is more the agent of an event than another.

3 Proximate causation Part of the law of “proximate causation” encodes a partial theory of intrinsic agency-of, part a partial theory of comparative agency-of. For many purposes, under the law, a person, such as a defendant in a lawsuit, is liable for a particular occurrence, most typically a harm, only if the defendant “proximately caused” it. Whether or not a particular person is a proximate cause of a harm, in the law’s technical sense, will turn on the question of whether she, or instead some other party, is more the agent-of that harm. The law uses the concept of proximate causation in part as a way of making it possible for legal actors like jurors and judges to provide a uniform answer to a more difficult moral question: Is the harm more the defendant’s doing or is it more someone else’s doing? Only if it is no more others’ doing than her doing is she potentially the appropriate party to remedy it. Seeing that criteria for being the agent-of an event are buried in the criteria for proximate causation requires a little background about the law. Under the law, a necessary condition for proximate causation is being a “cause-in-fact”, which is ordinarily operationalized in counterfactual terms. Typically, that is, the defendant is a cause-in-fact of an event if and only if the event would not have occurred had the defendant not performed some act or omission for which liability is being sought.1 But not every cause-in-fact is a proximate cause; various other conditions must be met. For our purposes there are two particularly important conditions. First, to be a proximate cause of an event, the agent’s action must not be a cause-in-fact of the event because of some freakish turn of events. It cannot be that the event came about, thanks to the act, in some entirely unpredictable way. Standardly, this is put as a requirement of “reasonable foreseeability”: only if the event was a reasonably foreseeable consequence of the act is the agent a proximate cause of the event. Recall that this very idea is a natural way to build a theory of intrinsic agency-of out of an account of action. If an action is a willed bodily motion or omission, then a person is the agent of anything that is caused by that, provided it does not come to pass in some entirely unpredictable way. The reasonable foreseeability limitation on proximate causation institutionalizes this view of intrinsic agency-of. Any legal doctrine that requires proximate causation for liability excludes from liability those who are not the agent of the relevant event for either of two reasons: their willed bodily motion (or omission) was not a cause-in-fact of the event, or it was a cause-in-fact of 450

Agency in the law

the event but in a freakish or unpredictable way, and so the agent did not enjoy even a thin sense of control over the event. Where the first requirement for proximate causation—non-freakish causation-in-fact— concerns the intrinsic side of agency-of, the second bears on the comparative side. The defendant is not the proximate cause of the event if it occurred, thanks to the voluntary intervention of another agent. Say that D1 acts and then, at a later time, D2 does. Both acts are causes-in-fact of a harm. Under the law, D1’s act is a proximate cause of the harm only if D2’s act does not constitute a “voluntary intervention”. If, for instance, D1 pays D2 to kill V, and D2 does so, D1 is not the proximate cause of V’s death, even though he is among the causes-in-fact of the death. The hitman’s act is a “voluntary intervention”. The idea is sometimes put informally by saying that other agents can break chains of proximate causation. Adjudicating what constitutes a “voluntary intervention” amounts to answering the following question: When more than one person causally contributes to an event, what makes one of them more the agent of it than the other? The question is which of the two is better suited to be put into service by the state in fulfilling its obligation to remediate the harm, given that, as a moral matter, ceteris paribus that job should be done by the person who is more the agent of the event. For instance, consider a well-known case.2 Vincent Acosta was sitting behind the wheel in a stolen car when approached by police. Acosta drove away and the police chased him. As the chase continued, two police helicopters joined the pursuit. One of the pilots made a peculiar maneuver, crashed into the other helicopter, and three people died. Acosta was charged with murder—a crime of which he was guilty only if he proximately caused the deaths. Acosta’s act of driving away from the police and leading them on a chase was a causein-fact of the deaths: had Acosta simply submitted to the police when they approached his car there would have been no chase, and so no involvement of helicopters, and so no pilot maneuver, and so no deaths. But did Acosta proximately cause anyone’s death? Under the law, the answer turns, in part, on whether the pilot’s maneuver, which was also a cause-in-fact of the deaths, was a “voluntary intervention”. The question worth asking is this: Was the pilot (who himself was among the dead) more the agent of the deaths than Acosta? If so, then Acosta is not to be held liable for the deaths. He is to do the state’s work of remediating the harm, by being subjected to criminal punishment for murder, only if no one is more the agent of the event than he. The most commonly accepted legal rule about such cases is that a person’s act is the kind of voluntary intervention that breaks the chain of proximate causation only if it is performed with the intention to bring about the relevant harm. Since the pilot did not undertake the maneuver with the intention of killing himself and the other two people, his maneuver was not a “voluntary intervention” of the sort that would strip Acosta of the title of “proximate cause”. That is, the pilot was no more the agent of the deaths than Acosta, thanks to the fact that he did not intend the deaths. This is what distinguishes the case from the murder for hire: the hitman intends the death she is paid to cause, which is why the person who pays is not the proximate cause of the death (although liable nonetheless through principles of liability, such as accomplice liability, for which proximate causation is not necessary).3 Notice that this does not imply that Acosta is to be held liable for the deaths. Liability requires more than being most-the-agent-of the setback to the legally protected interest in survival. There’s more to being a murderer than being the agent of a death. You also need to have some problematic mental state at the time of the action that leads to the death, such as an intention to kill or an awareness that one is doing something one shouldn’t and thereby risking a death. You need to have what the law labels “mens rea”, or a guilty mind. To deny 451

Gideon Yaffe

that the pilot is a “voluntary intervener” is to say only that Acosta cannot escape liability on the grounds that someone else is more the agent of the deaths than he is; he could perhaps escape on other grounds, such as on the grounds that he lacked mens rea. And, in fact, that is what the court ruled: it ruled that although Acosta was the proximate cause of the deaths, he did not act with the mental state needed to be guilty of murder. What theory of comparative agency-of is being assumed in the rule that the second, later actor breaks the chain of proximate causation only if she intends the harm? The rule isolates three factors as contributing to making a person more an agent of some events than others. The first factor is commitment to the occurrence of the event. The agent intends some of the events of which he will be an agent, thanks to his willed bodily motion. Among those events that are caused-in-fact by his act in a non-freakish way, some are intended. Others in this class the agent believes will come to pass, with varying degrees of certainty. Some he has no idea will happen. One is most committed to that which one intends. One has a lesser commitment to that which one believes will be caused by one’s act, and the commitment decreases with one’s degree of certainty. And one has little or no commitment to the occurrence of events that one does not expect to occur, thanks to one’s act. The thought is that the more one is committed to an event of which one is an agent, the more one is the agent of the event. So, if the second actor (the one who acts later, like the pilot in Acosta) intends the event, he is thereby at least as committed to its occurrence as the first actor. The result is that by that measure, the first is not more the agent of the event than the second, if the second intends the event. Notice, however, that if this were the only factor relevant to comparative agency-of, then intending agents would not invariably break chains of proximate causation. Rather, when both agents intend the event, as in the murder for hire case, the two would be equally the agent of the event. Since the first agent to act is released from liability on this ground only if there is another person who is more the agent of the event than she is, there must be some other relevant factors. The second relevant factor is temporal proximity. When two agents have an equally strong intention-based commitment to an event, and all else is equal, the one whose action is closer in time to the event is more the agent of it. This is why the hitman is more the agent of the death than the person who pays him, even though both intend the death. And it is why we know that the intending second actor is more the agent of the event than the first actor even if we do not know how committed the first actor is to the event. Since we know the second actor intends, we know that he is at least as much the agent of the event as the first actor on that score. And since we know that he is the second actor, and so his action is closer in time to the first, we know that he is more the agent of the event than the first actor. However, there must be at least one more factor if the legal rule is to make sense, given that it applies also in the case in which neither the first nor second actor intends the harm. One must be more the agent of events that one does not intend the more one is trying to do something wrongful. Neither Acosta nor the pilot is trying to kill anyone. But Acosta is trying to escape the police, which is wrongful, while the pilot is trying to apprehend a criminal, which is not. On this score, then, Acosta is more the agent of the resulting deaths than is the pilot. The legal rule that we are discussing—according to which later actors who intend harms undermine earlier actors’ claims to be the proximate cause of those harms—encodes no clear theory of how these three factors are to be weighed in the determination of who is more the agent of a particular harm. The court concluded that Acosta was more the agent of the deaths than was the pilot, thereby denying Acosta relief from murder liability on those grounds. For them to have reached this conclusion, they must have given much greater weight to 452

Agency in the law

the third factor—the wrongfulness of that which the agent is trying to do—than they did to the temporal proximity factor, which is independent. After all, the pilot’s act was much closer in time to the deaths than was Acosta’s, although much less wrongful. Was that the legally mandated conclusion? To know, we would need to know how these factors are to be weighed, and it is very hard to tell. Is the absence of a theory of weighting a flaw in the law? If the law were intended to be a philosophical theory, and so to be assessed as we do philosophical theories, it would be a serious flaw. Someone who wants to know what it is for one person to be more the agent of an event than another does not just want to know what qualities combine to constitute that—although identifying such qualities does mark philosophical progress; she also wants to know how they combine so as to constitute more or less agency-of. But the law of proximate causation is not intended as a philosophical theory of comparative agency-of. Rather, the law is a tool for helping people to overcome bias and self-interest in the allocation of the labor of discharging remedial obligations that governments and citizens owe. The question, then, is to what extent it does so adequately by constraining legal actors tasked with judging questions like whether Acosta was less the agent-of the deaths than was the pilot. To answer that question, we need to answer two others: Are the factors that the law takes to be relevant (commitment, temporal proximity, degree of wrongful trying), actually relevant to the question of how much an agent of an event a person is? And, can legal actors be trusted to weigh those factors as they ought to be weighed without guidance about how to do so? The second question, concerning the trustworthiness of legal actors, cannot be answered from the armchair; it’s an empirical matter, turning on the capacity and willingness of people to make accurate judgments about comparative agency-of, when constrained in what factors to consider and on which to predicate their judgement. But the first question, concerning whether the factors the law identifies as relevant actually bear on the question of comparative agency-of, is amenable to philosophical assessment. Answering it requires determining, from an extra-legal point of view, when one person is more the agent of an event than another (when they are both the agent of it non-comparatively), and what makes them so. The first two factors—commitment and temporal proximity—bear some intelligible relation to a thin notion of control. They are plausibly thought of as factors that increase the degree to which a person is intervening in a normal process in such a way as to direct it toward an event. Those who are committed to an event, stand behind it, as it were. The world aligns with them when the event comes to pass. This is part of what is, intuitively, involved in exercising control over an event. So, where there is greater commitment, there is greater control, ceteris paribus, in this thin sense. Temporal proximity, too, is linked to the idea of control. The farther in the future an event is, the more things can happen so as to derail it from occurring. The more one’s act ensures that the event will occur, the more one is exercising control over it. So, temporally proximate acts involve greater control over that which they cause than do less temporally proximate acts, ceteris paribus. To be clear: it is possible that in the end neither of these links will hold up philosophically. It is possible that the links just described between, on the one hand, the factors the law takes to be relevant to proximate causation and, on the other, our ordinary thin notion of control will turn out to be present even ceteris paribus. If not, then the law of proximate causation should be reformed. The point for our purposes here, however, is that it is not clearly an error to guide legal actors in making judgments of comparative agency-of through appeal to these two factors. There is at least a case to be made that where they are present to a greater degree, there is also more of the sort of control that we take to be present to a greater degree when one person is more the agent of an event than another. 453

Gideon Yaffe

By contrast, the third factor—where there is greater wrongful effort there is greater agency-of—is simply irrelevant to comparative agency-of. Some agents exercise control through permissible tryings others through impermissible; but there is no reason to think that there is greater control being exercised by the impermissible triers than the permissible, not even ceteris paribus. So, to the extent that the law predicates the presence of proximate causation on this factor, and so predicates the presence of greater agency-of on this factor, it is making a philosophical error. Recall, however, why the law enquires about proximate causation, when it does: the goal is to identify, among those who are the agents-of an event, which party is the most fitting, morally speaking, to bear the societal burden of remediating the harm. Ceteris paribus, this is the party who is most the agent-of the event. But all is not always equal. When neither of two parties is trying to do the harm, but both, nonetheless, are the agent of it—as in the Acosta case—something is off, morally speaking, if the party who was not trying to do anything wrong ends up bearing the societal burden of remediating the harm. It is better, in such a case, for the party engaged in wrongful trying to bear the burden. In other words, the legal doctrine of proximate causation, in its appeal to wrongful trying as a relevant factor, directs legal actors to conform to a distinct moral principle for the allocation of remedial labor when it does not make sense to give the job to the person who is most the agent-of the violation of a legally protected interest. The legal doctrine, that is, guides our practices toward assigning liability to the person who is most the agent-of the relevant event, except when it makes sense to assign liability to someone else. This is what legal doctrines should do, even if they are very easily mistaken for bad philosophical theories.

4 Conclusion It is plausible to hold that to be the agent of an event is to cause it in some special way. The hard philosophical problem is to specify what that special way is. What conditions need to be met, in addition to causation by action, for a person to be more the agent of an event than another? This problem is of deep intrinsic interest. It matters in itself to the basic project of understanding human nature. It is also of instrumental importance, without considering the law: given that agents of events are candidates for moral responsibility for those events, to understand morality and moral responsibility, we need to understand agency-of. But the concept is not just of importance intellectually. It matters also to the grand project of conforming legal practice to morality. As suggested here, it matters crucially to the allocation of a very particularly kind of morally justified forced labor: the labor of remediating setbacks to interests that are our collective obligation to remediate. All else equal, that labor should be allocated to the party who is most the agent-of the setback. Make no mistake: it is not only in its doctrine of proximate causation that the law enshrines theories of agency. But the project of mapping the law’s conception of agency requires taking one legal rule at a time. We need to determine, in each individual case, how the contours of the legal rule involve an effort to conform to an ordinary agential concept and therefore involve a theory of the structure of that concept.

Related topics Agency and responsibility; Agency and causation.

454

Agency in the law

Notes 1 This is what is sometimes called “the But-For” test. The hedge term “typically” appears here primarily because of overdetermination cases. If two assassins fire bullets at the president, either of which would suffice to kill him, and the bullets arrive simultaneously, both are causes-in-fact of the president’s death even though it is neither the case that the president would not have died but for the firing of the first bullet (since the second bullet would have done the trick) nor that he would not have died but for the firing of the second bullet (since the first would have sufficed). 2 People v. Acosta (284 Cal.Rptr. 117). 3 One might also ask whether the deaths came about, thanks to Acosta’s action, in a way that was sufficiently strange as to disqualify the deaths from being things of which he was the agent; that is, one might think that the first requirement for proximate causation, “reasonable foreseeability”, was not met, and so decline to consider the question of comparative agency-of. The court in the case thinks it normal for conduct that leads to police pursuit to end in death, and so takes this condition to be met. It is uncertain, and not easily determinable, whether this was the right conclusion to reach.

Further reading Feinberg, Joel (1970) Doing and Deserving: Essays in the Theory of Responsibility, Princeton: Princeton University Press. Moore, Michael (2009) Causation and Responsibility: An Essay in Law, Morals and Metaphysics, Oxford: Oxford University Press. Morse, Stephen (2007) “The Non-Problem of Freewill in Forensic Psychiatry and Psychology” Behavioral Sciences and the Law 25: 203–220. Pavlakos, George and Rodriguez-Blanco, Veronica (eds.) (2015) Reasons and Intentions in Law and Practical Agency, Cambridge: Cambridge University Press. Yaffe, Gideon (2012) “The Voluntary Act Requirement” in The Routledge Companion to the Philosophy of Law, Andrei Marmor (ed.), New York: Routledge.

455

42 AESTHETIC AGENCY Keren Gorodeisky

1 Introduction Engagements with aesthetic value pervade human life. We choose to wear these shoes because they beautifully match the dress; we travel to Petra on account of its beauty or we recommend Fiona Apple’s Fetch The Bolt Cutters as an excellent album, rich with allusions, poignant lyrics, and sweeping rhythms. Our aesthetic engagements appear to be as significant to our lives as human beings as our theoretical inquiries and practical engagements. But is there anything worthy of the name aesthetic agency? Even though the term ‘agency’ is said in many ways, one widespread assumption behind otherwise different philosophical discussions of agency might aptly be called the “Practical Approach to Agency.” The literature is characterized by this approach in extensively identifying ‘agency’ with the will, and, more specifically, with the capacity for intentional action. On this view, we exercise agency where we exercise voluntary control.1 This identification likely explains why until recently there has been no discussion of aesthetic agency. For traditionally, aesthetics has focused not on action, but on appreciation. But identifying agency with the will is unfortunate since it fails to do justice to the way that human beings are taken to exercise agency over much more than their voluntary attitudes. We standardly attribute beliefs, emotions, desires, and other conative and affective attitudes, including aesthetic appreciation to people’s agency, and rationally criticize them when these attitudes are inapt. The Practical Approach is unfaithful to this ubiquitous phenomenon. Fortunately, we need not abide by the Practical Approach, but can develop and adopt an alternative: the Authority Approach to rational agency. (This entry focuses on human rational agency since the particular aesthetic phenomena at stake are distinctively human phenomena.) This alternative does not only do justice to the widespread practice of rationally assessing, reactively responding to, and holding people responsible for beliefs, emotions, and acts of aesthetic appreciation—a practice to which the Practical Approach is unfaithful—but also preserves the grain of truth in the Practical Approach: the intuition that agency is shaped by rational sensitivity and activity, and that it is subject to rational assessment, criticism, and praise. This is mainly because this alternative approach preserves, explains, and deepens the relation between human agency and reason.

456

DOI: 10.4324/9780429202131-51

Aesthetic agency

2 Aesthetic agency in the literature It is only very recently that the notion of aesthetic agency has been explored and developed in different contexts. For example, John Dick (ms.) argues for what he calls Aesthetic Evaluative Agency. On his view, we exercise agency in aesthetic evaluation insofar as we can choose or decide upon one evaluation among several possible ones. We can and often do exercise the will, he argues, and have direct voluntary control over many (though not all) of our aesthetic evaluations. Thi C. Nguyen (2020) argues that reflection on the aesthetic experiences and values of games sheds light on the nature of practical agency in general, particularly on the unity of practical agency and the temporality and structure of motivation. Putting forward different notions of aesthetic agency, Susan Wolf and Jonathan Gingerich have challenged (independently of each other, and in different ways) what they regard as the unjustified moralized character of discussions of agency, namely, the narrow focus of these discussions on moral agency and moral responsibility. Gingerich (ms.) argues that philosophical work on agency has unjustifiably ignored the significance and ubiquity of actions that are pursued in light of the human drive toward creativity, and actions that express ideals of personal style. The result falls short of a pluralism about agency, which Gingerich deems more faithful to the phenomena. Wolf (2015, 2016) argues that the standard primacy of moral responsibility in philosophical discussions fails to do justice to the centrality and the depth of what she calls aesthetic responsibility, by which she refers primarily to the responsibility that artists bear toward the works they create, responsibility, she argues, that we respond to with reactive attitudes. Sometimes, though, Wolf suggests that aesthetic responsibility and thus aesthetic agency are wider than what artists exercise when they create, encompassing also the responsibility people bear, not for creatively acting but also for being charming, or having humor. Dominic McIver Lopes has done most to put the idea of aesthetic agency on center stage. In spite of the long history of dissociating aesthetic value from action, Lopes argues the best way to explore the value of aesthetic goods is in terms of the aesthetic actions of experts in different aesthetic practices. It’s time, he argues, for aesthetics to go practical. On Lopes’s view, the reasons that aesthetic value gives us (Lopes’s “aesthetic reasons”) are, like other kinds of reasons, reasons to act in ways that would count as an accomplishment in the relevant practice. We have reasons to pursue aesthetic goods, he argues, since, ideally, this pursuit would be an achievement (in a particular practice, given its norms etc.) (e.g., Lopes 2018: 127). Thus, for Lopes, aesthetic normativity is, what he calls, “plain vanilla normativity.” Aesthetic reasons are the same as the reasons we have for acting in any domain: reasons to excel in a particular practice in this domain. This is, on his view, why aesthetic value is a value: because it gives reasons for actions. When we act on such aesthetic reasons, we exercise aesthetic agency. We are aesthetic agents, then, when we act for aesthetic reasons, reasons to act in ways that would, ideally, count as an achievement. All these accounts put pressure either on philosophical aesthetics or on the literature on agency outside of aesthetics (and in some cases on both) in various valuable ways. None of them, though (other than an implicit suggestion from Wolf )2 challenges the Practical Approach: they all take for granted the assumption that agency is a matter of the will. But this assumption (a) is unfaithful to actual practice, and (b) must be challenged were we to develop an account of genuine aesthetic agency. This is because of the primacy of appreciation in aesthetics, where by appreciation is meant a cognitive-affective attitude that involves

457

Keren Gorodeisky

consciousness of the object as meriting appreciation and of itself as merited by the object (Gorodeisky and Marcus 2018; Gorodeisky 2021a). This primacy will here be explained in terms of (i) merit, warrant, success, and (ii) intelligibility.

2.1 Primacy of appreciation: merit, warrant, success The normative support for actions in pursuit of the aesthetic value of an item depends on, and is explained by, the normative support for the appreciation of the item. This does not mean that actions in pursuit of the aesthetic value of an object are warranted only if they support/promote/enhance appreciation. Rather, the argument is that an item that does not warrant aesthetic appreciation has no aesthetic value (since having it just is warranting appreciation). Given that it has no aesthetic value, no action in pursuit of such value is warranted qua pursuing this value. The primacy is transcendental: warrant of appreciation is the condition of the possibility of (un)warranted aesthetic actions. Without the former, the latter could not be (un)warranted as actions in pursuit of aesthetic value (though they may be warranted on other grounds). Consider reasons and justifications first. There are always reasons for appreciating aesthetically valuable objects (even if commonly through actions), but not reasons for acting on behalf of their aesthetic value independently of appreciation. Relatedly, aesthetically valuable objects fundamentally ground reasons for appreciating them, but only derivatively, insofar as they ground reasons for appreciation, they may also ground reasons for various different actions. In other words, actions in promotion of the object’s aesthetic value are justified insofar as, and because, aesthetically appreciating this object is justified. Actions and the normative support for them bear on whether an object is aesthetically valuable only indirectly via appreciation and the normative support for it. If facts about an object did not rationally support appreciation, they would not rationally support acting in ways that promote this value. For example, translating Elena Ferrante’s Neapolitan Novels to English on account of their aesthetic excellence is justified only if, and because, the facts that are taken to justify this action also justify aesthetic appreciation. This dependency of warrant is supported in part by the oddity of statements such as: “There are aesthetic reasons to exhibit Kara Walker’s ‘The Jubilant Martyrs of Obsolescence and Ruin’ on account of its aesthetic value3 but there are no reasons to aesthetically appreciate it.” They are odd plausibly because of the logic of merit and responsiveness in this domain. So consider merit now. Aesthetically valuable objects fundamentally merit appreciation, and only derivatively, insofar as they merit appreciation, they may also merit other actions. This may be grounded in the logic of success in this domain. Perhaps the normative support for actions depends on, and is explained by, the normative support for appreciation because nothing short of appreciation hits the mark of success in response to aesthetically valuable objects: nothing other than appreciation (e.g., mere action devoid of appreciation) counts as being responsive to an object’s aesthetic value just as nothing other than respect (e.g., mere action devoid of respect) counts as being responsive to the value of a person qua person. Consider the ways we respond to those who fail to appreciate what is to be appreciated and to follow our aesthetic recommendation with anything short of appreciation. Merely acting on behalf of an object’s aesthetic value without appreciating it falls short of hitting the mark of success of aesthetic recommendations.4 It is a reason for disappointment. But appreciating an aesthetically valuable object with no separate action on its behalf is no reason for disappointment. For example, if you tell me, following my recommendation, that you read Lucia Berlin’s A Manual for Cleaning Women, and believe me that it is aesthetically excellent, but did not like it, 458

Aesthetic agency

was not even inclined to finish the collection or to recommend it to others—in short if you don’t experience the cognitive-affective attitude that we standardly call appreciation—I’d most likely be disappointed, thinking that I failed to achieve the aim of my aesthetic recommendation. But if you loved the book, without seeking out Berlin’s other books or even seeking to deepen your appreciation through reviews, this is no reason for the same kind of disappointment. Your appreciation of the book marks proper responsiveness to its aesthetic value. The aim of aesthetic recommendations is appreciation, and only derivatively, on the basis of one’s appreciation, other, separate actions. Notice that the view that aesthetic value fundamentally merits aesthetic appreciation is not a denial that it merits other attitudes too, actions particularly. Clearly, aesthetically valuable objects standardly merit many actions in promotion of their value: some merit translation, some merit criticism, some merit editing, some merit showing, some merit currenting, some merit preservation, and so on and so forth. But the aesthetic value of items merits such ways of practically promoting it only insofar as, and because, it merits appreciation. This is manifested by our practices of recommending, conversing about, and critiquing aesthetically valuable items, by the fact that there are always reasons for appreciating aesthetically valuable objects (even if most commonly through actions), but not reasons for acting on their behalf independently of appreciating them, and by the related oddity of certain statements.

2.2 Primacy of appreciation: intelligibility Actions performed on aesthetic objects and in aesthetic practices depend on appreciation, not only in terms of warrant but also in so far as they are not intelligible as aesthetic actions and practices independently of appreciation. Nor can their objects, without great difficulty, be understood as aesthetic independently of appreciation. We cannot assume that curators, editors, and so forth exercise aesthetic agency because they act on aesthetic objects or properties since it is doubtful that we can specify what objects or properties are aesthetic independently of the aesthetic appreciation that these objects fundamentally merit. The ‘aesthetic’ as such, cannot, without great difficulty, be characterized by any set of properties or objects independently of a reference to appreciation. This often goes unnoticed. The standard characterization of the aesthetic is in terms of the properties that are taken to be paradigmatically aesthetic, like beauty, elegance, gracefulness, ugliness, and garishness (e.g., Meskin 2004; Lopes 2018). But this is doomed. First, most of the concepts of these properties can also be used non-aesthetically and the concepts of many other properties can be used aesthetically. Thus, they don’t capture what is aesthetic about aesthetic value, objects, and properties and derivatively about the actions and practices that center on these. In addition, these properties can both enhance and detract from the aesthetic value of any object: a dance piece can be aesthetically bad because it is elegant and a painting could be aesthetically great because it is garish (to mention just a couple). Thus, the so-called paradigmatic aesthetic value properties fail to characterize aesthetic value, and to distinguish positive from negative aesthetic evaluation. Finally, characterizing the aesthetic by this limited set of properties makes it extremely hard to accommodate the great variety of objects of aesthetic value and the great variety of the properties that realize this value, that is, the fact that briskness and sharpness and wit and even boringness and painfulness may contribute to or detract from the overall aesthetic value of an object, that is, could be used aesthetically. But we can characterize the aesthetic as that which merits a particular kind of (dis)appreciation. Accordingly, objects and properties are aesthetic insofar as they merit and fundamentally ground reasons for the cognitive-affective attitude of (dis)appreciation. 459

Keren Gorodeisky

Aesthetic actions and aesthetic practices are those performed on or center around these objects and properties. Of course, we will need to give an account of the distinctive nature of aesthetic appreciation (that does not beg the question) but this has been done (e.g., Kant [1790]2000; Gorodeisky and Marcus 2018; Gorodeisky 2021a). The essential role of appreciation in characterizing anything that may count as aesthetic and the fact that nothing short of it or independent of it counts as aesthetic responsiveness point to the derivative nature of actions and the primacy of appreciation in the aesthetic domain. This seems hard to deny. Even those who argue against the primacy of aesthetic appreciation admit: “appreciation is the key to many … aesthetic practices” (Lopes 2018: 209); “no one would accept the claim that ‘Ping is dainty’ but nobody ever has reason to appreciate him!” (Lopes 2018: 37). Since aesthetic actions depend on appreciation as explained above, we have genuine aesthetic agency only if we can exercise agency in acts of a capacity that is different from the will: the rational-affective capacity for appreciation. Without an account of agency exercised in appreciation, we lack any account of genuine aesthetic agency. At best, we would have accounts of practical agency, exercised upon a certain set of objects, properties, or in certain practices. But we saw that it is implausible that we can get a grip on what makes objects, properties, actions, and practices aesthetic independently of the (dis)appreciation that they merit. This primacy of appreciation requires that we challenge the Practical Approach to aesthetic agency, and search for an Aesthetic Approach: we must understand how we may exercise agency even when aesthetically appreciating. First, though, a few clarifications. First, the primacy of appreciation does not imply that aesthetic value is the value of the appreciation that it affords, as held by proponents of what is known as “empiricism of aesthetic value” (e.g., Shelley 2010). The Aesthetic Approach to aesthetic agency is not committed to such empiricism. Rather, the proposed view denies that what makes objects of aesthetic value aesthetically valuable is their capacity to afford finally valuable experiences. Second, a detailed account of what aesthetic appreciation is, how it differs from other kinds of appreciation, and how it can characterize objects of aesthetic value and aesthetic value properties is clearly needed. But, again, such an account is available. (e.g., Gorodeisky and Marcus 2018; Gorodeisky 2021a). Third, appreciation is primary but not exclusive. The proposed view neither undermines nor discounts the significance of aesthetic actions and practices. Of course, actions are of utmost importance in the aesthetic domain. The aesthetic would be unthinkable and unrecognizable without artists, acts of artistic creation, and the actions of curators, editors, landscapers, and so forth. Rather than undermining these, the emphasis on the agency we exercise in aesthetic appreciation is meant, first, to pave the way for a more adequate approach to rational agency in general, and, second, it alone can make room for a genuine aesthetic agency since, as we saw, actions are “aesthetic” only derivatively insofar as they center around those that merit (dis)appreciation. A genuine form of aesthetic agency is a form of agency exercised in aesthetic appreciation.

3 Is agency practical? Perhaps the identification of agency with the will is partly grounded in the apt thought that agency is required for a kind of responsibility that goes with rational assessment: one can be assessed for an appropriate/inappropriate, merited/unmerited attitude, and is responsible for this attitude if and only if this attitude is an exercise of one’s (rational) agency. Notice 460

Aesthetic agency

that attitudes can be legitimately criticized or praised along the lines just mentioned only if they are sensitive to considerations of what is appropriate, merited, or called for. Perhaps the Practical Approach to agency is based on a slide from these apt considerations, through the thought that intentional actions are shaped by deliberation, to the conclusion that agency is reduced to intentional action. But this is infelicitous: there are human attitudes besides intentional actions that are similarly sensitive and evaluative. The relevant responsibility is not merely causal as the kind of responsibility alluded to when we say that the crushing tree limb was responsible for smashing Ashley’s car. Nor is it meant third-personally to refer to the kind of responsibility that we have for things we own, like a car, or things we can exert external influence over, like a dog under our care.5 Rather, the kind of responsibility that requires rational agency is the one reserved only to human beings 6 as authors:7 responsibility over what we can (dis)avow when subject to rational reflection or criticism. Non-human animals clearly act, and perhaps can have the other attitudes for which we take human beings to be responsible. But we do not praise or criticize them for these attitudes; we do not take them to authorize these attitudes in a way that allows for (dis)avowal; and so, we don’t hold them responsible for their actions and attitudes in the same way that we hold people responsible. But we do, pervasively, assess people, respond to them reactively, and hold them responsible for attitudes that we take to be exercises of their agency. This sense of agency likely gets to the core of our human nature. But why do we think that it is reduced to the capacity for intentional action? Identifying agency with voluntary control excludes an important dimension of the relevant kind of assessment and responsibility. After all, we assess and criticize the man who believes that homosexuality is a sin. We take him to be responsible for his false belief and for changing it no less than people are responsible for their actions. We can legitimately demand that he disavow this belief and reconsider what to believe. We view Michael as responsible for feeling contempt against all people of color. Rational criticism, blame, moral outrage, and a demand to disavow these attitudes are as appropriate in these cases as they are in cases of intentional actions. And the relevant assessment is not necessarily moral. Even if we don’t necessarily blame people for, say, an inapt grief (over the loss of the person’s favorite soccer team), inappropriate anger (say with one’s mom), or an unmerited aesthetic appreciation (say, disappreciating Knausgård’s My Struggle or appreciating Downtown Abbie), we do assess, respond reactively, and often criticize people for having a variety of inapt, unwarranted, or unmerited, though amoral, emotions, feelings, desires, aesthetic appreciations, and perhaps other affective and conative attitudes, and demand that they disavow those attitudes by reconsidering what to feel, desire, or appreciate. We take these (amoral) attitudes that are not under our voluntary control to be expressions of our rational agency: we act, interact, speak, and react pace the Practical Approach to agency. I focus here on emotions since aesthetic appreciation is affective: a cognitive kind of feeling that presents its object as meriting appreciation and involves an awareness of itself as merited by its object (Gorodeisky and Marcus 2018; Gorodeisky 2021a). But an increasing number of philosophers convincingly argue that other attitudes that are not under voluntary control, like beliefs, desires, and even our noticing and neglecting, are exercises of rational agency, for which we are responsible (Moran 2001; Smith 2005; Tenenbaum 2007; Brewer 2009; Boyle 2010; Boyle and Lavin 2010; Yao 2019, and Marcus in this volume). On the face of it, identifying agency with voluntary control is unwarranted. What is, then, the sense of agency that allows only people to be (i) subject to rational assessment and reactive attitudes, and (ii) responsible, not merely causally, for various attitudes? 461

Keren Gorodeisky

4 The authority approach to agency My answer draws on the work of Stuart Hampshire, Charles Taylor, Richard Moran, and Matthew Boyle. This is the “Authority Approach” to agency. On this view, an attitude is an exercise of rational agency, not insofar as it is formed ‘at will,’ but insofar as it is constituted by consciousness of its own appropriateness. Thus, beliefs and many of our emotions and desires are active exercises of agency insofar as they are constituted by the awareness that they are called for, merited, or appropriate responses. We ordinately talk and act on the assumption, which we have built into the normal forms of speech, that many of our emotions and attitudes, desires and interests, were formed, and can be altered by our thinking about the appropriateness or inappropriateness of their objects. (Hampshire, 1965: 94; Cf. Taylor, 1985) These attitudes that are exercises of rational agency are not merely related to one’s consciousness of the world and what is called for, but they are so related because they are constituted by this consciousness. In Hampshire’s words, my consciousness that “the object is dangerous, or of its saddening features, constitutes my fear of it, or my sadness about it” (Hampshire 1965: 95). Hampshire and Moran often speak as if what constitutes those agential attitudes is a belief about the relevant value (e.g., Moran 2001: 54), and even suggest that such a belief is a component of emotions and desires (e.g., Hampshire 1965: 95, 97). But there is no reason to postulate a separate belief as either external or internal to rationally agential attitudes. Without either being (affective or conative) beliefs or including beliefs as components, emotions, desires, and acts of aesthetic appreciation can themselves be direct modes of experiencing the world and what it merits: they are affective and conative presentations of the world as meriting themselves.8 On this view, belief is an exercise of rational agency since to believe is to take the corresponding proposition to-be-believed (that is, true). An action is an exercise of rational agency because to act is to take the action to-be-done (that is, good in some way). Paradigmatic emotions and acts of aesthetic appreciation are exercises of rational agency because to have them is to take their objects as to-be felt or to-be-appreciated in a particular way, a way that is merited by the relevant object’s value. These attitudes could be (dis)avowed and be subject of rational assessment because they embody the subject’s rational assessment: the subject’s take on what is to be believed, felt, desired, or appreciated. If on reflection this take turns out to be mistaken, the subject is under rational pressure to change it. And changing one’s take on what to believe, feel, desire, or appreciate is changing the belief, feeling, desire, or appreciation itself. Because exercises of human agency are constituted by this awareness, they are always legitimately vulnerable to a certain “why” question, a question that does not ask what causally or psychologically led one to have the attitude, but whether it is appropriate. In contrast to sensations, agential attitudes are sensitive to a question about their normative grounds. While the actuality of sensations is exhausted by how subjects are affected by an object, the actuality of agential attitudes depends on their being taken to be appropriate. This dependence of an attitude on its self-awareness as the proper attitude to have is a central mark of rational agency (cf. Boyle 2010). Thus, this Authority Approach of agency can preserve the intuition that agency has something to do with “activity” while denying the misleading idea that it is identical to voluntary control. 462

Aesthetic agency

This is not to deny that emotions, desires, and acts of aesthetic appreciation seem receptive in a way that actions and even beliefs are not. Surely, we cannot change our emotions, desires, and aesthetic appreciations at will. It is also true that those who acknowledge, say, that their anger or jealousy are unmerited but are unable to change their emotion often, though not always, remain intelligible to others in a way that those who acknowledge that their belief is false but are unable to change it usually are not. There seems to be place for hope that they will feel appropriately in the former case, but not in the latter case. The Authority Approach does not deny this receptivity of emotions, desires, and acts of aesthetic appreciation. What it denies instead is that this receptivity excludes them from being exercises of rational agency. On the Authority Approach, the kind of receptivity that characterizes our emotions, desires, and aesthetic appreciations is compatible with rational agency: feeling, desiring, and aesthetically appreciating are the ways through which we, human beings, exercise our rational agency. This alternative account of agency helps us recognize that we may be aesthetic agents not merely as having other action-producing roles (e.g., creators, editors), but even as aesthetic appreciators. Our aesthetic appreciations are constituted by an awareness of their objects as meriting appreciation, and of themselves as responsive to reasons for appreciation (however inchoately and without articulation at first).9 Aesthetic appreciation is a stance on the question: “what is to be appreciated?” just as beliefs are a stance on the question: “what is to be believed?”, actions and intentions are a stance on the question: “what is to be done?”, and emotions are a stance on the question: “what is to be felt?” It is for this reason that aesthetic appreciation is an exercise of human rational agency, albeit it differs in form—not only in content—from the theoretical and practical exercises of human agency.

5 Aesthetic agency: the authority approach That agency is irreducible to intentional action is good news for aesthetics. For actions are intelligible as aesthetic actions and make sense as human actions in general only insofar as they are performed on objects and in practices whose values merit appreciation. But is aesthetic appreciation reasons-responsive and active in the way required for rational agency on the Authority Approach? One might argue that even if the Authority Approach is to be preferred to the Practical one, this is insufficient for showing that aesthetic appreciation is an exercise of rational agency. One may worry that aesthetic appreciation is not in the space of reasons. But this is mistaken. Acts of aesthetic appreciation are capable of reasons-responsiveness and are thus exercises of rational agency insofar as they are (a) assessable as either supported or unsupported by good reasons, (b) subject to an explanatory/normative ‘why’ question, and (c) criticizable for being irresponsive to reasons. Like other exercises of rational agency, acts of aesthetic appreciation are standardly assessed as either supported or unsupported by reasons. Should I enjoy Fleabag—does it merit appreciative aesthetic enjoyment? Doesn’t Bela Bartok’s Concerto for Orchestra merit appreciation no matter how unenjoyable I find it? Arguably, we criticize acts of aesthetic appreciation for lacking reasons as often as we criticize unjustified actions and beliefs. Relatedly, the ‘why’ question that asks for an (inseparable) explanation of both what led one to appreciate the object and what renders this appreciation appropriate is one of the commonest responses to the expression of aesthetic appreciations. Hearing that a friend dislikes the new Spike Lee, or that another friend marvels 463

Keren Gorodeisky

at Rachel Cusk’s latest novel Second Place, we standardly ask for reasons: Why do you appreciate this work but dislike this one? We also form reactive attitudes toward our (dis)appreciating interlocutors: we’re disappointed by a good friend who dislikes a poet we love, or admire a critic who constantly helps us appreciate excellent works. And notice the difference between, on the one hand, appreciation and, on the other hand, sensation. When hearing the expression of appreciation, we standardly ask the same kind of ‘why’ question that we ask in response to avowals of the agential attitudes mentioned above, but not in response to an avowal of sensation; we ask about the normative reasons for them. This is never true of sensations, which are opaque rather than transparent to the world and to whether the world is such as to give reasons for a certain response. It makes no sense to ask a person why she feels a headache, where this why question concerns the normative grounds of the sensation, whether to feel it. Even though they are not producible at will, acts of appreciation are reasons-responsive and active: unlike sensations, appreciations are beholden to considerations about whether and what to appreciate, and about their appropriateness. They are not episodes we undergo, but expressions of our reasons-responsive capacity. Acts of appreciation are thus required to be well-supported by reasons, and so they are surely capable of reasons responsiveness—should implies can. When the explanation fails to show them to be appropriate, we criticize the subject, expect her to acknowledge the reasons against her appreciation, and to reconsider what to appreciate. If this is what is required in order for an attitude to be an exercise of rational agency, as I argued above, then aesthetic appreciation is an exercise of rational agency.

Conclusion I’ve argued that genuine aesthetic agency must be agency exercised in the very appreciation of objects. The Practical Approach to Agency that reduces agency to willing forecloses the possibility of such appreciative aesthetic agency. Luckily, we need not endorse the Practical Approach. There are very good reasons to reject it and replace it by the Authority Approach. On this latter approach, an attitude is an exercise of rational agency insofar as it is constituted by an awareness of what is called for or appropriate or merited. And so, not only actions, but also beliefs, emotions, desires, and acts of aesthetic appreciation are exercises of rational agency. This allows us to carve out space for a genuinely Aesthetic rather than Practical Approach to aesthetic agency.

Related topics Agency and emotions; Rational agency; Agency and responsibility; Agency and games; Agency and normativity.

Notes 1 E.g., “In very general terms, an agent is a being with the capacity to act, and ‘agency’ denotes the exercise or manifestation of this capacity” (Markus Schlosser, Stanford Encyclopedia of Philosophy, 2019). 2 Although Wolf does not explicitly consider the view I defend below that we exercise agency in aesthetic appreciation, one of her passages acknowledges the importance of regarding our very susceptibility to beauty as part of our responsible agency: “But if moral responsibility is not a part

464

Aesthetic agency

3 4 5 6 7 8 9

of some larger or more general feature of human agency, it will not be or relevant to our capacity for humor or creativity or to our susceptibility to nature or to beauty” (2015: 141). Reasons for acting not on behalf of, but regardless of, an object’s aesthetic value need not depend on appreciation. Such lack of appreciation also falls short of the mark of success in response to criticism, see Gorodeisky (2021b). Cf. Hampshire (1965), Moran (2001). Cf. Taylor (1985), Wolf (2016), Smith (2005). Cf. Moran (2001). On desires as presentations of goodness, see Brewer (2009), Tenenbaum (2007), Boyle and Lavin (2010). On the emotions as presentations of value (independently of having a belief as a component), see McDowell (1998), Milona (2016), Tappolet (2016), and Gorodeisky (2021a). The claim is not that every act of aesthetic appreciation is responsive to reasons. I might appreciate a shade of violet color for no other reason than its particular color, which as such merits appreciation. This in no way undermines the rationality of aesthetic appreciation just as the fact that some actions are performed for no reason does not undermine the rationality of action. In both cases, the relevant rationality is not a matter of a positive answer to the rational question “why” but of the applicability of this question (cf. Anscombe 2000: 25). Moreover, even though I appreciate this shade of violet for no other reason than this very color, my appreciation is constituted by my sensitivity to this shade of color as meriting appreciation.

Further reading Boyle, Matthew 2009. Active belief. Canadian Journal of Philosophy 39 (S1): 119–147. Argues that even though we cannot believe “at will,” our beliefs are exercises of rational agency insofar as they are constituted by making up our minds about what to believe. Hampshire, Stuart 1965. Freedom of the Individual. Princeton: Princeton University Press. Argues that there are two kinds of self-knowledge: knowing one’s mental states by observing them and knowing one’s mental states by making up one’s mind about their appropriateness, and locates the freedom of the individual in the latter. We often come to know our beliefs, intentions, desires, and emotions by asking the normative question concerning their appropriateness since these mental states are constituted by evaluating them to be appropriate to their objects. Lopes, Dominic McIver 2018. Being for Beauty: Aesthetic Agency and Value. Oxford: Oxford University Press. Argues that aesthetic value plays an important role in the good life. It is a value insofar as it provides agents with practical reasons to act in ways that would ideally amount to achievements in the relevant practices. Nguyen, C. Thi 2020. Games: Agency as Art. Oxford: Oxford University Press. Argues that games belong to a distinctive medium of art, the art form of agency, insofar as they offer players aesthetically (and practically) valuable experiences of adopting variety of different agencies, and of engaging in different motivational structures. Smith, Angela 2005. Responsibility for attitudes: Activity and passivity in mental life. Ethics 115 (2): 236–271. Argues that we are (morally) responsible for our attitudes not insofar as they are under our voluntary control, but insofar as they reflect our evaluative judgments.

References Anscombe, G.E.M. (2000) Intention. Cambridge, MA: Harvard University Press. Boyle, Matthew 2009. Active belief. Canadian Journal of Philosophy 39 (S1): 119–147. Boyle, Matthew and Doug Lavin (2010) ‘Goodness and Desire’ in Tenenbaum, S. (ed.) Desire, Practical Reason, and the Good. Oxford: Oxford University Press: 161–202. Brewer, Talbot (2009) The Retrieval of Ethics. Oxford: Oxford University Press. Dyck, John (ms.) “Choosing Beauty.” Gingerich, Jonathan (ms.) “Aesthetic Agency.” Gorodeisky, Keren (2021a) “The Authority of Pleasure,” Nôus 55(1): 199-220.

465

Keren Gorodeisky ——— (2021b) “Must Reasons be Either Theoretical or Practical? Aesthetic Criticism and Appreciative Reasons,” Australasian Journal of Philosophy. Gorodeisky, Keren and Eric Marcus (2018) “Aesthetic Rationality,” Journal of Philosophy, 115(3): 113–140. Hampshire, Stuart (1965) Freedom of the Individual. Princeton: Princeton University Press. Kant, Immanuel [1790] (2000) Critique of the Power of Judgment, ed. and trans. Paul Guyer and Eric Matthews. Cambridge: Cambridge University Press. Lopes, Dominic McIver (2018) Being for Beauty: Aesthetic Agency and Value. Oxford: Oxford University Press. McDowell, John (1998) “Aesthetic Value, Objectivity and the Fabric of the World,” in Mind, Value and Reality, Cambridge, MA: Harvard University Press: 112–130. Meskin, Aaron (2004) “Aesthetic Testimony: What Can We Learn from Others about Beauty and Art?” Philosophy and Phenomenological Research, LXIX(1): 65–91. Milona, Michael (2016) “Taking the Perceptual Analogy Seriously,” Ethical Theory and Moral Practice 19(4): 897–915. Moran, Richard (2001) Authority and Estrangement: An Essay on Self-Knowledge. Princeton: Princeton University Press. Nguyen, C. Thi (2020) Games: Agency as Art. Oxford: Oxford University Press. Schlosser, Markus (2019) Agency, in The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta, URL = . Shelley, James (2010) “Against Value Empiricism in Aesthetics,” Australasian Journal of Philosophy 88(4):707–720. Smith, Angela (2005) “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics 115(2): 236–271. Tappolet, Christine (2016) Emotions, Values, and Agency. Oxford: Oxford University Press. Taylor, Charles (1985) Philosophical Papers: Volume 1, Human Agency and Language. Cambridge: Cambridge University Press. Tenenbaum, Sergio (2007) Appearances of the Good: An Essay on the Nature of Practical Reason. Cambridge: Cambridge University Press. Yao, Vida (2019) “The Undesirable & the Adesirable,” Philosophy and Phenomenological Research 99(1): 115–130. Wolf, Susan (2015) “Responsibility, Moral and Otherwise,” Inquiry: An Interdisciplinary Journal of Philosophy 58(2): 127–142. ——— (2016) “Aesthetic Responsibility,” The Amherst Lecture in Philosophy 11: 1–25. .

466

INDEX

Note: Bold page numbers refer to tables, italic page numbers refer to figures, and page numbers followed by “n” refer to endnotes. ability: and disability 128; expert agency 126–127; and mistakes 127–128; pathologies of agency 128–129; powers and skills 125–126 access consciousness 201, 205 achievement play 26, 81–82 acquaintance 30, 261n4 action 203–204, 403; causal theory of 101–103; commanded 272; explanation of 130–132; goodness of 443–446; standard story of 22, 24, 47–48, 131; structure for 202–203; theory, place of will in 277 active thoughts and attitudes 254–255 adequacy criteria 412–413; attributability criterion 413; evaluation criterion 413; mentality criterion 412–413; practicality criterion 413 aesthetic agency 401–402; agency practical 460–461; authority approach 462–464; intelligibility 459–460; in literature 457–460; merit, warrant, success 458–459; primacy of appreciation 458–460 Aesthetic Evaluative Agency 457 agency: automaticity and control 203; and autonomy 266–267, 298–305; and body 180–181; as capacity 3; as causal initiation 21, 28; as causal process 21, 28–29; and causation 20–21, 27–33; as creation 9; defined 2, 3, 6, 15, 19, 21, 37, 73, 159; as determination 25; and disability 128, 160–165; and embodiment 5–6; and emotions 268–269, 317–325; and evidence 223–224, 244–250; and games 25–26, 77–84; as goal-directedness or purposiveness 395; and identification 266, 288–296; and intentionality 6; and mistakes

127–128, 149–156; and morality 400, 440–446; narrative and mortality 334–335, 385–391; and normativity 398–399, 421–428; as outcome 21, 28; and personal identity 333–334, 376–383; and philosophy 1–16, 2–3; powers and skills 125–126; and practical knowledge 222–223, 234–242; and practical reasoning 397–398, 412–418; as productive power 4–5; as psychological causality 11–13; as reason responsiveness 14–15; and reasons 396–397, 403–410; and responsibility 265–266, 279–285; as self-constitution 10–11; and self-knowledge 224–225, 253–260; and time 330–335, 336–391; time and rationality 331–332, 357–364; varieties of 6–7; and (the limits of ) volitional conflict 267–268, 307–315; and will 270–271; will and freedom 264–265, 270–277; see also separate entries agency in the law 401; agent-of 449–450; bodily motions, or omissions 448; criminal domain 448; proximate causation 450–454 agency of non-persons (wantons) 288 agent-causalists 28 agent-causal strategy 32–33 agent, defined 448 agentialism 225, 254, 256, 258 agential models of self-knowledge 256–260 agential powers: agent’s practical reasoning 131; dispositions 134; and explanation of action 130–132; one-way power 134; specification of 132–133; two-way powers 134–135 agentive guidance, notion of 208 agent-of 449–450 agent’s maxim 442

467

Index the agent’s real reason 410 Aguilar, Jesus H. 20, 21 aim of agency 399–400; and aim of belief 431; content of normative standards 434–435; explanatory ambitions of aim theories 431–432; formal vs. substantial aim theories 433–434; teleological vs. efficient aim theories 432–433; theorists explaining normativity 435–437 aim of belief 431 aim theories: explanatory ambitions of 431–432; Presupposition Objection 437; Shmagency objection 435, 436; Why Bother Objection 435–437 alexithymia 161–162 Alien hand syndrome 169 Allen, W. 72 Altshuler, R. 334, 335 Alvarez, M. 396, 397 Amaya, S. 127, 128 Andreou, C. 332 animal agency 88–89; animal kingdom 103–105; causal theory of action 101–103; moral responsibility 106–107; rationality 105–106 Anscombean views 240–241 Anscombe, G.E.M. 23, 31, 49, 121, 122, 154, 170, 193, 222, 223, 235–238, 240–242, 248, 348, 369 apathy 173–174 apotemnophilia 198 Aristotle 93, 106, 121, 135, 228, 270, 272, 442, 443, 445 Armstrong, D. 143, 236 Arpaly, N. 292, 293, 413 artificial and machine agency 333, 366–374; attitudes for simple planning agent 368; deliberate reasoning 367; interactions between data sources and reasoning services 370; IRMA architecture and its descendants 368–371; matching actions to goals 367–368; reasoned desires 371–373 artificial intelligence (AI) research 333 art of games 83–84 atelic action sentences 55 atelic ends 345n1 attention 181–182; and action 203–204; and blindness 204–206; characterization of 203–204; Many-Many Problem 204 attitudes: active thoughts and 254–255; agent’s direction 353; based explanations 154–155; committal attitudes 254; content-bearing attitude 274; existing attitudes 254; objective attitude 280; reactive attitudes 280; for simple planning agent 368 attributability criterion 413 autism 174 auto-activation deficit (AAD) syndrome 129, 173–174

autonomous agent 92 autonomy and agency 266–267, 289; agents’ actions 299; answerability-centered views 303; appropriate receptivity 304–305; authenticity-based accounts 300–301; autonomous action 298; forms of interference 299; higher-order desire 299–300; identificationist approach 298, 300; self-governance 302; self-governing policies 300, 301; take or accept responsibility 302–303 background framework 423 Bagnoli, C. 268, 269 Barandiaran, X. E. 92 Barnes, Elizabeth 160 Bartok, Bela 463 basic actions 135 BDI (belief–desire–intention) architecture 333, 368 belief/desire psychology 12 Bereiter, Carl 145 Berger, Ronald J. 160 Berlin, Lucia 458, 459 Bilgrami, Akeel 258, 260 blameworthy 281 blaming 281 blindness 204–206 body and agency 180–181, 183; embodied agency 195–198; as enabling condition for agency 192; as structuring agency 193–194 body identity integrity disorder 198 body ownership, sense of: first-order forms 197; illusory sense of 197; loss of sense 197; ownership constraint 197–198 Boghossian, P. 120 BOID architecture 373 bounded agency 25, 68–74; agential architecture 71–73; feedback-driven 71; in-principlecorrect way 70; Nietzschean version of agent 69; polymorphic agent 75n15; self-governing system 68–69 bounded rationality 25, 70 Boyle, M. 106, 228, 462 Brandom, R. 105 Bratman, M. E. 71, 153–155, 211, 274, 291–293, 331, 337, 352, 362–364, 368–370, 418 Brewer, T. 341 Broersen, Jan N. 373 Broome, J. 353, 397, 398, 414–417 Broome’s account 398, 414–415 Brunero, J. 417 Buckareff, Andrei A. 20, 21 Bukoski, M. 435 Burge, T. 91, 256, 257, 260 Buss, S. 267, 268 the But-For test 455n1

468

Index Calvo, Rafael A. 373 Campbell, L. 240 Carlson, Licia 160 causal initiation, agency as 21, 28 causal process, agency as 21, 28–29 causal reduction and deviant causal chains 38–42 causal theory of action (CTA): bodily movements 101–102; extension of agency 103; mental states 102 causation in agency: events as causes 32; as mere difference-making 27, 29–30; powersin-process as causes 33; as production or generation 31; proximate causation 450–454; substances as causes 32–33 Chakravartty, Anjan 32 Charles, D. 50–53 Clark, A. 206 Clarke, R. 23, 24, 135 Clayton, P. 92 cognitivist 239 Cohen, Philip R. 369 Coliva, Annalisa 258, 260 commanded actions 272 compulsive – behaviour 111 Comrie, B. 52 conscious attention 143, 146n3, 181, 188, 195 consciousness 181–182, 201; access 201; behavioural possibilities 202, 202; ManyMany Problem 202; non-attentional visual awareness 208–209; phenomenal 201; purely reflexive behaviour 202; structure for action 202–203; unconscious guidance (see zombie actions) consequentialists 440, 441 constitutive standards 150; and goodness of action 443–446 constitutivism 396, 427 constitutivist accounts 400 constrained maximization 332 content of normative standards 434–435 continental tradition 17 continuous and atelic activity 341 control vs. irresistibility 172 creation, agency as 9, 12, 14 Crowther, T. 53, 54 Csikszentmihalyi, M. 126, 127, 140, 144 Cusk, Rachel 464

death, concept of 390; see also mortality decision-theoretic approach 245–246 decision theory 245–246 deliberate practice 141–142 deliberative externalism 423–424 Dennett, D. C. 92, 104 deontological account 442 deontologists 440–441 deviant causal chains, problem of 12, 107n1 deviant causation 131 diachronic agency 330; continuity vs. unity 339–340; dynamics of practical standpoint 342–343; kinds of temporal agents 338; planning capacity 337; plan-states 337; structure of extended actions and activities 339–341; telic vs. atelic ends 340–341; temporal agents 336–338; temporal goods and values 341–342; temporal integration of actions, activities, and the agent 345 dialectical activities 341 difference-making theories 21 disability and agency 128; alexithymia 161–162; capacities and abilities 159; emotional blunting 161–162; exercise of self-control 163–164; impairment and disability 160–161; intellectual disability 162–163; nature and approach to disability 160–161; Parkinson’s disease 161; physical disabilities 166n2; socially and ecologically dependent 164; social models of disability 160; successful human agency 164–165 disappearing agent problem 13, 113 discounting: exponential discount functions 360, 360; future utility 357–358; hyperbolic discount functions 359, 359–360; partiality toward oneself 359; possibility of altruism 358 disembodied agent 185 dispositions 134 distinctive form of agency 119 Doherty, Patrick 368 Dowe, Phil 63 doxastic views 239–240 Dreier, J. 432, 434 Dretske, F. I. 92 Dreyfus H. L. 141 Dreyfus, S. E. 141 Dworkin, Gerald 300, 302

Dancy, J. 246, 398, 416, 417 Danish String Quartet (DSQ) 142 Darwin’s theory of evolution 42 Davidson, D. 38–41, 43, 44, 48–50, 56, 60, 120, 151, 152, 179, 235, 236, 240, 241, 273, 275, 348, 349, 407–409, 417 deafferentation: and automaticity of everyday agency 193; and non-observational knowledge of action 193–194

Ebels-Duggan, Kyla 303, 304 effortful expertise 140 elicited agency 272 Ellis, J. 175 Elster, Jon 324 embodied agency: at will character 5; body as effector 195–196; mental agency, issue of 5–6; proprioception as condition on 196–197; role of body 5; sense of body ownership 197–198

469

Index emotional blunting 161–162 emotions and agency: as action-oriented 319–320; cooperative schemes 323–324; emotional engagement 321–322; emotional resonance 323–324; and moral agency 323; as motivational packages 318–319; as normative prompts 320–321; paradigmatic scenarios 319–320; perceptualist model 319; problem of emotional occlusion 322; rational and moral appropriateness 317; in rational deliberation 321–322; recalcitrance of 322; shared actions 323–324; as springs of action 318–321; transmutation and temporal agency 324; understanding 318 empiricism of aesthetic value 460 empiricist views: non-inferential 236; non-observational 236; observational knowledge 237 Enoch, D. 426, 427, 436 episodic refraining 64 epistemic agency 222; basic and non-basic action matters 229; belief properties sortals or attributes 227; bodily actions and activities 228; epistemic activity and indirect epistemic control 229–230; ethics of belief 230–231; making up one’s mind and direct epistemic control 228–229; ontology of believing 227–228; overview 226–227 Ericsson, K. A. 141, 142 error cases 397 ethical responsibility 7 etiological theory of function 94 evaluation criterion 413 events and processes: progression and completion 47–50; realization 55–56; two entities 22, 52–55; two properties 22, 50–52 events as causes 32 everyday expertise 140–142 evidence and agency 223–224; decision theory 245–246; practical knowledge 248–249; practical reasoning 246–248; problem 244–245 evolutionary etiology 94–95 exercises of agency: agency and embodiment 5–6; agency and intentionality 6; agency as a productive power 4–5; agency as capacity 3; goal-directedness 4; telic and atelic goals 5 experience 234 experience sampling method 144 experiential present 54 expert agency 126–127; Act without acting 144–145; agency in flow 143–144; everyday expertise 140–142; final effort 145; hypothesis of effortful expertise 140; post-performance amnesia 142–143; professional-level expertise 140–142 expert skill, defined 139

expert system 366 explanatory role approach 92–93 Explicability of the Purposeful Behaviour of an Intentional Agent (EPIA) 110 Ferrero, L. 246, 330, 436 Fetch The Bolt Cutters (Apple) 456 first-person authority 255–256; agential authority 256; epistemic authority 255–256 First-Person Perspective (FPP) 115–116 Fischer, J.M. 289, 294, 389–391 fitness estimation approach 88; behavioral dispositions 96–98; evolutionary fitness 97; goal-directedness 97; probability distribution 96 Fodor, J. 42 Ford, Henry 213 formal vs. substantial aim theories 433–434 Frankfurt, H. 28, 102, 103, 105, 152, 211, 268, 290, 291, 293, 299, 300, 302, 308, 309, 310, 323, 376, 377, 378, 386 Freud, S. 154, 155, 212 Frey, C. 240 Frey, J. 240 frontotemporal dementia (FTD) 162 Frost, K. 223, 240 Fulda, F. C. 93 full-blooded agency: distinctive social form 8; epistemic position 7; ethical responsibility 7; freedom and free will 8; intentional conduct of fellow human beings 7; normative or deontic statuses 8; structural complexity 7 function and teleology: belief and desire 38, 41; causal reduction and deviant causal chains 38–42; description of behaviour 42–43; mental states 38–39; notion of guidance 39–40; proportionate cause 40–41; selectional explanations and teleology 37, 42–44 Galton, A. 53, 54 games and agency: art of games 83–84; on games and roles 77–80; goal of game-playing 79–80; in Homo Ludens 77–78; kind of agency-switching 79; layered agencies 82–83; magic circle approaches 78, 84n1, 84n2; striving and achievement play 81–82; suits on nature of 80–81 Gauthier, D. 152 generic agency 424–426 genuine agency 87 Gertler, Brie 224, 225 Gibbard, A. 354 Ginet, C. 215, 237 Gingerich, Jonathan 457 goal-directedness 4 goal-direction: content-bearing attitude 274; desires and emotions 274–275; Hobbesian

470

Index model of agency 275–276; and intentionality 272–273; and motivating power 273–276; self-determination or freedom 274 Goldie, P. 387 Goodale, M. A. 206, 208 good reasoning 417–418 goods of mere continuity 342 Gorodeisky, Keren 401, 402 guidance, notion of 39, 204 Guthrie, E. R. 139

event-causalists 113–114; explanation 109, 110; flexible behaviour 111; FPP (First-Person Perspective) 115–116; independence 112; intentional states 109, 110; kind of intellectualism 110; preliminary profile 109–110; purposefulness 109, 110; rational revision 111–112; unity in category 114–116 intentional agent 109, 110–111 intentional states 109–110 internalism 405 International Classification of Impairment, Disability, and Handicap, WHO’s 161 intrinsic power 93–94 inverse akrasia 292 Ismael, J. T. 68

Haase, M. 22, 23 Haffenden, A. 206 Hampshire, S. 462 Haraway, Donna 213 Haslanger, Sally 423 heteronomy 445 Hieronymi, P 3, 173, 228, 265, 266 historical goods 342 Hobbes, T. 272–277 Hobbs, Jerry R. 368 Holton, R. 152, 171, 332, 338, 361, 362 Hornsby, J. 49, 53, 54, 172 Horty, John F. 333, 371 Huizinga, Johan 77–79 human agency: practical organization 348, 349; social organization 348 Hume, D. 29, 405 Hunter, D. 222 hyperbolic discounting 332 hypothetical imperative 445

Jackson, S. A. 144 James, W. 203, 204 Jamison, K. R. 175 Jaworska, A. 266, 291, 300 justifying reasons 403

identification and agency: participantidentification 293–295; stance-identification 290–293; two concepts of identification 289–290 imperfection, facing 155–156 imperfective paradox 22, 48 individual planning agency: and diachronic organization 349–351; future-directed intentions 349; Lockean cross-temporal ties 350; ordinary desires and beliefs 349–350; planning psychology 351; self-governance 350 inference 238 inferentialist views 238–239; inference and evidential support 238 initial contrast 253 instantiation 56 intellectual disabilities 162–163, 174 intelligent agent 366 Intelligent Resource-bounded Machine Architecture (IRMA) 333; beliefs, desires, and intentions 370; and descendants 368–371 intentional action, concept of 48 intentional agency 6, 89–90, 109–116; agent it/ her/himself 112–114; designer-given desire-like state 112; desire 112; disappearing agent 113;

Kant, Immanuel 250, 400, 425, 426, 441–445 Katsafanas, P. 425, 431, 436, 437 Kauffman, S. 92 Kavka, G. S. 363 Kennett, J. 163, 174 Kim, J. 31 kinds of agency: animal agency 88–89, 101–107; full-blooded agency 7–8; intentional agency 89–90, 109–116; minimal agency 88, 91–99; rational agency 90, 118–123; varieties of agency 6–7 knowledge and agency: agency and evidence 223–224, 244–250; agency and practical knowledge 222–223, 234–242; agency and self-knowledge 224–225, 253–260; epistemic agency 222, 226–231 Kolodny, N. 418, 436 Korsgaard, C.M. 72, 211, 276, 385, 400, 418, 425, 426, 434–436 Kvarnstrom, James 368 Laird, John E. 366 layered agencies 82–83 Lee, Spike 463 Levesque, Hector J. 369 Levy, R. 173 Levy, Y. 184 Lewis, D. 29, 30, 378 Lewthwaite, R. 139 Lindeman, K. 399, 437 Lopes, Dominic McIver 457 Lumsden, D. 388, 391 MacIntyre, A.C. 164, 388 Mackie, J. L. 29

471

Index Maier, J. 132 A Manual for Cleaning Women (Berlin) 458–459 Marcus, E. 90 Marušić, Berislav 223, 224 matching actions to goals 367–368 material agency 182, 211–218; account’s plausibility 216; act-types 216; agential capacity 211–212, 217; bodily movement, or movement at all 213; Deleuzian assemblage 213; functional body 212; general and specific forms 213–214; individuation of capacities 216–217; lacks of capacity 215; neurological deficit 215; temporal character 214; theory of human agency 217, 218; versions of knowhow 215–216 McDonnell, N. 40–42 McDowell, J. 122, 313 McGeer, Victoria 259, 260 Meditations (Descartes’) 185 Mele, A. R. 39, 152, 156, 171 mental actions 184 mental agency 180; anti-Cartesian views of mind 186; and bodily agency 183; bodily and mental paralysis 188–189; brain imaging technology 184; causal pedigree view 188; disembodied agent 185; epistemological considerations 184; mental actions 184; overt actions 184–187; practical self-awareness 187–188 mentality criterion 412–413 mental retardation 166n4 mere difference-making: causation in agency 27, 29–30; non-productive theories of causation 29; Ramsey Lewis view 30 mere happenings vs. doings 1–2 Metaphysical Disputations (Suarez) 276 metaphysics of agency: agency and causation 20–21, 27–33; agency and games 25–26, 77–84; agency, events, and processes 22–23, 47–56; agency, functions, and teleology 21–22, 37–44; bounded agency 24–25, 68–74; negative agency 23–24, 59–64 mid-life crisis 330, 343 Miller, George A. 367 Millgram, E. 20, 24, 25 Millican, P. 139 Mill, J.S. 441, 443, 445 Milner, A. D. 206, 208 minimal agency 88, 91–99; authorship and behavioural freedom 91; autonomous agent 92; beyond minimal agency 98–99; goaldirectedness 91–92; individuality 91; theories of (see theories of minimal agency) The Minority Body (Barnes) 160 mistakes and agency 127–128; attitude-based explanations 152–155; causal deviance 151; constitutive standards 150–151; distinctively

human agency 150; facing imperfection 155–156; idealization 151–152; intentional actions 150; performance mistakes 153–154; planning agency 150; psycho-physiological approach 156n4; standards of agency 149–150 Montero, B. G. 126, 127 morality and agency 400; action and structure 440–443; agent’s maxim 442; consequentialist 440, 441, 443; constitutive standards and goodness of action 443–446; deontological account 440–442; Kantian maxim 442; Nicomachean Ethics 442; virtue theorists 440, 443 moral psychology and agency 263–264; agency and autonomy 266–267, 298–305; agency and identification 266, 288–296; agency and responsibility 265–266, 279–285; agency and the emotions 268–269, 317–325; agency and (the limits of ) volitional conflict 267–268, 307–315; agency, will, and freedom 264–265, 270–277 moral responsibility 106–107 moral theory, action and structure of 440–443 Moran, R. 257, 260, 462 Moreno, A. 92 mortality 390–391 Morton, Adam 319 Morton, J. M. 319, 397, 398, 434 Mossio, M. 92 Mother Teresa 379, 380 motivating reasons 403 motivational hierarchy (Frankfurt) vs. evaluation (Watson) 290–291 Mourelatos, A. 53, 55 Muller, G. E. 143 Mullin, Amy 300 Murdoch, Iris 319 Nagel, T. 358 narratives: actions and circumstances 385; and agency 387–390; constitutive of personal identity 385–386; emotional structure of 334–335; internal structuring of actions 335; meta-narrative of whole life 389; metaphysical identities 387; and mortality 335, 390–391; motivational structure 388–389; self-constitution view 388; selfdeception 386; self-understanding 389 nature and kinds 404 Neapolitan Novels (Ferrante) 458 necessarily less-than-successful agency, notion of 129, 169–171, 176 negative agency 4; actions negatively described 60; differences 64; refraining 59, 62–64; unwitting omissions 59, 60–62 Nelkin, Dana 165 neo-Aristotelian intrinsic power theory 88

472

Index Newell, Allan 366, 367, 373 Nguyen, C. T. 20, 25, 26, 457 noise 96 non-agential powers, one-way powers 134 normative reasons 403, 404–406 normative standards, content of 434–435 normativity and agency 398–399; from agency to reasons 421–423; captain’s agency 421–423; deliberative externalism 423–424; generic agency 424–426; objections 426–428; theorists explaining 435–437 Nozick, Robert 323 Nussbaum, M.C. 84 O’Brien, L. 89, 90, 237 omissions: causal theory of 60; unintentional 60; unwitting 59, 60–62 operational closure 95 organizational theory 88, 95–96 O’Shaughnessy, B. 189, 196, 237 outcome, agency as 21, 28 Pacherie, E. 140 paradigmatic aesthetic value properties 459 Parfit, D. 358, 385, 408 Parkinson’s disease 161 Parsons, T. 50–52, 55 participant-identification 293–295, 294; role of reason-responsiveness (Velleman vs. Smith) 293–295 pathologies of agency 128–129; actual motivating reasons 175; allocating responsibility 173–174; answerability and moral community 172–175; autonomy in context of mental disorder 175–176; implications for decisional capacity 175–176; intention and intentional action 170; notion of necessarily less-than-successful agency 171–172, 174–175; planning theory of intention 170; practical rationality and knowledge 170; range of behaviours 173; responsiveness to reasons 172–175 Paul, S. 17, 238, 239 Perez de Calleja, Mirja 303, 304 performance mistakes 153–154 personal identity and agency: addiction and addictive behavior 382–383; agential identity view 378, 380–383; diachronic aspect 378; higher-order endorsement 377; human history, movement in 376; identity politics 381; integration and separation 377; kind of passivity 382; motivational profile 379–380; patterns of thought, behavior, experience, or commitmenT 379; psychological continuity theories 378 perspectivists 396 phenomenal consciousness 201

The Philosophy of Action (Paul) 17 philosophy of agency: agency and philosophy 2–3; distinction 1–2; exercises of agency 3–6; four pictures of agency 8–15; kinds of agency 6–8; philosophy of agency 15–16 pictures of agency: agency as creation 9; agency as psychological causality 11–13; agency as reason responsiveness 14–15; agency as selfconstitution 10–11 Pilzecker, A. 143 Pink, T. 264, 265 Pippin, Horace 389 planning agency 330–331; brute shuffling 364; constrained maximization 363; continued incommensurability 353–354; and diachronic organization of individual agency 349–351; fecundity of 331, 354; human practical organization 348–349; individual planning agency 331; instrumental rationality 363; linking principle 361; and organization of shared intentional activity 351–352; partial-plan-framed practical reasoning normative reasons 354; planning and selfgovernance 352–354; rational planning 361; self-governance 353; straightforward maximization 363 Pollack, M. 371 post-performance amnesia 142–143, 143 power of reason 276–277 powers 125–126 powers-in-process as causes 33 practicality criterion 413 practical knowledge and agency 222–223, 223; Anscombean views 240–241; doxastic views 239–240; empiricist views 236–237; General 236; inferentialist views 238–239; intentional action exhibits self-consciousness 234; moral knowledge 235; reason and knowledge 241–242; skeptical reactions 235–236 practical reasoning and agency 246–248, 397–398; Broome’s account 414–415; four adequacy criteria 412–413; good reasoning 417–418; reason 415–417 practical reasons 403 pre-commitments 338–339 primitive agency 91 privative manifestations of agency 125 processes vs. events 23 processual goods 342 professional-level expertise 140–142 progression and completion 47–50 progressive paradox 49 proportionate cause, notion of 40 pro-tanto reason, concept of 405 proximate causation 450–454; cause-in-fact 450–451; ceteris paribus 453–454; commitment and temporal proximity 453; legal rule

473

Index 452–453; reasonable foreseeability 450–451; temporal proximity 452; theory of weighting 453; trustworthiness of legal actors 453; voluntary intervention 451–452 psychic economy 12, 264 psychological causality, agency as 11–13 psychological-causation 13 psychologism 407 purposefulness 109–110 Puzo, Mario 69 quality of will 280 Radden, J. 175 Radoilska, L. 128, 129 Ramsey Lewis view 30 rational agency 90, 118–123; acts of reflection 119; causationconstituting normative judgments 123; critics of the Taking Condition 120; deliberation 119–120; doubts about 259–260; epistemic entitlement to self-knowledge 256–257; The Knowledge Gap 121–122; rational causation 119; self-conscious engagement 118 rationality 105–106, 331–332, 357–364 Raz, J. 418 reactive attitudes 174, 280, 282 realization 55–56, 56 reason responsiveness, agency as 14–15 reasons and rationality 415–417; and actions 403; and explanation of action 406–409; and knowledge 241–242; motivating reasons 406–409; nature and kinds 404; normative reasons 404–406; reasoned desires 371–373 reasons why 172 reconsideration 361–364 refraining 59, 62–64; episodic refraining 64; quasi-causation of non-occurrence 63 Reiter, Raymond 368 resoluteness 361–364 responsibility 3; being at fault 282; blaming 281; child’s fault 282; everyday uses of 279; extreme or unusual circumstances 280; interpersonal relationships 280, 281; as mattering 283; objective attitude 280; problem of free will and moral responsibility 282–283; reactive attitudes 280, 281, 283–284; responsible agency 265–266, 282–285; the strains of involvement 280; use of responsible 282 Ringuette, Marc 371 Rodl, S. 236 Rorty, A. O. 152, 323 Rosati, C. S. 432 Rules of Play (Salen and Zimmerman) 78 Russell, B. 28

Russell, Stuart 369 Ryle, G. 52, 53, 133, 136, 228, 229, 272 Salen, Katie 78 Salomone-Sehr, Jules 397, 398 Salvucci, Dario D. 374 Scanlon, T. M. 302, 415, 416 Scardamalia, Marlene 145 scepticism 272 Schafer, K. 435 Schechtman, M. 333, 386 Scheffler, S. 343 Schroeder, T. 292, 293, 413 Schwenkler, J. 223, 224 Schwitzgebel, E. 175 Sehon, S. 21, 22 Seidman, J. 291, 300 selectional explanations and teleology 42–44 self-constitution, agency as 10–13 self-governing policies (Bratman) vs. caring (Seidman) 291–292 self-knowledge and agency 224–225; active thoughts and attitudes 254–255; agential models of self-knowledge 256–260; as assured by normative commitments 258; doubts about rational agency 259–260; first-person authority 255–256; initial contrast 253; rational agency 256–257 self-maintenance, agency as 10–11 self-motion, agency as 10 Sellars, W. 237 Seltzer, Mark 213 Setiya, K. 236–240, 248, 330, 343, 435 Shah, N. 431 Shakespeare, Tom 161 Shanahan, Murray P. 370 shared intentional activity (SIA) 351–354 shmagency challenge 399 Shoemaker, D. 165, 174, 291, 300 Shusterman, Richard 141 Silverstein, M. 435, 437 Simon, Herbert A. 366, 367 Simon, Robert L. 78, 79 Simple View 134 simple view of emotions 318 skeptical reactions 235–236 skills 125–126; vs. abilities 133; dispositional conception of 137n15; expert skill, defined 139; powers and 125–126 Slingerland, Edward 144 Small, W. 125, 126, 136, 214, 240 Smith, A. 295 Smith, M. 60–64, 132 Smith, Matthew Noah 180, 182 Snedegar, J. 354 sophisticated rationality 332 Sorensen, R. 155

474

Index sortal 227 Soteriou, M. 180 stance-identification 290–293; alternative interpretations fall short 293; motivational hierarchy (Frankfurt) vs. evaluation (Watson) 290–291; self-governing policies (Bratman) vs. caring (Seidman) 291–292; wholism vs. special attitude-types (Arpaly and Schroeder) 292–293 standards for success 150 standards of agency 149–150 standard story of action 12, 19, 22, 24, 47–48, 131 Stanley, J. 215 Sterelny, K. 92 Steward, H. 50–52, 54, 56, 88, 89, 93 Stout, R. 28, 53 straightforward maximization 332 Strawson, Galen 188, 265 Strawson, Peter F. 280–284 striving and achievement play 81–82 striving play 81–82 structured, non-processual goods 342 structuring agency 193–194; deafferentation and automaticity of everyday agency 193; deafferentation and non-observational knowledge of action 193–194; vestibular disturbance and egocentric perception and action 194 sub-intentional actions 103 substances as causes 32–33 Taylor, Charles 462 teleological explanations: causal explanations in disguise 22; vs. efficient aim theories 432–433; functional or selectional explanations 22; general account of 43; goal-oriented behaviors 21 telic and atelic goals 5 temporally additive goods 342 temporally extended agency 57n4 temporal positional goods 342 Tenenbaum, S. 433, 434, 436 theories of minimal agency: evolutionary etiology 94–95; explanatory role 92–93; features of 98, 98; fitness estimation 96–98; intrinsic power 93–94; organizational theory 95–96 Thomason, Richmond H. 333, 372 Thompson, M. 23, 49, 56, 235 Thomson, Judith Jarvis 422 time and agency 331–332, 357–364; agency and personal identity 333–334, 376–383; artificial and machine agency 333, 366–374; diachronic agency 330, 336–345; narrative and mortality 334–335, 385–391; philosophical questions and challenges

329–330; planning agency 330–331, 348–354; and rationality 331–332, 357–364 transcending reflective agency 143 transparency method 257 two entities view (TEV): accounts of composition 53–54; atelic action sentences 55; experiential present 54–55; processes and events 53; substantive disagreement 53 two properties view (TPV): category of individual processes 50–51; interruption of actual progression 51–52; progressive paradox 52 type-token model of instantiation 56 Ulatowski, J. 388, 391 unconscious attention 143 unconscious guidance, see zombie actions van Hateren, J. H. 88, 92 varieties of agency 6–7 Velleman, J.D. 143–145, 211, 248, 293–295, 296, 386, 389, 391, 425, 433–436 Vermazen, Bruce 63, 65 virtue theorists 440 Vogler, C. 71 volitional conflict and agency 267–268; allthings-considered normative judgment 312–313; conflicting goals or commitments 308; dilemmatic conflict 311; exercise of agency 314; experience of conflicting ends 310; normative conflict 311; perfect and imperfect requirements 311; psychic conflicts 307–308; self-imposed substantive requirement 312; substantive commitments 309, 310; volitional necessity 308–310 Walden, K. 398, 399, 435 Watson, G. 152, 290–293 Wedgwood, R. 431 Wefald, Eric 369 Westlund, Andrea C. 267 wholism vs. special attitude-types (Arpaly and Schroeder) 292–293 Wigley, Mark 213 will and freedom 264–265; agency and will 270–271; goal-direction and intentionality 272–273; goal-direction and motivating power 273–276; place of will in action theory 277; power of reason 276–277; purposiveness 270; rationality or reasonableness 270–271; scepticism 272; self-determination 271 Williams, B. 272, 390 Williamson, T. 215 Wilson, G. 39 Wolf, Susan 457 Wooldridge, M. 139 Wooldridge, Michael J. 139, 371

475

Index Wulf, G. 139 wu-wei, Taoist concept of 126 xenomelia 198 Yablo, S. 40 Yaffe, G. 401

Zimmerman, Erich 78 zombie actions 182; conditions of illusion 207; dorsal and ventral stream representations 208; experience-based control 206; subject-level visual states 207–208; subpersonal states 207

476