The Routledge Handbook of Phenomenology of Agency [1 ed.] 9780367568696, 9781138098978, 9781315104249

Phenomenology has primarily been concerned with questions about knowledge and ontology. However, in recent years the ris

135 46 6MB

English Pages 436 [437] Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Cover
Half Title
Series Page
Title Page
Copyright Page
Table of Contents
Notes on Contributors
Introduction
Part I: Important Figures: From Brentano to Tengelyi
1 Franz Brentano’s Critique of Free Will
2 Phenomenology of Willing in Pfänder and Husserl
3 Alexander Pfänder’s Phenomenology of Motivation
4 Scheler’s Phenomenology of Freedom and his Theory of Action
5 The Intentionality and Positionality of Spontaneous Acts: Adolf Reinach’s Account of Agency
6 Dietrich Von Hildebrand on the Will and Intentional Agency
7 The Varieties of Activity: Hans Reiner’s Contribution
8 Martin Heidegger: From Fluid Action to Gelassenheit
9 Edith Stein: Psyche and Action
10 Action in the Phenomenology of Alfred Schütz
11 Determined to Act: On the Structural Place of Acting in Sartre’s Ontology of Subjectivity
12 Emmanuel Levinas: Freedom and Agency
13 Hannah Arendt: Plural Agency, Political Power, and Spontaneity
14 Merleau-Ponty and Agency
15 Paul Ricœur: A Phenomenological Hermeneutics of Meaningful Action
16 Operari Sequitur Esse: Hermann Schmitz’s Attitudinal Theory of Agency, Freedom, and Responsibility
17 Hubert Dreyfus: Skillful Coping and the Nature of Everyday Expertise
18 Life is an Adventure: Làszló Tengelyi’s Phenomenology of Action
Part II: Systematic Perspectives
Phenomenology of Agency 1: General Issues
19 On the Satisfaction Conditions of Agentive Phenomenology: A Dialogue
20 Ambulo!: Structures of Phenomenology and Ontology in Action
21 Will-Power: Essentially Embodied Agentive Phenomenology, by Way of O’Shaughnessy
22 The Phenomenology of Agency and the Cognitive Sciences
Phenomenology of Agency 2: Aspects of Agency
23 The Phenomenology of Free Agency
24 The Phenomenology of Rational Agency
25 Acting, Choosing, and Deliberating
26 Involuntariness: Actions and their Context
27 Moral Experience: Its Existence, Describability, and Significance
Index
Recommend Papers

The Routledge Handbook of Phenomenology of Agency [1 ed.]
 9780367568696, 9781138098978, 9781315104249

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

THE ROUTLEDGE HANDBOOK OF PHENOMENOLOGY OF AGENCY

Phenomenology has primarily been concerned with questions about knowledge and ontology. However, in recent years the rise of interest and research in phenomenology and embodiment, the emotions and cognitive science has seen the concept of agency move to a central place in the study of phenomenology generally. The Routledge Handbook of Phenomenology of Agency is an outstanding reference source to this topic and the first volume of its kind. It comprises 27 chapters written by leading international contributors. Organised into two parts, the following key topics are covered: • • • • • • • • •

major figures the metaphysics of agency rationality voluntary and involuntary action moral experience deliberation and choice phenomenology of agency and the cognitive sciences phenomenology of freedom embodied agency

Essential reading for students and researchers in phenomenology, philosophy of mind, metaphysics and philosophy of cognitive science, The Routledge Handbook of Phenomenology of Agency will also be of interest to those in closely related subjects such as sociology and psychology. Christopher Erhard is Lecturer at the Faculty of Philosophy, Philosophy of Science and the Study of Religion at the Ludwig-Maximilians-University of Munich, Germany. Tobias Keiling, currently an Alexander von Humboldt Research Fellow at Somerville College, Oxford, UK, is based at the University of Bonn, Germany.

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 OF SKILL AND EXPERTISE Edited by Ellen Fridland and Carlotta Pavese THE ROUTLEDGE HANDBOOK OF PHENOMENOLOGY AND PHENOMENOLOGICAL PHILOSOPHY Edited by Daniele De Santis, Burt Hopkins and Claudio Majolino THE ROUTLEDGE HANDBOOK OF THE PHILOSOPHY AND SCIENCE OF PUNISHMENT Edited by Farah Focquaert, Elizabeth Shaw, and Bruce N. Waller THE ROUTLEDGE HANDBOOK OF PHENOMENOLOGY OF AGENCY Edited by Christopher Erhard and Tobias Keiling THE ROUTLEDGE HANDBOOK OF FEMINIST PHILOSOPHY OF SCIENCE Edited by Sharon Crasnow and Kristen Intemann For more information about this series, please visit: https://www.routledge.com/Routledge-Handbooks-in-Philosophy/book-series/RHP

THE ROUTLEDGE HANDBOOK OF PHENOMENOLOGY OF AGENCY

Edited by Christopher Erhard and Tobias Keiling

First published 2021 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2021 selection and editorial matter, Christopher Erhard and Tobias Keiling; individual chapters, the contributors The right of Christopher Erhard and Tobias Keiling to be identified as the authors 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: Erhard, Christopher, 1980– editor. | Keiling, Tobias, 1983– editor. Title: The Routledge handbook of phenomenology of agency / edited by Christopher Erhard and Tobias Keiling. Description: New York, NY: Routledge, 2021. | Series: Routledge handbooks in philosophy | Includes bibliographical references and index. | Identifiers: LCCN 2020021236 (print) | LCCN 2020021237 (ebook) | ISBN 9781138098978 (hbk) | ISBN 9781315104249 (ebk) Subjects: LCSH: Agent (Philosophy) | Phenomenology. Classification: LCC B105.A35 R675 2021 (print) | LCC B105.A35 (ebook) | DDC 128/.4—dc23 LC record available at https://lccn.loc.gov/2020021236 LC ebook record available at https://lccn.loc.gov/2020021237 ISBN: 978-1-138-09897-8 (hbk) ISBN: 978-1-315-10424-9 (ebk) Typeset in Bembo by codeMantra

CONTENTS

Notes on contributors

viii



PART I

Important Figures: From Brentano to Tengelyi



v

5

Contents

10 Action in the phenomenology of Alfred Schütz Michael Barber 11 Determined to act: on the structural place of acting in Sartre’s ontology of subjectivity Simone Neuber

121

134

12 Emmanuel Levinas: freedom and agency Michael L. Morgan

147

13 Hannah Arendt: plural agency, political power, and spontaneity Marieke Borren

158

14 Merleau-Ponty and agency Thomas Baldwin

175

15 Paul Ricœur: a phenomenological hermeneutics of meaningful action Timo Helenius

189

16 Operari Sequitur Esse: Hermann Schmitz’s attitudinal theory of agency, freedom, and responsibility Henning Nörenberg

208

17 Hubert Dreyfus: skillful coping and the nature of everyday expertise Justin F. White

219

18 Life is an adventure: László Tengelyi’s phenomenology of action Tobias Keiling

235

vi

Contents PART II

Systematic Perspectives

261

Phenomenology of Agency 1: General Issues

263

19 On the satisfaction conditions of agentive phenomenology: a dialogue Terry Horgan and Martine Nida-Rümelin

264

20 Ambulo!: structures of phenomenology and ontology in action David Woodruff Smith 21 Will-power: essentially embodied agentive phenomenology, by way of O’Shaughnessy Robert Hanna

300

314

22 The phenomenology of agency and the cognitive sciences Shaun Gallagher

336

Phenomenology of Agency 2: Aspects of Agency

351

23 The phenomenology of free agency Galen Strawson

352

24 The phenomenology of rational agency Roberta De Monticelli

362

25 Acting, choosing, and deliberating John J. Drummond

376

26 Involuntariness: actions and their context Günter Figal

388

27 Moral experience: its existence, describability, and significance Uriah Kriegel

398

Index

415

vii

CONTRIBUTORS

Thomas Baldwin is an Emeritus Professor of Philosophy in the Department of Philosophy at the University of York, UK. He edited Merleau-Ponty: Basic Writings (Routledge 2004), Reading Merleau-Ponty (Routledge 2007), and The Cambridge History of Philosophy 1870–1945 (Cambridge University Press 2003). He was editor of MIND 2005–2015. Michael Barber is Professor of Philosophy at Saint Louis University in the USA, and has authored seven books and eighty articles on the phenomenology of the social world. His book, The Participating Citizen, won the 2007 Ballard Prize. Marieke Borren is Assistant Professor in Philosophy at the Open University, the Netherlands. Her research expertise lies at the intersection of continental political philosophy and phenomenology. She has published widely on Hannah Arendt’s political phenomenology and takes an interest in the phenomenology of gender and race. Antonio Calcagno is Professor of Philosophy at King’s University College, London, Canada. He is the author of The Philosophy of Edith Stein (2007) and Lived Experience from the Inside Out: Social and Political Philosophy in Edith Stein (2014). He is the editor of the recent volume Contemporary Italian Political Philosophy (2015). His latest edited volume is titled Gerda Walther’s Phenomenology of Sociality, Psychology, and Religion. Roberta De Monticelli was Professor of Modern and Contemporary Philosophy at the University of Geneva, Switzerland (1989–2004), and is currently Professor of Philosophy of Personhood at San Raffaele University, Milan, Italy. She is Chief Editor of Phenomenology and Mind – The Online Journal of the Faculty of Philosophy, San Raffaele University. Among her books are Towards a Phenomenological Axiology: Discovering What Matters (Palgrave, 2020); Il dono dei vincoli – Per leggere Husserl (Garzanti, Milan, 2018) (trans.: The Gift of Bonds – Husserl’s Phenomenology Revised, forthcoming, Springer); and La novità di ognuno. Persona e libertà (Milan, 2009). Francesca De Vecchi is Professor of Theoretical Philosophy at San Raffaele University (Milan, Italy). Her current research topics focus on qualitative ontology and eidetics in the fields of the phenomenology of social reality, intersubjectivity and gendered personal identity. With reference to the seminal works of Adolf Reinach, Edith Stein, and Max Scheler, she has published several papers contributing to the analysis of the various types of intentionality viii

Contributors

involved in the experience of social reality – what she labels “heterotropic intentionality” She is co-editor of the journal Phenomenology and Mind, member of the governing board of PERSONA – Research Centre in Phenomenology and Sciences of the Person, and director of the Interfaculty Centre for Gender Studies at San Raffaele University. John J. Drummond is the Robert Southwell, S.J. Distinguished Professor of Philosophy and the Humanities at Fordham University, New York, USA. He is the author or editor of eight books and the author of numerous articles on phenomenology, intentionality, the emotions, value theory, and ethics. Christopher Erhard is a Lecturer at the Faculty of Philosophy, Philosophy of Science, and the Study of Religion at the Ludwig-Maximilians-University of Munich. He is working on the intersections between phenomenology and analytic philosophy of mind (esp. intentionality) and theory of action. He is the author of Denken über nichts. Intentionalität und Nicht-Existenz bei Husserl (De Gruyter 2014). Günter Figal is Professor Emeritus of Philosophy at the University of Freiburg, Germany. He has held many appointments as visiting professor, and in 2017 he was awarded the International Chair of Philosophy Jacques Derrida at the University of Torino. His recent books include Philosophy as Metaphysics. The Torino Lectures (2014). Shaun Gallagher is the Lillian and Morrie Moss Professor of Excellence in Philosophy at the University of Memphis, and Professorial Fellow at SOLA, Faculty of Law, Humanities and the Arts, University of Wollongong (AU). He held the Humboldt Foundation Anneliese Maier Research Fellow (2012–2018). His latest book is Action and Interaction (OUP, 2020). Sacha Golob is a Senior Lecturer in Philosophy at King’s College London and the CoDirector of the Centre for Philosophy and the Visual Arts (CPVA). Before joining King’s, he was a Fellow at Peterhouse, Cambridge. He has published extensively on modern French and German Philosophy. His current research explores conceptions of ethical change. Robert Hanna is an independent philosopher and philosophical generalist; he’s held research/ teaching positions at the Universities of Cambridge UK, Colorado-Boulder USA, Luxembourg, PUC-PR Brazil, Yale, and York CA; and he’s the author/co-author of 11 books, including, most recently – the five-book series The Rational Human Condition (2015–2018) and The Mind-Body Politic (co-authored with Michelle Maiese, 2019). Timo Helenius received his PhD in philosophy from Boston College in 2013 and was a Visiting Scholar at the Department of Religious Studies in Brown University from 2016 to 2019. Helenius has taught philosophy and ethics at Boston College, Mount Ida College, and – most recently – at the University of New Brunswick Saint John as a Visiting Assistant Professor of Philosophy. Helenius’s research and many publications have focused on contemporary Continental philosophy in general and on Paul Ricoeur in particular. His work, Ricoeur, Culture, and Recognition: A Hermeneutic of Cultural Subjectivity, was published by Lexington Books (Rowman & Littlefield) in 2016. Helenius holds the title of Docent in Philosophy at the University of Turku. Terry Horgan is Professor of Philosophy at the University of Arizona. He works in philosophy of mind, metaphysics, epistemology, and metaethics. His publications include Connectionism and the Philosophy of Psychology with J. Tienson (MIT, 1996), Austere Realism with M. Potrč (MIT, 2008), The Epistemological Spectrum with D. Henderson (Oxford, 2011), and Essays on Paradoxes (Oxford, 2017). ix

Contributors

Tobias Keiling, currently an Alexander von Humboldt Research Fellow at Somerville College, Oxford, is based at the University of Bonn, Germany. He holds a PhD in philosophy from Boston College, USA, and the University of Freiburg, Germany. He has published numerous articles on themes in Post-Kantian philosophy and on Heidegger in particular. Eugene Kelly is Professor of Philosophy at the New York Institute of Technology. Among his publications are Structure and Diversity: Studies in the Phenomenological Philosophy of Max Scheler (Kluwer, 1997); Material Ethics of Value: Max Scheler and Nicolai Hartmann (Springer, 2011); The Basics of Western Philosophy (Greenwood Press, 2004; reprint: Prometheus Books, 2006); and a translation into English of Nicolai Hartmann’s Ästhetik (De Gruyter, 2014). Uriah Kriegel is a Professor of Philosophy at Rice University. He is the author of Subjective Consciousness: A Self-Representational Theory (OUP, 2009), The Sources of Intentionality (OUP, 2011), The Varieties of Consciousness (OUP, 2015), and Brentano’s Philosophical System: Mind, Being, Value (OUP, 2018). Karl Mertens is Professor of Philosophy at the Institute of Philosophy in Würzburg University. Mertens studied Philosophy, German Literature, and History at the Universities of Freiburg i. Br., Zürich, and Cologne. From 2005 to 2017 he was co-editor in chief of the journal Phänomenologische Forschungen/Phenomenological Studies (Meiner). He has published several articles on theory of action, social philosophy, ethics, theory of knowledge, philosophy of mind, and phenomenology. Michael L. Morgan is the Chancellor’s Professor of Philosophy and Jewish Studies (emeritus) at Indiana University and the Senator Jerahmiel S. and Carole S. Grafstein Chair in Jewish Philosophy and Jewish Studies (emeritus) at the University of Toronto. He has taught at Northwestern, Yale, Stanford, Toronto, and Princeton. He is the author and editor of 22 books – most recently The Oxford Handbook of Emmanuel Levinas (2019), Levinas’s Ethical Politics (Indiana, 2016), and Michael L. Morgan: History and Moral Normativity (Brill, 2018). Simone Neuber is Lecturer at the University of Heidelberg. She is the author of a monograph on issues related to the “Paradox of Fiction” and has published various articles on a.o. Sartre and Heidegger. Currently, she is working on a book on self-deception. Martine Nida-Rümelin is Professor of Philosophy at the University of Fribourg, Switzerland. She works mainly on topics in the philosophy of mind such as phenomenal consciousness and phenomenal concepts, personal identity, and the phenomenology of agency and freedom. In 2019, she was awarded the renowned Jean Nicod Prize. Her book Conscious Subjects – Sketch of a Theory will be published 2021 by Oxford University Press. On the topic of this volume she published several articles, e.g. “Doings and Subject Causation” (Erkenntnis 67, 2007) and “Freedom and the Phenomenology of Agency” (Erkenntnis 83, 2018). Henning Nörenberg is a Postdoctoral Researcher at the University of Rostock. He works on various topics in phenomenology, social ontology, political philosophy and philosophy of religion, e.g., shared normative background orientations and their affective dimension. Alessandro Salice is Lecturer at the Department of Philosophy, University College Cork. Previously, Alessandro held postdoctoral positions at the University of Graz, University of Basel, University of Vienna, and at the Center for Subjectivity Research in Copenhagen. His expertise stretches from classical accounts of collective intentionality and joint action to phenomenology, philosophy of mind, and moral psychology. His papers have been published in x

Contributors

several journals, including Phenomenology and the Cognitive Sciences, European Journal of Philosophy, Journal of Consciousness Studies, Review of Philosophy and Psychology, Frontiers in Psychology, Topoi, and International Journal of Social Robotics. Denis Seron is FNRS Senior Research Associate and Head and head of the Center for Phenomenological Research at the University of Liège, Belgium. He has published multiple books and articles in the field of contemporary philosophy. David Woodruff Smith is Distinguished Professor of Philosophy at the University of California, Irvine. Addressing aspects of phenomenology, ontology, and philosophy of mind, his books include Husserl (2013), Mind World (2004), The Circle of Acquaintance (1989), and Husserl and Intentionality (with Ronald McIntyre, 1982). He is co-editor of Phenomenology and Philosophy of Mind (2005, with Amie L. Thomasson) and The Cambridge Companion to Husserl (1995, with Barry Smith). Galen Strawson holds the President’s Chair of Philosophy at the University of Texas at Austin. His books include Freedom and Belief (1986) and Mental Reality (1994). His most recent books are The Subject of Experience (2017) and Things That Bother Me (2018). Genki Uemura is Associate Professor at Graduate School of Humanities and Social Sciences, Okayama University, Japan. His research interests include Husserl’s phenomenology, phenomenology in the Munich and Göttingen schools, and the opposition between them. Justin F. White is an Assistant Professor of Philosophy at Brigham Young University. His research focuses on 19th- and 20th-century European philosophy and philosophy of agency. His recent work has been published in Midwest Studies in Philosophy, The Southern Journal of Philosophy, and The Kierkegaardian Mind (Routledge).

xi

INTRODUCTION Christopher Erhard and Tobias Keiling

The stimulus for editing this volume on the “phenomenology of agency” can be traced back to what Terry Horgan, John Tienson, and George Graham wrote in 2003: In philosophy of mind... there has been a wide-spread, and very unfortunate, tendency to ignore the phenomenology of doing altogether – and to theorize about human agency without acknowledging its phenomenology at all, let alone seeking to accommodate it. It is time to get beyond this major philosophical blindspot. (Horgan/Tienson/Graham 2003: 332) The first challenge for an attempt to overcome this blindness, however, is to say more about what the desired addition to the theory of action is since there are at least three meanings associated with “phenomenology”. First, the term can be used as an historical designation to refer to a philosophical tradition founded by Edmund Husserl which is still alive today. Second, there is the methodological sense, according to which “phenomenology” refers to a certain way of philosophizing. Importantly, and with a view to the history of the phenomenological tradition in particular, there are several methodological ideas that may all be phenomenological in spirit but are by no means synonymous or equivalent. Just to mention some of them: “saving the appearances”, “returning to the things themselves (Sachen selbst)”, “describing the essential structures of lived-experiences (Erlebnisse)”, “theorizing from the first-person perspective”, “letting something be seen”, “intentional analysis”, “intuiting essences”, “constitutive analysis”, “transcendental analysis”, “hermeneutics of facticity”, “analysis of the normativity of meaning”, “eidetic variation”. Clearly, different approaches to the problem of human agency emerge relative to these ideas that have been associated with “phenomenology” as a philosophical method, and it is quite unclear how they relate. For instance, the attempt to identify essential, “eidetic” structures of agency can be associated with but does not follow by necessity from a methodological preference for the “first-person perspective”, and although these ideas have as a matter of historical fact been combined by authors such as Husserl, argument would be needed to show why a phenomenological methodology should include these two distinct ideas. Although often taken as self-evident, the relation between the historical and the methodological implications of “phenomenological” philosophy becomes the more controversial the closer one 1

Christopher Erhard and Tobias Keiling

looks. In particular, it cannot do to concentrate on merely one author as representative of “phenomenology” in the historical and the methodological sense. To further complicate matters, a third, weaker and not necessarily related meaning must be distinguished. According to this experiential sense of “phenomenology of agency”, the term refers to an agentive phenomenal dimension, the “what-it-is-likeness” of performing an act or action, if such can be distinguished. This is the sense in which Horgan, Tienson, and Graham are using the term in passage quoted earlier. Although there is clearly a connection between what they require of a philosophical account of agency and the starting point of many accounts of phenomenological method, to specify this connection is a related, but different task, as is working to this end with the help of authors from the phenomenological tradition. All chapters of this volume in one way or another relate to at least one of these senses of the term “phenomenology”, but it should not be assumed that the term is used synonymously or equivalently. Editing this collection, we rather hope to highlight the range and diversity of approaches to human agency that are phenomenological in one sense or another, even if this entails showing that the question as to a positive and definite proposal for how to understand phenomenology is still an open issue. We would expect this volume to stir rather than settle a discussion of that question. There are two further goals for this volume, associated more closely with its two parts. First, we wish to portray the richness of phenomenological accounts of human agency. It is for this reason that perhaps little-known phenomenologists such as Alexander Pfänder, Edith Stein, Adolf Reinach, Dietrich von Hildebrand, and Hans Reiner are also covered in the chapters of the first part. Because these authors worked in the first decades of the 20th century, recognizing that they have made substantial contributions to the question of agency in their own right modifies a received view according to which the phenomenological tradition only began to take human agency seriously when more “enactivist” philosophers such as Max Scheler, Martin Heidegger, Hannah Arendt, Jean-Paul Sartre, Maurice Merleau-Ponty, and Paul Ricœur entered the stage. Quite the contrary: not only did Husserl work extensively on the notion of action (Handlung) and related lived-experiences such as volitions, strivings, and intentions; many of his early followers developed succinct analyses of the realm of human activity. In this vein, one important aim of Part I of our collection (Important Figures: From Brentano to Tengelyi) is to challenge Uriah Kriegel’s recent claim that “the golden decade of conative phenomenology … arrives only in the French Philosophy in the 1940s” (Kriegel 2013: 538). As scholars have become increasingly aware through more recent editions and translations of Husserl and his first-generation followers, an earlier “golden era” of conative and agentive phenomenology already took place roughly between the publication of Husserl’s Logical Investigations (1900/1) and 1927, when Heidegger’s Being and Time was published. Engaging with more recent discussions and basic questions concerning human agency, the chapters in Part I show that the history of phenomenology is not a mausoleum of dead ideas and thinkers but occasion for a fruitful dialogue with the past. Second, Part II of our collection (Systematic Perspectives) gathers papers by eminent scholars with quite different backgrounds all of whom make use of at least one of the three notions of “phenomenology” in their work. In this way, we want to underline that a phenomenological approach is relevant, if not indispensable in the treatment of fundamental problems in the philosophy of action. While the papers in roughly the first half of this second part focus on general issues concerning the phenomenology of action, the contributions in the second section address specific aspects or dimensions of human agency such as freedom, rational action, deliberation, choice, and morality. 2

Introduction

Some of the contributors to this Handbook also participated in a conference entitled “Phenomenology in Action” that took place in 2016 at the University of Munich (LMU). We would like to thank the Deutsche Forschungsgemeinschaft (DFG) and the Münchener Universitätsgesellschaft (MUG) for funding this conference. Moreover, we would like to thank the anonymous reviewers for their help during the evaluation process of the contributions. We also want to thank the contributors for all the work they have put into this project and also for their patience. Finally, we thank Tony Bruce and Adam Johnson from Routledge who were invaluable in realizing this volume. Christopher Erhard, Munich Tobias Keiling, Oxford/Bonn August 2020

References Horgan, T., Tienson, J. and Graham, G. (2003) “The Phenomenology of First-Person Agency,” in S. Walter and H.-D. Heckmann (eds.), Physicalism and Mental Causation: The Metaphysics of Mind and Action, Exeter: Imprint Academic, pp. 323–340. Kriegel, U. (2013) “Understanding Conative Phenomenology: Lessons from Ricœur,” Phenomenology and the Cognitive Sciences 12, 537–557.

3

PART I

Important Figures From Brentano to Tengelyi

1 FRANZ BRENTANO’S CRITIQUE OF FREE WILL Denis Seron

The German philosopher Franz Brentano is usually portrayed not only as one of the founding fathers of the so-called phenomenological tradition but also as having played a significant role in the history of contemporary ethics, through his theory of value and will. In spite of this, Brentano offered no proper theory of action in the vein of later attempts by direct or indirect followers. His ethics is basically about feelings, and feelings can be ethically right or wrong even if they intrinsically involve no reference to action. This is so even in the case of desire: for example, you can desire that the weather gets warmer (Montague 2017: 113). However, Brentano’s account of volitions does make reference to action.1 Insofar as Brentano bequeathed us a fully developed theory of the will of his own, it can be said that he indirectly contributed to the theory of action. This is particularly true of his critique of free will, which this chapter aims to discuss and explore. Brentano intended to investigate free will in the fifth of the six planned books of the Psychology from an Empirical Standpoint, devoted to emotion and will (Brentano 1924: 1, engl. trans.: xxvii, 1925: 110, ftn., engl. trans.: 254) – a book which, unfortunately, he never wrote. Most of Brentano’s reflections on this topic are found in Part 3 of his 1876–1894 Vienna lectures on practical philosophy that were posthumously published as The Foundation and Construction of Ethics by Franziska Mayer-Hillebrand. Other relevant sources include the second volume of the Psychology and the 1889 lecture The Origin of Our Knowledge of Right and Wrong. Before beginning, three points should be underlined. First, Brentano considers the will neither as a sufficient nor even as a necessary condition for freedom: there can be volitions that are not free as well as free acts – for example “the pangs of remorse over an earlier transgression, malicious pleasure, and many other phenomena of joy and sorrow” – that are not volitions (Brentano 1925: 110–111; engl. trans.: 254–255). Second, the problem of freedom, as Brentano sees it, has to do with time and causation. As such, it belongs to genetic rather than to descriptive psychology.2 Third and finally, Brentano is interested in the freedom of will rather than freedom of action. In his Vienna lectures on ethics, he makes a distinction between the actus a voluntate imperatus and the actus elicitus voluntatis (Brentano 1952: 235; engl. trans.: 147).3 The former is the action that you desire to bring about; the latter is your act of will itself. As we will see, Brentano’s extended discussion of the determinism-indeterminism debate is mostly about the actus elicitus voluntatis. 7

Denis Seron

Determinism and empiricism The most striking feature of the mature Brentano’s approach to free will is its uncompromising determinism.4 A crucial aspect of this determinism is what could be called Brentano’s “actualism” – an epistemological view which follows from his radical empiricism (another aspect, his so-called necessitarianism, will be considered later on). At first glance, it seems natural to think that the free will issue has to do with the possibility or impossibility to act or will otherwise than we actually do. Our will is said to be free insofar as we could, at least in some cases, want to act differently. But Brentano rejects this idea – which he attributes to Descartes (Brentano 1952: 242; engl. trans.: 151). The reason for this is that, at a more general level, he considers reference to possible events incompatible with his empiricism. Thus, he heavily criticizes how John Stuart Mill appeals to “possibilities” in order to dispense with the non-empirical assumption of real things beyond appearances (Brentano 1930: 130; engl. trans.: 77). In his critique of indeterminism, Brentano repeatedly emphasizes that, since merely possible volitions are not given in experience, no empirical argument can support the indeterminist’s principle of alternative possibilities: It is just as impossible for the determinists to perceive that we cannot do otherwise as for the indeterminists to perceive that we can. Only concrete facts can be perceived; not possibilities, impossibilities, or necessities. (Brentano 1952: 269; engl. trans.: 167) The indeterminist contends that we are not determined to will this or that, and that this indetermination of the will is empirically evident. In other words immediate consciousness, that is, inner perception, tells us that (at least when the opposite motives are of equal weight) we are able to will the opposite of what we are willing now. To this, Brentano objects that “we can only perceive what is actual, not what is merely possible” (Brentano 1952: 246; engl. trans.: 154). Knowledge of possibility is derivative, that is, acquired on the basis of the experience of what is actual. But how could the indeterminist validly infer that she can will non-A from her experience that she wills A? In the same vein, an argument in favor of indeterminism may be that “our will occasionally withstands a passionate desire with resolution and energy, but then leaves off the battle in a state of exhaustion, only to renew its resistance after a pause” (Brentano 1952: 251; engl. trans.: 156). In such cases, it seems that the will is not determined, but resists motives that would determine it if it didn’t resist them. Brentano’s objection to this argument is that no experience can ever support it and hence that it is purely fictional. What does the indeterminist mean when she says that the will struggles against desires? She cannot have in view here the (present) act of will, for the act of will must come after the motives and thus does not exist yet. Does “the will” denote the ability to will? The ability to will is not the sort of thing that can act or offer a resistance to anything: “It is simply a potentiality.” Hence, Brentano concludes, the correct interpretation is not that the will battles against the motives, but that “one motive battles against another.”5 Two points should be noted before we move into a closer discussion of Brentano’s determinism. First, Brentano constantly assumes that, if the free will dispute makes sense, then it can and must be resolved on the basis of experience alone. It is worth noticing that the same concern prompted Brentano to warn against the use of abstract phrasings such as “I have overcome my passion” or “My inclination to do my duty overcame my leanings towards 8

Franz Brentano

pleasure.” Because abstracts are no more than fictions, we should say instead: “One part of myself gained a victory over another part of myself ” (Brentano 1952: 251; engl. trans.: 156). Second, Brentano’s empirical approach to free will does not involve abandoning all talk of dispositions as nonsensical. The key point is rather that, since habits and inclinations cannot be perceived, but only derived from perceptual experience (Brentano 1968: 119, 238–239; engl. trans.: 93, 172), they have no (empirical) reality except insofar as they function as actual motives for actual actions.6

Brentano’s demonstration of determinism Brentano provides a detailed demonstration of determinism in his Vienna lectures on ethics. The issue at stake concerns the will rather than action. By determinism, Brentano means the view that the will is causally determined. Thus, he seems to take it for granted that, within certain limits that are fixed by internal and external conditions, the actus a voluntate imperatus is free (Brentano 1952: 235 ff.; engl. trans.: 147 ff.). In short: you can do what you want, within certain limits. Brentano’s contention is that not the willed action, but the will itself is causally determined. A second point to keep in mind is that the determinism-indeterminism dispute is not about freedom in general, but only about freedom in the sense of being “free from necessity” (Brentano 1952: 235; engl. trans.: 147). The question is whether the actus elicitus voluntatis – the act of will – is “necessarily determined,” that is, caused in such a way that its cause is not the will itself and that it makes it necessary for it to be so and so. The third part of the Vienna lectures is devoted to establishing successively that indeterminism is very improbable (Brentano 1952: 276 ff.; engl. trans.: 172 ff.), that it lacks explanatory power (Brentano 1952: 279 ff.; engl. trans.: 173 ff.), and that it is necessarily false. I shall here restrict myself to the final argument. Brentano distinguishes between an extreme and a moderate form of indeterminism. The former is the view that the will simply has no cause. The latter holds that the will has no cause that makes it necessary (Brentano 1952: 267; engl. trans.: 166). The moderate indeterminist admits that the will may be determined, but denies that it is determined necessarily (Brentano 1952: 283; engl. trans.: 175). In Brentano’s view, both extreme and moderate indeterminism thus defined involve the assumption of an “absolute accident,” that is, of something that is not necessarily determined. Brentano’s aim is to demonstrate that this assumption is self-contradictory (Brentano 1952: 281 ff.; engl. trans.: 174 ff.). His argument takes the form of a reductio ad absurdum and runs as follows: 1

2

As an empiricist, Brentano requires his determinism as well as his critique of indeterminism to be empirically grounded. If determinism holds that the will is causally determined, then necessary causal relations must be somehow accessible to experience. Let us first consider causality. Brentano agrees with Hume that outer experience does not make us directly acquainted with anything such as causation, and hence that physical causality is no more than a relation of temporal antecedence between events. However, unlike Hume, he claims that we do experience causal relations in inner perception. Brentano cites inference and will as examples of such relations (Brentano 1952: 283; engl. trans.: 175). Your act of drawing a conclusion based on premises is caused by previous acts of judgment; your act of willing something as a means is caused by your act of willing the end it is a means to. Inner perception teaches us that these causal relations are such that if a process A causes another process B, then B is made necessary by A. A conclusion necessarily follows 9

Denis Seron

3

4

from given premises; a volition necessarily derives from other volitions. That is why both are amenable to a priori treatment: volitions are subject to genetic laws that are the “main psychological foundations of ethics”; acts of reasoning are subject to genetic laws that provide the “principal psychological foundations of logic” (Brentano 1925: 67–68, 109; engl. trans.: 224, 253). Since it is conceptually true that the effect must take place when the cause occurs, the concept of a non-necessitating cause is self-contradictory. Therefore, “moderate indeterminism contradicts itself when it says that motives are causes that can bring about acts of will, but need not necessarily do so” (Brentano 1952: 283; engl. trans.: 175). Accordingly, the only option left to the indeterminist is to endorse the stronger view that the will has no cause whatsoever, namely extreme indeterminism. But Brentano claims to demonstrate that this latter view, too, involves contradiction. More generally, Leibniz’s principle of sufficient reason – the “law of causality” to the effect that it is impossible for something to exist out of nothing – is an analytic truth. Brentano’s demonstration proceeds (very roughly) as follows (Brentano 1952: 287 ff.; engl. trans. 177 ff., 1970: 137 ff.): (3a) “Something is” is necessarily equivalent to “something is present.” (3b) Necessarily, everything that is (present) is subject to continuous temporal changes. In Brentano’s terms, this means that everything that is must be a beginning, an end, or an inner border of a time lapse, and that between beginning and end there must always be some interval, however small: nothing can abruptly begin and end at the same time. (3c) Brentano then asks us to imagine a white dot on a black table, and to suppose that it is contingent absolutely speaking. Then, there is no necessity for the dot to exist, and it is even much more likely to cease to exist, a moment later. But if so, the dot is both beginning and end at the same time – which is assumed to be impossible. Therefore, the concept of an “absolute accident” and hence extreme indeterminism are self-contradictory.7 Since both extreme and moderate indeterminism are necessarily false and are the only possible forms of indeterminism, determinism is necessarily true.

Obviously, the argument is controversial in many respects. First, even supposing that, as claimed in step (2), causality involves necessity, one may object, following Thomas Reid, that motives are not efficient causes and that a motive, by its nature, does not make the motivated action or volition necessary, with the consequence that the agent can be regarded as morally responsible even though determinism is false and her actions and volitions are contingent. Brentano, however, does not take the objection seriously and considers the motive-cause distinction as a sophism: To be sure, Thomas Reid thought we ought to distinguish between motives and causes, for motives can move us to act, but cannot themselves act. But was this anything more than a mere sophism? If motives impel us to action, they are clearly working in conjunction with other factors; that is, determining along with other factors. Hence, as motives, they are operative causes of the will. (Brentano 1952: 262; engl. trans.: 163–164) Second, all of the first three steps are supposed to be intuitively evident on the basis of inner perception. This entails, for Brentano, that they are necessary in virtue of the empirical concepts being used. For example, Brentano views the “law of causality” as an analytic judgment, that is, as deriving from the analysis of concepts abstracted from experience. It is all but clear, however, why inner perception supports the law of causality, or the view that causality involves necessity, and not the indeterminist’s alternative possibilities view. 10

Franz Brentano

An ethics of Bildung An objection Brentano considers crucial is that the determinist runs the risk of falling into a kind of fatalism. What Brentano calls “fatalism” is either the view that actions are entirely independent of the will or the view that actions are dependent on the will but that the will itself is determined by the character conceived of as a set of inclinations that are given once for all. Brentano attributes the former – “pure or Asiatic fatalism” – to Oedipus and the latter – “modified fatalism” – to Schopenhauer (Brentano 1952: 261; engl. trans.: 162–163). The problem with fatalism is that it absolves us from any moral responsibility for our actions and thus “throttles all effort and ambition” (Brentano 1952: 245; engl. trans. 153): the will either has no control over actions or is innately determined to will this or that. Since modified fatalism involves the view that the will is necessarily determined, it is tempting to think that the same problem arises for determinism as well. But Brentano claims that the objection does not hold up to scrutiny, or at least that it does not compel us to embrace indeterminism. In fact, Brentano’s variety of determinism differs not only from pure fatalism by maintaining that actions are dependent on the will but also from modified fatalism in its denial that the character is entirely innate and out of the will’s control. In opposition to both pure and modified fatalism, Brentano claims that “not only our behavior but also our character is partially dependent upon our will” (Brentano 1952: 261; engl. trans.: 163). To recapitulate: pure fatalism is the view that the will does not determine action; modified fatalism holds that the will determines action, that it is determined by the character, and that the character is not determined by the will; Brentano’s determinism holds that the will determines action, that it is determined by the character, and that the character is determined by the will. Brentano is quite emphatic on this latter point, which he takes to be decisive. In his estimation, the idea that the character is determined by the will is a knock-down argument not only against fatalism, but against indeterminism as well. Indeed, by denying that inclinations causally determine the will, the indeterminist, too, must hold that we have no control over our will and thus that every effort to improve it is useless. Whether determined by immutable inclinations or not determined at all, our will remains something we have no causal influence over. Thus, the agent cannot be held as responsible for her action; she is “not the actor, but merely the stage upon which the act takes place” (Brentano 1952: 280; engl. trans.: 174). In this sense, indeterminism “is the doctrine that the will is not free” (Brentano 1952: 280; engl. trans.: 173–174). By contrast, Brentano’s view is that, since our will is causally determined by our inclinations, we do possess the power to improve it by improving our inclinations, i.e. our character. This leads him to propose what we may call an ethics of Bildung. Determinism, as Brentano sees it, holds not only that the will is determined by motives but also that “it is the necessary result of the circumstances at hand and of our intrinsic disposition, which is constructed primarily by previous exercises of the will” (Brentano 1952: 265; engl. trans.: 165). The underlying idea is this: the will has some influence on the motives that causally determine it; therefore it can indirectly act upon itself through exercise and self-discipline. Taking up Aristotle’s concept of habit as a “second nature” (Brentano 1952: 261; engl. trans.: 163), Brentano claims that it is in our power “to work on our character by forming (Bildung) good habits or by improving our natural inclinations” (Brentano 1952: 265; engl. trans.: 165). Of course, such effort towards self-improvement would make no sense if determinism were false, that is, if our character and inclinations did not somehow determine our will and, hence, our actions. This entails that education as well as all forms of self-improvement through practice must play a prominent role in moral life as Brentano conceives it. For example, asceticism should 11

Denis Seron

not be practiced as an end in itself, but only as a means of learning moderation (Brentano 1952: 384; engl. trans.: 234; also 237–238; engl. trans.: 147).8 Likewise, Brentano makes a plea for “good manners” as a means to self-improvement: Good manners can also be regarded as a practice ground; we ought to maintain them even where neglecting them would give no offence – even, indeed, when we are alone. Bear in mind the story of the Englishman who always changed for dinner, even though he was the only white man living among the natives. We ought to take care about our clothing even at home, for such matters should become second nature. (Brentano 1952: 384–385; engl. trans.: 234–235) Finally, education is given a central place in Brentano’s account. Indeed, a modified fatalist may object that Brentano’s argument is somehow circular: the effort towards selfimprovement itself needs to be willed and thus motivated accordingly, with the consequence that you cannot make such effort unless you already have good inclinations. The objection falls away, however, if good inclinations can be acquired through education. As Brentano puts it, one’s will can influence not only one’s own thoughts and affects but also “the powers of other people” (Brentano 1952: 236; engl. trans.: 147).

Concluding remarks To conclude with, Brentano presents his determinism as fully compatible with both the law of causality and the view that we are morally perfectible and responsible for our actions. In fact, his claim is even stronger, since he maintains that determinism is the only way to make sense of our moral perfectibility and responsibility. Brentano proposes an interesting variant of virtue ethics based on the idea of self-improvement. Opposing the view that moral life consists in resisting inclinations that would otherwise cause the agent to act badly, he asks us to conceive of the will as being necessarily determined by inclinations and having to strive actively to improve them through self-discipline.

Related topics Chapters 23 (Strawson), 27 (Kriegel).

Notes 1 Cf. Brentano (1925: 115; engl. trans.: 257; Montague 2017: 114): “Every volition or striving in the strict sense refers to an action. It is not simply a desire for something to happen but a desire for something to happen as a result of the desire itself. An act of will is impossible for someone who does not yet know, or at least suspect, that certain phenomena of love and desire directly or indirectly bring about the loved object.” Brentano’s theory of will is well-documented in the literature. See, among recent works, Montague (2017), Kriegel (2017), and Kriegel (2018: Ch. 7). By contrast, there has been very little research on Brentano’s critique of free will. The only study I am aware of is Modenato (1981). 2 The distinction between “descriptive” and “genetic” (or “physiological”) psychology is central to Brentano’s epistemology. The latter states laws that apply to the “succession of mental phenomena.” The former’s task is to clarify psychological concepts. See Brentano (1982) and my comments in Seron (2017). 3 This distinction is drawn from Aquinas. See Summa Theologiae, Part II-1, Quest. 6, Art. 4, and the helpful comments in McInerny (1992: 20 ff.).

12

Franz Brentano

13

Denis Seron transl. by P. Remnant and J. Bennett (1996): New Essays on Human Understanding, Cambridge: Cambridge University Press. McInerny, R. (1992) Aquinas on Human Action: A Theory of Practice, Washington, DC: The Catholic University of America Press. Modenato, F. (1981) “Libertà e necessità in Franz Brentano,” Verifiche: Rivista Trimestrale di Scienze Umane 10/4, 437–453. Montague, M. (2017) “Brentano on Emotion and the Will,” in U. Kriegel (ed.) The Routledge Handbook of Franz Brentano and the Brentano School, New York: Routledge, pp. 110–123. Nathan, N.M.L. (1971) “Brentano’s Necessitarianism,” Ratio 13/1, 44–55. Seron, D. (2017) “Brentano’s Project of Descriptive Psychology,” in U. Kriegel (ed.) The Routledge Handbook of Franz Brentano and the Brentano School, New York: Routledge, pp. 35–40.

Further reading U. Kriegel (ed.), The Routledge Handbook of Franz Brentano and the Brentano School (New York: Routledge, 2017) is a fine collection of essays devoted to all aspects of Brentano’s thought, including his ethics and his theory of the will (chapters 11, 20, and 21). Although somewhat out of date due to recent research advances, D. Jacquette (ed.), The Cambridge Companion to Brentano (Cambridge: Cambridge University Press, 2004) is still useful. Written by a leading figure in the field, U. Kriegel, Brentano’s Philosophical System: Mind, Being, Value (Oxford: Oxford University Press, 2018) offers an excellent and in-depth overview of Brentano’s philosophy as a whole (on his theory of the will, see Chapter 7).

14

2 PHENOMENOLOGY OF WILLING IN PFÄNDER AND HUSSERL Karl Mertens

“We are human beings, freely willing subjects who are actively engaged in our surrounding world, who continuously have a part in shaping it” (Husserl 1989a: 4, translated in Bernet et al. 2005: 157; cf. Husserl 1989a: 24). This is how Husserl, in his first article for the Japanese journal The Kaizo, published in 1923, describes human action. Speaking of ourselves as “freely willing subjects” and as “actively engaged in our surrounding world,” Husserl concurs with a traditional view of action, arguing that it is our free will that enables us to actively influence the course of the world. In other words actions are constituted by free will. In the following, I will mainly focus on reflections elaborated by Husserl in 1914 in a section entitled On the Phenomenology of Will, which he added to his lecture course on Basic Questions of Ethics and Value Theory held in 1908/09 and 1911 (Husserl 1988: 102–125). A major impetus for this section can be traced back to Husserl’s intense interest in Alexander Pfänder’s article Motives and Motivation, which was published in 1911. Husserl carefully studied and annotated Pfänder’s article (cf. Schuhmann 1973: 94ff.).1 Therefore, in the first part, I begin by sketching the decisive conceptual distinctions that allow us to understand both the similarities and the differences between Husserl’s and Pfänder’s positions. In the second part, I address a controversial aspect of Pfänder’s analysis: his assumption that willing “includes the immediate consciousness of self ” (Pfänder 1963a: 135, 1967: 22). The third part outlines and critically discusses Husserl’s concept of action as realization of the will as developed in his 1914 considerations. In this context, I pinpoint what I consider to be the relevant shortcomings in Husserl’s analysis of willing and acting by discussing some aspects of our experience of action which he seems to neglect and which should instead be integrated into a proper phenomenological account of action.

Theory of action as a theory of willing: Husserl’s reading of Pfänder Both Pfänder and Husserl analyze the sphere of willing in relation to its temporal structure. Recognizing some crucial and structural temporal differences, they operate distinctions within this sphere. Thus, in contradistinction to the willing directed toward a future action, which Husserl calls “resolution-will” (Entschlusswille) or “purpose-will” (Vorsatzwille), Husserl refers to the willing that is constitutive of the execution of a current action as 15

Karl Mertens

“action-will” (Handlungswille) (Husserl 1988: §§ 15f., pp. 106ff.). Regarding the willing that temporally precedes the action, “purpose” (Vorsatz) and “resolution” (Entschluss) describe different relations between the action-will and the willing that actually precedes the constitutive willing of an action. As a “purpose,” the willing appears as a sudden act referring to a future action. However, if the willing is preceded by a process of reflection, the so-called “deliberation of the will” (Willenserwägung), then the willing can be characterized as a “resolution” (Entschluss) directed toward a future action or, if conflicting possibilities are rejected, as a “decision” (Entscheidung) (cf. Husserl 1988: 111). A somewhat similar distinction, albeit formulated in different terms, can be found in Pfänder, who distinguishes “action of willing” (Willenshandlung) and “act of willing” (Willensakt). The act of willing, also described as the “act of volitional proposing,” is directed toward a future action. The action of willing, the “volitional execution of what is willed” (Pfänder 1963a: 126, 1967: 15), instead indicates for Pfänder a willing that cannot be separated from the execution of the action.2 Whereas action, according to Husserl, necessarily implies an action-will (which more or less corresponds to Pfänder’s action of willing), Husserl’s purpose-will or resolution-will don’t necessarily have to precede the action. In contrast, Pfänder’s act of willing precedes the action of willing (cf. Pfänder 1963a: 134, 1967: 21), even though the action of willing possibly may follow “immediately, without hesitation, upon the act of willing” (cf. Pfänder 1963a: 126, 1967: 15). Therefore, we can say that, according to Pfänder, an action of willing is based on an act of willing. Besides, a purpose, resolution, or act of willing do not need to be performed at all.3 Moreover, at least in Husserl’s phenomenology of willing, a form of willing that is a constitutive condition of action is differentiated from a form of willing that is not a constitutive condition of action. It is for this reason, according to Husserl, that also the willing directed toward a future action can only be carried out by the decisive action-will, directly linked to the action. Therefore, in terms of a theory of action, the determination of the action-will is the essential point, since every action is constituted as an action through the present realization of an action-will. By contrast, Pfänder primarily focuses on the determination of the psychic sphere of the willing. In this context, Pfänder emphasizes the character in virtue of which willing is fundamentally separated from “mere strivings and counterstrivings” (Pfänder 1963a: 133, 1967: 20). Accordingly, we can distinguish a phenomenological approach to willing as an autonomous phenomenon of consciousness (in Pfänder) and a phenomenological approach to willing as constituens of human action (in Husserl). These different interests in the sphere of willing can also be related to a remarkable terminological divergence: both Pfänder and Husserl are using the term “act of willing” (Willensakt). However, while Pfänder in his fundamental distinction of acts of willing (Willensakte) and actions of willing (Willenshandlungen) classifies the willing itself as a form of action, Husserl only makes use of the term “act of willing” (Willensakt) within the overall terminological classification of the sphere of acts he is interested in; “act of willing” covers purpose, resolution and action-will (cf. Husserl 1988: 103). Compared with this rather artificial use, when it comes to an explication of the aspects of willing that are constitutive of action, Husserl does not terminologically refer to the willing itself as an act or an action. In fact, willing seems to characterize the action as an action in the first place. Husserl concisely expresses this by terminologically referring to the willing constitutive of action as action-will (Handlungswille). Apart from such differences in the overall conception of Husserl’s and Pfänder’s phenomenology of willing both Husserl’s and Pfänder’s elaborations are of importance for a phenomenological theory of action. Indeed, both investigate the structures of intentionality 16

Alexander Pfänder and Edmund Husserl

which are constitutive for actions, and discussing what is distinctive for both views and possible shortcomings in the overall approach may help to generally improve the phenomenological account of action.

The volitive experience – a problem in Pfänder’s analysis From a methodological point of view, Pfänder and Husserl consider willing to be a phenomenon of subjective experience accessible to descriptive analysis, which is the task of phenomenological reflection. Yet, in turning to Pfänder’s studies on acts of willing, one encounters at first glance a highly problematic description of the phenomenon of willing. According to Pfänder, willing “includes the immediate consciousness of self ” (Pfänder 1963a: 135, 1967: 22) – i.e. willing means being aware of oneself as the one who is willing, grasping one’s own ego and turning it into the subject of the action that follows the will. Precisely because the ego is the subject of a willful action, and thus because the action relevantly depends on the ego, willing should be distinguished from striving. This is expressed by Pfänder in the following passage: The act of willing (Willensakt) refers to one’s own ego. If . . . it is to be a genuine act of willing, then one’s own ego must not be merely thought about but must itself be immediately grasped and must be made a subject referent of a practical proposing. Thus willing, but not striving, includes the immediate consciousness of self. (Pfänder, 1963a: 135, 1967: 22) This definition of willing given by Pfänder seems to be far too ambitious, for willing doesn’t necessarily imply explicit self-consciousness. If, for instance, I want to get bread in the morning, I am instantly aware of the objective of my voluntary action (which is to get some bread). However, a rather exceptional context would be required for me to become somewhat aware – not to mention explicitly aware – of myself and my willing of the project. For instance, the family morning chaos leads me to forget about what I actually want (will), or rather, wanted (wollte). In that case, I could remind myself and recall what was actually willed. But even in such cases, I’m not necessarily conscious of myself. More specific situations will be required. On the way to the bakery, I might meet a friend and explain to her why I absolutely want to buy the bread at the organic bakery. In doing so, I may become aware of myself as someone who does not simply want to buy some bread, but at the same time wants to be ecologically aware. Self-awareness does not generally come along with volitive experience, but is manifest only in forms of more distinct higher-order willing.4 When Pfänder considers self-awareness to be an integral moment of willing, at best he can refer to very specific experiences, whose impact on a philosophical analysis of willing, i.e. an analysis interested in general structures of willing, should be questioned. However, Pfänder does indeed care about the determination of such general structures. This becomes obvious with regard to the systematic role that the recourse to self-awareness plays in Pfänder’s phenomenology of willing. For it is self-consciousness – as the quoted passage indicates (cf. Pfänder 1963a: 135, 1967: 22) – that provides the decisive criterion for the distinction between acts of willing (Willensakte) and strivings (Strebungen). However, with this determination of willing, Pfänder seems to disregard his own concept of a carefully descriptive, phenomenological-psychological analysis of the phenomena in question.5 The outlined criticism is backed up by Husserl’s own annotations of Pfänder’s article, where he writes: “Is it appropriate that in every actual ‘I will’, the I itself must be an object? 17

Karl Mertens

An actual ‘self-consciousness’ must be accomplished (as Pfänder states)? . . . I can will, just as I can judge without thematically ‘objectifying’ the I.”6 In line with this thought, Husserl points out that I become aware of myself in and through an act of reflection, explicitly reminding myself – even in a linguistic form – that it is me who wants (will) this or that. In this context, the self-consciousness of a willing person – in most cases retrospectively – becomes thematic and explicit in situations of communicative self-assurance as described earlier.7 That in willing the “I” is not objectified, according to Husserl, does not mean that willing is not experienced by an “I.” However, this is a common feature of every kind of act-consciousness and its modalities and does not uniquely characterize willing: The peculiarity of willing lies of course in its entanglement with the I (Ich-Hereingezogenheit). But first of all, this is a common feature of (gegenüber) all ‘acts’ in their concise meaning. . . . There are different ‘positionings’ (Setzungen) and ‘sentences’. And there are fundamental theses and modalities of fundamental theses everywhere.8 However, looking at Pfänder’s statement in the context of other remarks, one could come to the conclusion that Pfänder himself might even basically agree with the objection.9 The method he is aiming at in his Phenomenology of Willing is described as a “‘subjective’ method.” Hence, Pfänder’s analysis refers to the consciousness of the willing subject. But for Pfänder, this method does not amount to an introspective approach that requires the possibility of a “direct observation of the immediately experienced” (Pfänder 1963a: 7; own translation). The self-observation in question rather refers to a remembered consciousness. Consequently, instead of referring to an introspective approach, Pfänder rather refers to a “retrospective method.”10 In doing so, he remains consistent with Husserl’s concept (mainly explicated in the 1920s) of the generally retrospective character of reflection.11 In the light of such a retrospective self-attention, we can also understand the following observation made by Pfänder: “The act of willing . . . is an act of self-determination in the sense that the ego is both the subject and the object of the act” (Pfänder 1963a: 135, 1967: 23). This is not to imply that the willing subject, while performing her act of willing, is aware of the self-centered character of its willing, but only to show that it can retrospectively become aware of this reference to herself: In this context, ‘consciousness of willing’ is not to be understood as the knowledge of the fact that one now wants (will) something, but as the state of consciousness that is given when someone wants (will) something, irrespective of the fact whether s/he states or notes that s/he wants (will) this now.12 (Pfänder 1963a: 6; own translation) Pfänder even goes one step further. In conjunction with the earlier quoted passage taken from Motives and Motivation, he elaborates that willing is linguistically expressed in sentences with the structure “I will P” or “I will not P.”13 From this angle, light can be thrown on the term “subject referent” (Subjektsgegenstand), which, in the just outlined critique, has been interpreted as thematic self-consciousness. This, however, is a misunderstanding – at least if one consults Pfänder’s 1921 Logic. In this context, the term “subject referent” gains significance in connection with the analysis of the judgment (Urteil) and the respective state of affairs (Sachverhalt) (cf. Pfänder 1963b: 38ff.).14 With that in mind, the self-consciousness which is linked to willing can be taken as subject referent from an analysis of sentences about willing. Regarding the ontological state of affairs underlying the logical judgment expressed in 18

Alexander Pfänder and Edmund Husserl

a sentence like “I will P” or “I will not P,” the I has to be understood as subject referent which pronounces its willing. Therefore, regarding the reference to an explicit self-consciousness, the (earlier quoted) claim that “willing . . . includes the immediate consciousness of self ” (Pfänder 1963a: 135, 1967: 22) should be carefully interpreted. However, this does not change the crucial point: according to Pfänder, willing is essentially connected with the possibility of a practical and free self-determination of the willing person. And concerning this point, Husserl fully conforms with Pfänder as, for instance, becomes clear in the passage from Kaizo quoted at the beginning (Husserl 1989a: 4; cf. Husserl 1989a: 24; cf. Ubiali 2010: 245ff.). This consideration can be understood as phenomenological, as long as the linguistic reference is interpreted in a heuristic sense. In this sense, Pfänder makes some occasional remarks on the meaning of our common speech for phenomenological-psychological analysis, when he considers linguistic expressions as a heuristic starting point for the examination of consciousness. However, according to Pfänder, what is linguistically expressed must be exposed in the phenomenological analysis of consciousness (cf. e.g. Pfänder 1963a, 6, 140, 1967: 26). How we linguistically express our willing, therefore, reflects a phenomenological state of affairs which is supposed to be revealed as a fact of consciousness by a retrospective and linguistically directed analysis. For this to succeed, the memory and verbalization of an experience must, for its part, be measured by the remembered and the verbalized, that is to say, by the experience itself. This basically seems possible when we admit that we can even criticize the adequacy of a retrospective linguistic determination of our willing, if an adequate linguistic expression of our pre-linguistic experience is not or not yet available. This possibility shows that what is outlined by means of memory and verbalization cannot simply be identified with our pre-reflexive experience. Retrospective attention, reflection, and linguistic expression are not sufficient when we want to explore our actions and willing. In that respect, Pfänder’s recourse to self-consciousness is under-determined from the phenomenological point of view; it only designates a place holder for a more precise determination of the experience of willing. Conversely, Pfänder’s methodological reflections point to the difficult claim of a purely descriptive phenomenological analysis, particularly when willing is concerned. Apparently, what is elaborated in phenomenological reflection is not merely a description of our experience. Rather, phenomenological reflection establishes something, creates its object, especially if – which is often the case with phenomenologists – it is emphatically understood as distinguished, original, or even actual experience. Husserl’s elaborations on the constructive nature of the analysis of essences or ideas (cf. e.g. Husserl 1959: 218, 245f., 456, 504, 1962: 282, 1970: 304) and particularly Merleau-Ponty’s observations on the phenomenological reflection as a “truly creative act (une véritable création)” (Merleau-Ponty 1945: iv, 2005: xi) in the preface of Phenomenology of Perception clearly outline this. But nevertheless these approaches are of a phenomenological nature insofar as they enable us to better understand the phenomenal world of our experience. In the following, I will critically discuss the interrelation between a phenomenological-productive reflection and its phenomenal adequacy with a view to Husserl’s elaborations on the phenomenology of willing in more detail.

Husserl’s concept of action as realization of the willed and the question of its phenomenal adequacy In a brief and concise examination of the difference between willing and wishing, which can be found in his analysis of the sphere of willing from 1914, Husserl claims that the realization of the willed is crucial when studying action from a philosophical point of view:15 According 19

Karl Mertens

to Husserl, wishing cannot be defined as a practical act, or as willing in the broadest sense (Husserl 1988: 103). The wished is known not to be feasible in practical terms. Conversely, the willed is known to be practically realizable in some way – whether it be certainly, probably, possibly, etc. The concept of realizability or feasibility can be specified here if we assume the following: first, the willed is to be understood as being feasible for me, the agent; willing does not aim for the willed to merely somehow happen, but to happen because of me. Second, it is not a question of the actual feasibility of the willed, but rather of me, the acting person, becoming aware of an imagined possibility that I can actualize, i.e. I (the agent) have to be aware of the willed as feasible (cf. Husserl 1988: 104). Since I can actually wish for something without willing to realize it, even if I am aware of the fact that I could (actually) realize it, the awareness of the feasibility can only be a necessary, but not a sufficient condition for the willing.16 Therefore, third, when differentiating between willing and wishing, the meaning of I am willing to realize something must be explained. However, this kind of phrasing, in its almost tautological form, does not say much. It can rather be interpreted as a kind of blank space, to be filled by a more precise analysis of willing. And fourth, willing and wishing, in the actual act of willing, are intertwined in many ways; their differentiation is merely of heuristic and explanatory importance. This is not to imply that wishing and willing can be distinguished in every case.17 In his reflections on the meaning of the fiat Husserl attempts to determine more precisely the characteristic of the willing that actualizes, mentioned as third point, and with a certain emphasis.18 He writes: Consciousness in a way does not say: ‘It will be, and therefore, I want it’ (‘Es wird sein, und demgemäß will ich es’); but: ‘Because I want it, it will be.’ (‘Weil ich es will, wird es sein.’) In other words, the will pronounces its creative ‘let there be!’ (‘Es werde!’). The establishment of the will equals the establishment of the realization. But realization here does not only stand for something becoming real, but making something real, the achievement of realization. But this is something inherently foundational (Ureigenes), and its origin lies in the peculiarity of the consciousness of willing, and can only be understood from there.19 (Husserl 1988: 107; own translation) If one wants to know how paradigmatic examples of this type of realization look like, one can think of cases like the following: a person wants to do something and does it. For instance, I want to go to the movie tonight, and then I go to the movie; I want to call someone and then do so. What is to be realized in the near or distant future, toward which the will of a person is directed, is given as the goal of the realization previous to acting. With respect to this goal, the action as realization of the willed can be judged as successful or unsuccessful. Trivially, these examples also illustrate that the realization of every willing directed toward the future, i.e. in Husserl’s notion a purpose or resolution, can be basically jeopardized. Suppose I want to go to the movie, but meet a good friend in the foyer that I haven’t seen for a long time. We decide to have a beer together. This shows that the initial willing can be modified or even revised between the moment we plan and the realization of this plan. According to the interpretation outlined earlier, Husserl agrees. Nevertheless, for Husserl, every action is to be understood as the realization of a willing, more precisely, of an action-will, which is the constitutive moment of the action itself. The decisive realization which is constitutive of the action is therefore to be understood as the immediate realization of the willing. Wherever a modification or revision of the resolution-will or purpose-will is noticeable, we should, for 20

Alexander Pfänder and Edmund Husserl

logical reasons, bring into play an appropriate (new) action-will. For instance, now, I don’t want to go to the movie anymore, but go for a beer with my friend. Since in this case the model of a willing directed toward future action is applied to the case of the current willing, the problem of the explanation of revising and modifying the action shifts to the current voluntary action itself.20 The relation between action-will and the current action (constituted by the action-will) thus turns into a microscopic copy of the relation between resolution- or purpose-will and future action. It is to be assumed that the action-will that underlies the action takes a temporal-processual course.21 Therefore, Husserl considers the willing of action as a perpetual creative source which from phase to phase, from moment to moment, creates the willed as a current reality. Here, a now is to be marked as the beginning of the creative production and another now is to be marked as the end of the action. Furthermore, by considering the temporal horizons of the past and future of the action as voluntary horizons, i.e. as horizons of the creative production, Husserl is able to describe the realization of the action-will during acting by means of the constant metamorphosis of these horizons. With the initial phase, the future horizon of the action is given as the “creative future” of the yet to be realized. Since every current now of the action passes over into an ever new now, the horizon of the creative production changes. The “creative production” becomes a “creative present,” something actually created, and finally turns into a “creative past.” By completing the action, the finished action as a whole sinks into ever more pasts. Husserl expresses this specific dynamic of the creative activity passing from one present to another by means of his terminology of intention and fulfillment which he developed in relation to the concept of evidence. In every phase, there is an intention of the will directed toward the creative content of the next phase; at the same time, the former intentions are being fulfilled (cf. Husserl 1988: 111). In this case, the term “intention” (Intention) does not only refer to the particular moment of the yet unfulfilled state of being directed toward (something) as it can be found in every intentional act, but it also denotes, as the literal sense of “intention” implies, a voluntary, intentional state of being directed toward the realization of the willed. In this context, the term of fulfillment corresponding to the term of intention designates the active realization of the intention of the will. The individual intentions of the will are being kept together by the continuous unity of the willed action. In this sense, Husserl states: In every now, the direction of the will and the creative ‘Let there be!’ pass through the continuity of the moments of the will; with every new current moment of creation, a prior intention of the will, directed toward its content, is being fulfilled. (Husserl 1988: 111; own translation) This description of the realization of the action-will in the performance of the action undoubtedly seems quite artificial. This could be the case because Husserl is interested in a description of typical characteristics of voluntary action comparable to Weber’s ideal type. However, according to the previously outlined concept of phenomenological reflection, we can only judge the phenomenological acceptability or even merely the plausibility of such a construction if we ask ourselves in what way such a theoretical analysis of the will contributes to a better understanding of voluntary action. Well, it seems that there are phenomena of action which can be well understood by way of Husserl’s analysis. In order to illustrate this, I’d like to take up an example Searle uses in his book Intentionality. Here, Searle describes the action of a skiing beginner who (intentionally) takes a left turn as follows22: “For a beginner to make a left turn, he must put his weight on the downhill ski while edging it into the slope, 21

Karl Mertens

stem the uphill ski, then shift the weight from left to right ski, etc., all of which are reports of the content of his intentions in action” (Searle 1983: 100). When describing this action by the means Husserl’s phenomenology of willing provides, one will have to distinguish different phases of willing, which are directed toward the individual phases of action. The skier, while taking the left turn, first wants to shift his weight to the downhill ski, and while doing so, his intention of will is already directed toward the execution of the next phase, edging the downhill ski. Moreover, he will be anticipating the further phases of stemming the uphill ski and the shifting of weight. At the moment when his action-will focuses on edging the downhill ski, his intention of will already focuses on stemming the uphill ski, while shifting the weight on the downhill ski is retentionally present to him as a past, voluntarily produced action, etc. The acting of a beginner who is learning a challenging motion or technique provides a good example for the detailed analysis of willing, going from phase of action to phase of action. Whoever wants to learn a complicated movement such as skiing will oftentimes be explicitly encouraged to structure her action in phases and to carry it out step by step, in a very conscious way. The previously outlined continuum of current phases of willing seems to correspond to the conscious experience of a skiing beginner.23 However, the consciousness of an untrained individual might just give an example for a lack of clarity and sovereignty in the voluntary action. In the way that the current phase of willing pushes to the fore, the proceeding willing of the next and subsequent action phases vanishes. The skiing beginner, in his/her acting, again and again will be fully oriented by the upcoming partial action. So, while edging the downhill ski, all of his/her effort may be focused on stemming the uphill ski. Especially, if this doesn’t succeed due to his/her insufficient skills s/he may hardly be aware of the proceeding willing which is directed toward the next motion sequences, aiming at finishing the left turn. Given the focus on willing about what is to come next, the skiing beginner will have to revise his/her original willing again and again. The nature of the external conditions as well as his/her insufficient skills will then undermine the sovereign performance of the originally willed action and the willing will be limited to getting back in control of the action. In view of such cases, Husserl’s description of a willing which takes into account several phases of action simply appears to be cognitively overload. What someone intends to do in essential parts depends on the situation and the skills of an agent. Thus, despite the appearance, the action of the beginner, whose willing must be characterized as momentarily due to insufficient control over his/her own action, cannot be considered a suitable example for Husserl’s analysis of willing. But even looking at the confident willing of a skilled skier doesn’t provide a better example for Husserl’s analysis. Compared with the narrowness of the willing in the case of the beginner, it is now the vastness and complexity of the willing that cannot be covered by Husserl’s analysis. Hence it is unlikely that the willing of the experienced individual is directed toward the step-by-step execution of a left turn, but instead, for example, toward going down the slope by making some elegant turns on the edge. But it is not just a matter of wider phases of action that are characteristic here. The willing of the experienced individual is rather characterized by a fundamental complexity and openness. For instance, the experienced skier, while navigating down the slope, positions himself/herself in front of a less experienced skier and tries to lead him/her down the slope quickly in order to catch the last lift with him/her. At this point, multiple desires conjoin and cannot be separated to gradually pass through individual phases of willing. Occasionally, the experienced skier becomes aware of current phases of action and phases of willing, often when special incidents

22

Alexander Pfänder and Edmund Husserl

require his/her attention and, if necessary, force him/her to modify or even revise his/her willing and acting. Thus, as opposed to the case of a beginner, the confident willing of the experienced individual doesn’t turn out to be the execution of a will as a whole passing through different phases, but a complex and basically open form of willing. The complexity and dynamics of such willing and our practical experience cannot be covered sufficiently by the means of Husserl’s meticulous analysis of the willing constitutive of the action. According to Husserl, the action-will which is constitutive of the action is clearly defined with respect to what is willed. The realization of the willing is to be understood as the practical execution of the willed, in relation to which the action is determined as fulfillment of an intention of the will. However, this analysis becomes problematic wherever in the course of the action aspects arise that lead to an understanding of the executed action which characterizes it not just as a mere realization of the willed. If one attempts to understand cases of specification, modification, and revision of a will using the means provided by Husserl’s analysis, there is always a critical moment when the former intention of will is being replaced by a new one. Every modification of the will necessarily has the character of an abrupt change, of a new initiative of the will, which cannot be incorporated into the consistent course of action together with the former willing. These difficulties ultimately result from Husserl’s neglect of the constitutive significance of situational conditions of acting, which can interfere with initiated action, accentuate it, redefine it or even provide the opportunity for initiatives of new action. One could object against this interpretation that, according to Husserl, willing as a founded act is dependent on the perception of a situation (cf. Husserl 1988; 109). This is right. And therefore, it is possible to understand by means of Husserl’s analysis how the perception of a new situation may influence the willing.24 However, taking recourse on the structure of intention and fulfillment, an action must in every case be well determined by a corresponding willing or, more precisely, by the “fiat” which is defined by an intention, the willed content, going to be fulfilled in advance. As a consequence of this conceptual frame, every development of willing must be traced back to a concretized, modified, or even revised “fiat” induced by the perception of the situation, providing a new intention going to be fulfilled in advance. However, on the basis of Husserl’s analysis, we cannot understand an action lacking a well-defined willing. At best, we understand it as an action merely referring to the intention of a superordinate willing.25 And if a defined willing is modified or changed, Husserl’s analysis is even obliged to introduce a new will constituting a new action. In short following this line, there is no analytic possibility to understand the unity of an action which in its process is modified or even revised.26 The microscopic splitting of the unity of acting in always new moments and phases of willing gets into trouble wherever we modify our willing and acting due to what happens while performing a voluntary action. In this respect, events, bodily behavior, and situations befalling the agent play a constitutive role for the willing of agents, too. Therefore, Husserl’s emphatic understanding of human agency by taking recourse on “freely willing subjects” (Husserl 1989a: 4, translated in Bernet et al. 2005: 157) at least must be completed by taking into account the constitutive meaning of aspects of passivity and situatedness which are involved in the constitutive process of shaping an agent’s willing. Willing subjects are essentially also affected from what is not under their control.27

Related topics Chapters 3 (on Pfänder), 17 (on Dreyfus), 20 (Smith), 25 (Drummond).

23

Karl Mertens

Notes

24

Alexander Pfänder and Edmund Husserl or more distant memories. Hence this ‘introspective’ method or method of ‘inner observation’, as it has been erroneously called, is not actually ‘introspective’, but mostly ‘retrospective’. (Pfänder 1963a: 7; own translation) 11 Cf. especially the analyses of phenomenological reflection in Husserl (1959: 86ff.). 12 Cf. Pfänder 1963a: 16f.; own translation: This feeling of the self (Ichgefühl) . . . at the same time always is a feeling of relatedness to something. The feeling of the self and the “objective” content are always given at the same time, and in the course of this, the feeling of the self already involves the relation of the self to the content. Self (Ich), relation of the self to a content and “objective” content are only the separated, distinguishable moments of the state of one consciousness. For the state of consciousness of the existence of a content can be considered from two sides, the side of the self and the “objective” side. . . . It is for the fear of the conscious-self (Bewußtseins-Ich) (that is associated with something which is of metaphysical character, and therefore despised) that some psychologists, without being aware, rather use the term of consciousness instead of the term of self. 13 Pfänder 1963a: 134; own translation (slightly deviating from the English translation in Pfänder (1967: 21)). 14 Cf. Spiegelberg in his annotation 13 in Pfänder (1967: 22), who, in relation to the earlier quoted passage (Pfänder 1963a: 135, 1967: 22), points to Pfänder’s distinction “between subject-concept (Subjektsbegriff ) and predicate-concept (Prädikatsbegriff ), as constituents of the logical proposition (Urteil),” on the one hand and between “subject-referent (Subjektsgegenstand) and predicate-referent (Prädikatsgegenstand), as constituents of the ontological state of affairs (Sachverhalt) projected by the proposition,” on the other hand. 15 Cf. for the following Mertens (2010: 469ff.); cf. also Nenon (1990: 302ff.). 16 Husserl seems to point to this indirectly when he writes (Husserl 1988: 104; own translation, italics from me): “Anything which is possible can be wished for, but not just anything which is practically possible.” This implies: what is practically possible can be wished for as well. Of course, apart from this, we can also wish for things that are not practically feasible. 17 In fact, simple examples will illustrate a difference that allows us to explain what it means that our willing, as a rule, is related to wishing as well. According to Husserl, willing includes wishing as “reasonable implication” (Implikation im Sinne der Vernunft), i.e. willing something without wishing for it would be unreasonable, if it is not willed for the sake of something else we wish for. In this sense wishing can be considered as founding the act of willing (cf. Vargas Bejarano 2006: 123). However, it is not a “real implication of the wishing within the willing” (reelle Implikation eines Wünschens im Wollen). For I can will something that I don’t wish for. In the first case, it is about the desirability of the willed (i.e. that which is willed is desirable); the second case refers to the question whether a current wishing is actually required for the willing. De facto, this does not have to be the case (Husserl 1988: 105; cf. Vargas Bejarano 2006: 124, who emphasizes the difference between willing and wishing). 18 Husserl adopts the notion of “fiat” from William James (Melle 1992: 287f., 1997: 176f.). It seems that the function of Husserl’s “fiat” roughly corresponds to Pfänder’s concept of “initiating impulses” (Willensimpulse) which is only indicated in Motives and Motivation (cf. Pfänder 1963a: 126, 1967: 15f.). 19 Husserl’s elaborations match Searle’s distinctions of directions, more precisely, directions of fit as well as directions of causation of visual perception and intentional action. Searle characterizes visual perception by its “mind-to-world” direction as well as the reverse direction of causation, inasmuch as mental states in perception must comply with the world, because mental states are being caused by things and events that happen in the world. Conversely, he characterizes intentional action by the direction “world-to-mind” and the direction of causation “mind-to-world” (Searle 1983: 97). 20 Cf. Husserl (1988: 107f.; own translation): The creative “Let there be!” ( “Es werde!”), which is essential to the establishment of the will (Willenssetzung), can be (something) currently creating – the will is action-will, executive will, actually creating –, or, it can only be directed toward creating, a future creation. 21 Cf. for the following especially Husserl (1988: 110). 22 I think it is worth mentioning that Searle’s description applies to a historical state of ski technique (approximately of the early 1980s); with the introduction of Carver-Ski, Searle’s description would have to be modified considerably.

25

Karl Mertens

26

Alexander Pfänder and Edmund Husserl Husserl, E. (1988) Vorlesungen über Ethik und Wertlehre 1908–1914, hg. v. U. Melle (Husserliana XXVIII), Dordrecht, Boston, MA, London: Kluwer Academic Publishers. Husserl, E. (1989a) Aufsätze und Vorträge (1922–1937). Mit erg. Texten hg. v. T. Nenon und H.R. Sepp (Husserliana XXVII), Dordrecht, Boston, MA, London: Kluwer Academic Publishers. Husserl, E. (1989b) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Second Book: Studies in the Phenomenology of Constitution, translated by R. Rojcewicz and A. Schuwer, Dordrecht: Kluwer. Husserl, E. (2002) Einleitung in die Philosophie. Vorlesungen 1922/23, hg. v. B. Goossens (Husserliana XXXV), Dordrecht, Boston, MA, London: Kluwer Academic Publishers. Husserl, E. (2004) Einleitung in die Ethik. Vorlesungen Sommersemester 1920/1924, hg. v. H. Peucker (Husserliana XXXVII), Dordrecht: Springer Science + Buisiness Media. Husserl, E. (2012) Einleitung in die Philosophie. Vorlesungen 1916–1920, hg. v. H. Jacobs (Husserliana Materialien IX), Dordrecht, Heidelberg, New York, London: Springer. McDowell, J. (2011) “Some Remarks on Intention in Action,” The Amherst Lecture in Philosophy 6, 1–18. Melle, U. (1992) “Husserls Phänomenologie des Willens,” Tijdschrift voor Filosofie 54, 280–305. Melle, U. (1997) “Husserl’s Phenomenology of Willing,” in J. Hart and L. Embree (eds.), Phenomenology of Values and Valuing, Dordrecht: Kluwer Academic Publishers, pp. 169–192. Melle, U. (2015) “‘Studien zur Struktur des Bewusstseins’: Husserls Beitrag zu einer phänomenologischen Psychologie,” in M. Ubiali and M. Wehrle (eds.), Feeling and Value, Willing and Action. Essays in the Context of a Phenomenological Psychology, Heidelberg, New York, Dordrecht, London: Springer, pp. 3–11. Merleau-Ponty, M. (1945) Phénoménologie de la perception, Paris: Éditions Gallimard. Merleau-Ponty, M. (2005) Phenomenology of Perception, transl. by C. Smith, transferred to digital printing, London, New York: Routledge. Mertens, K. (1998) “Husserl’s Phenomenology of Will in His Reflections on Ethics,” in N. Depraz and D. Zahavi (eds.), Alterity and Facticity. New Perspectives on Husserl, Dordrecht, Boston, MA, London: Kluwer Academic Publishers, pp. 121–138. Mertens, K. (2010) “Möglichkeiten und Grenzen einer phänomenologischen Theorie des Handelns: Überlegungen zu Davidson und Husserl,” in C. Ierna, H. Jacobs, and F. Mattens (eds.), Philosophy, Phenomenology, Sciences. Essays in Commemoration of Edmund Husserl, Dordrecht, Heidelberg, London, New York: Springer, pp. 461–482. Nenon, T. (1990) “Willing and Acting in Husserl’s Lectures on Ethics and Value Theory,” Man and World 24, 301–309. Pfänder, A. (1963a) Phänomenologie des Wollens. Eine psychologische Analyse/Motive und Motivation, dritte, unv. Aufl. mit einem Vorwort von H. Spiegelberg, München: Verlag Johann Ambrosius Barth. Pfänder, A. (1963b) Logik, dritte, unv. Aufl. Mit einem Vorwort von H. Spiegelberg, Tübingen: Max Niemeyer Verlag. Pfänder, A. (1967) “Motives and Motivation,” in A. Pfänder, Phenomenology of Willing and Motivation and Other Phaenomenologica. Translated, with an introduction and supplementary essays by H. Spiegelberg, Evanston, IL: Northwestern University Press, pp. 12–40. Schuhmann, K. (1973) Die Dialektik der Phänomenologie I: Husserl über Pfänder, Den Haag: Martinus Nijhoff. Searle, J. R. (1983) Intentionality. An Essay in the Philosophy of Mind, Cambridge, New York, Melbourne: Cambridge University Press. Searle, J. R. (2000) “The Limits of Phenomenology,” in M.A. Wrathall and J. Malpas (eds.), Heidegger, Coping, and Cognitive Science. Essays in Honor of Hubert L. Dreyfus, Cambridge, MA: MIT Press, pp. 71–92. Searle, J. R. (2001) “Neither Phenomenological Description Nor Rational Reconstruction: Reply to Dreyfus,” Revue Internationale de Philosophie 217, 277–284. Searle, J. R. (2005) “The Phenomenological Illusion,” in M.F. Reicher and J.C. Marek (eds.), Experience and Analysis. Erfahrung und Analyse, Wien: ÖBV & HPT, pp. 317–336. Stein, E. (1964) On the Problem of Empathy. Translated by W. Stein with a foreword by E.W. Straus, Dordrecht: Springer Science + Business Media. Stein, E. (2016) Zum Problem der Einfühlung. Eingeführt und bearbeitet von M.A. Sondermann OCD (Edith Stein Gesamtausgabe 5), 3., durchges. Aufl. Freiburg, Basel, Wien: Herder.

27

Karl Mertens Summa, M. and Mertens, K. (2019) “Shaping Actions. On the Role of Attention and Ascription in the Formation of Intentions within Behavior,” in M. Summa and J. Müller (eds.), Intention and Intentionality in Phenomenology and Medieval Philosophy (Phänomenologische Forschungen 2018/2), Hamburg: Meiner, pp. 177–196. Ubiali, M. (2010) “Die Willensakte und der Umfang der Motivation – Eine Gegenüberstellung von Pfänder und Husserl,” in P. Merz, A. Staiti, and F. Steffen (eds.), Geist – Person –Gemeinschaft. Freiburger Beiträge zur Aktualität Husserls, Würzburg: Ergon Verlag, pp. 241–267. Vargas Bejarano, J.C. (2006) Phänomenologie des Willens. Seine Struktur, sein Ursprung und seine Funktion in Husserls Denken (New Studies in Phenomenology 3), Frankfurt a.M. et al.: Peter Lang.

28

3 ALEXANDER PFÄNDER’S PHENOMENOLOGY OF MOTIVATION Genki Uemura

Introduction We often activate our volitional capacity for some reasons. For instance, I go bed because it is late at night; you decide to escape from the university campus because the weather is too lovely to attend the afternoon seminar. What is, then, the relation that holds between these voluntary activities and reasons? According to the Munich phenomenologist Alexander Pfänder, this problem calls for phenomenological considerations. In his 1911 paper “Motive und Motivation,” he writes: “The relation of the grounds [i.e., reason] to the decision of the will is not put into it by the interpretation of an outside observer but is experienced in the actual facts, hence is present in them phenomenally” (Pfänder 1911: 141 [tr. 27], translation modified, our italics). Since Pfänder calls the relation between the volition and its grounds motivation, his contention here is that phenomenology is indispensable for understanding motivation. In fact, his phenomenological analysis of motivation is meant to support the following claims which must have some imports for our theorizing about actions in general. 1 2 3

Motivation is not causation. Motives (i.e., grounds of the will) are mind-transcendent entities. A judgment on one’s own ought does not necessarily yield motivation.

The aim of this chapter is to reconstruct and assess Pfänder’s phenomenology of motivation and the three claims he defends through it.1 In section “Pfänder’s method of phenomenological description,” we outline Pfänder’s method of phenomenology in so far as it concerns with our topic. In section “The structure of Pfänder’s argument in ‘Motive und Motivation’,” we clarify the structure of his argument in “Motive und Motivation.” With these preliminary considerations at hands, we reconstruct his phenomenology of motivation in section “Phenomenology of motivation.” Sections “The nature of motives and motivation” and “The connection between ought-judgment and motivation” are devoted to the exposition and examination of his arguments for (1), (2), and (3).

29

Genki Uemura

Before starting the discussion, let us make clear what we will not do in the sections to follow. We will not take issue with skeptics about the very existence of a phenomenal character (or a set of such characters) peculiar to volitional experience. In what follows, we assume that there is something it is like for us to have volitional experience and that it is irreducible for it is barely fruitful to look for an argument for the so-called agentive or conative phenomenology in Pfänder and other classical phenomenologists. As Kriegel observes, “[i]n the phenomenological tradition, the existence of both cognitive and conative phenomenology is often taken to be unproblematic” (Kriegel 2015: 83; see also 75–83 for his own defense of the irreducibility of conative phenomenology). The philosophical significance of classical phenomenology for this issue, if any, lies in their phenomenological analyses of voluntary experiences, of which Pfänder’s discussion of motivation is an example – or so we shall argue.

Pfänder’s method of phenomenological description As Moritz Geiger, another prominent figure in the Munich school of phenomenology, points out, Pfänder appeals to an analogy with space and material objects in his phenomenological descriptions of various experiences (cf. Geiger 1933: 10). Pfänder sheds light on his target phenomena by utilizing terms that we usually associate with space and matter. For instance, Geiger writes: That love is characterized by Pfänder as “looking-up [aufblickend]” does not mean that he regards love as something spatial. Rather, the concept of direction [Richtungsbegriff ] is something entirely general that is never confined to space but realized in sentiment as well as in space. (Geiger 1933: 10–11) In this section, we give a brief outline of what Geiger calls the method of “analogical description” (cf. Geiger 1933: 11) in so far as they are necessary for the discussion in the rest of the chapter. (As we will discuss at the end of this section, Geiger’s interpretation of Pfänder’s method is not without problems. For the sake of the simplicity, however, for a while we assume that this interpretation is correct.) A key idea behind Pfänder’s method of phenomenology is that the world consists of two different kinds of reality (Wirklichkeit). This point is made clear for the first time in Einführung in die Psychologie. There he claims that while natural sciences investigate on one kind of reality called “material,” phenomenology or psychology has to do with the other kind of reality (cf. Pfänder 1904: 8). Note, however, that the issue is not whether such “mental [psychisch]” reality exists, but what it amounts to (cf. Pfänder 1904: 12). The existence of the second reality is now taken for granted, because, even though it is easily missed in natural sciences, we never doubt it in our ordinary life (cf. Pfänder 1904: 13). Pfänder further holds that the mental reality makes humans and animals something more than merely material entities like machines (cf. Pfänder 1904: 13–16). In other words, the mental reality is made up of varieties of experience, which must be absent in the machines. Another important difference between the material and mental realities is that the latter is not spatial: “The application of these spatial predicate to mental reality has no sense” (Pfänder 1904: 52). This implies that the mental reality is immaterial since every material object is in space. Contrary to what one might expect, however, Pfänder does not mean to make the mental reality entirely isolated from the material reality. Instead, like many other classical

30

Alexander Pfänder

phenomenologists, he emphasizes that consciousness is not like a capsule because it involves the world that transcends it. But how does this cohere with his claim that the mental reality is not spatial? To this question, Pfänder answers that a portion of the material reality, which is not in the mental reality, is nevertheless on the periphery (or limit) of it, when a subject has the object-consciousness (Gegestandsbewußtsein) such as perception (cf. Pfänder 1904: 208; see also Uemura and Yaegashi 2012: 255–258). It is beyond the present aim to examine whether and to what extent this claim is convincing. What is important for our purpose is that it obviously presupposes the analogy between the mental reality and space. Pfänder develops his method of description in “Motive und Motivation.” In this short piece, he makes some distinctions by means of the analogy of the mental reality with space and material objects. Among these distinctions, what matters for the present concern are one (a) between real versus phenomenal causes, and another (b) between the “I-center [Ich-Zentrum]” or “I-core [Ich-Kern]” and the “I-body [Ich-Leib].” (a) Pfänder illustrates his notion of phenomenal cause through some examples (cf. Pf änder 1911: 130 [tr. 18]). When a heard-noise (gehörtes Geräusch) arouses in a subject a striving (Streben) to look at a particular position of her surrounding, her striving has the heard-noise as its phenomenal cause. This phenomenal cause should not be identified with the real cause of her striving, which is a complex set of certain psychophysical conditions. Even though Pfänder himself contends with giving examples in this context, we can clarify his claim in terms of his idea of analogical description. His discussion seems to be led by the observation that we sometimes experience an object as arousing something in us. To capture such an experience, we can describe its object as the cause of experience: just as material objects are causally connected to each other, an analogous connection is also found in the mental reality and its periphery. The analogy with material objects helps us to realize that the concept of cause is also applicable to items in the mental reality. (b) According to Pfänder, a subject of experience, which he calls “the I [das Ich]” is structured by the I-center and the I-body that surrounds it (cf. Pfänder 1911: 130 [tr. 18]; note that they are called “ego-center/core” and “ego-body,” respectively, in the English translation). Nowhere in his 1911 piece, however, does he explain what this distinction amounts to. As he makes explicit later, his claim could be justified well only after considering various mental facts with that distinction (cf. Pfänder 1916: 67).2 In this chapter, we do not deal with this issue in detail. For our purpose, it suffices to highlight a role, perhaps among others, that is played by the distinction. Let us start with one of the reasons why Pfänder brings the notion of the I into his phenomenology in the first place. As pointed out in the earlier passage from Geiger, Pfänder holds that the direction is found not only in the space of the material reality but also in the mental reality. For him, the having-an-object of an experience, which is a characteristic of the non-spatial mental reality, could be described as a direction. Being a direction, however, it must have departure and end points. According to Pfänder, the I serves as either of those points, depending on the kind of experience it has. While, for instance, an experience of striving has a “centrifugal” direction from the I to its object loved, an experience of the arousal (Erregung) is a “centripetal” one from an object to the I (cf. Pfänder 1911: 128–130 [tr. 16–18]). It is at this point Pfänder appeals to the distinction between the center and the body of the I. By this distinction, he describes different ways in which strivings figure in our experience (cf. Pfänder 1911: 130–131 [tr. 18–19]). Most of our – namely grown-up humans’ – strivings, he claims, take place in the I-body. Such a striving, which is called eccentric or off-center

31

Genki Uemura

(in German: exzentrisch), tends to get hold of the I-center, but this tendency is realized without the voluntary control of the subject; the realization is experienced only as phenomenal causation. By this series of descriptions, Pfänder seems to capture situations like this: a striving for water, which has happened to me, comes to hold me so that I now do want some water. Therefore, we can understand Pfänder’s notions of the I-body and the I-center as means to pin down the difference between an experience that figures as something happening to me and one that figures as something I do. Now, with those (and some other) distinctions at hands, Pfänder provides the phenomenological analysis of motivation. Before dealing with this topic, however, we make a remark on how to interpret his phenomenological descriptions. If we are to follow Geiger’s interpretation, we should read “periphery,” “cause,” and other terms in Pfänder’s writings as straightforwardly applicable to the mental reality as well as the physical reality. Then, we should read those terms as literarily expressing the concept of periphery, cause, and so on. This line of interpretation may be attractive, but some passages from Pfänder seem to resist it. In Einführung in die Psychologie, he sometimes holds that terms he uses for the phenomenological description are metaphorical. When he is discussing the “distance” between the I and an object, for instance, he claims that spatial determinations are applied to experiences only in non-literal sense and suggests to take the terms “near” and “far” as “metaphorical instruction [bildlicher Hinweis]” (cf. Pfänder 1904: 357; see also 274–275; we owe this reference to Marbach 1974: 245 n. 40). If we are to emphasize this point, we could not accept Geiger’s interpretation of Pfänder. In this chapter, however, we do not go further into this otherwise important issue. Our reconstruction of Pfänder is neutral about whether his terms for the phenomenological description are metaphorical.

The structure of Pfänder’s argument in “Motive und Motivation” In the introduction of “Motive und Motivation,” Pfänder defines motive as the grounds or reason (Grund) of the will, which is never identical with the cause of the will (cf. Pfänder 1911: 125 [tr. 15]). This implies (1) above, namely that motivation is not causation. It is not charitable, however, to interpret Pfänder as drawing the non-causality of motivation trivially from his definition of motive. Rather, we should understand him as justifying his definition of motive through phenomenological descriptions of motivation. In other words, by giving some good reason to accept the non-causality of motivation, he justifies his definition of motive as a reason rather than a cause. In this way, Pfänder turns to the phenomenology of motivation, but he actually does not deal with the whole range of this topic. Dividing volitional experience into three sub-class, namely act of willing, voluntary action (Willenshandlung), and impulse of willing, he confines himself to the grounds of acts of willing, i.e. volitional experiences of deciding (cf. Pfänder 1911: 126 [tr. 15]). For, he claims, we must have solved the question concerning grounds of decisions if we are to ask what are grounds of voluntary actions or impulses of action. Even though he does not explain why he gives such a primacy to deciding, we can find two reasons for his claim. First, to argue for (1) and thus for his definition of motive, he does not have to deal with the motivation in the other two classes of volition. If he succeeds in showing that the motivation involved in deciding is not causation, he would obtain a sufficient counter-example to those who oppose to his position, for it is implausible to assume that motivation is either causation or not, depending on which class of volition is at issue. Second and more importantly, deciding may well serve as a paradigmatic case if we are to investigate motivation phenomenologically. Even if there is something it is like to be motivated, 32

Alexander Pfänder

it must be much more elusive for the subject of experience than something it is like, say, to see something. While most of us would agree that colors and shapes are figuring in (normal) visual experiences, we are not that sure about what are figuring in experiences we (normally) have when we are motivated to do something. To capture the phenomenal character of motivation, therefore, the best available point to start is a set of cases in which motivation is experientially most salient to us. Having delimitated the field of research, Pfänder divides his argument into two steps (cf. Pfänder 1911: 127 [tr. 16]). In the first step, he shows that deciding is different in kind from striving. In the second step, which includes the phenomenological analysis of motivation, he sorts out various experiences that precede an act of deciding. The first step helps him with clarifying further what is deciding, the second with getting rid of the confusion between the motive and the cause of deciding. In the reminder of this section, I give a brief reconstruction of (i) the first step and (ii) some part of the second in so far as it enables us to see how Pfänder gets prepared for the phenomenology of motivation. i

ii

Pfänder differentiates acts of willing from strivings by two factors (cf. Pfänder 1911: 133–134 [tr. 20–21]): (1) while the acts of willing are always performed by the I-center and thus always take place at that center, the same does not hold for strivings. This description illuminates that deciding figures necessarily as something I do, whereas striving may be experienced as something that happens to me (see Section 1). (2) Of the two types of experience, only acts of willing always involve the awareness of their object, which he calls “projects” (cf. Pfänder 1911: 135 [tr. 23]). To sum, Pfänder takes it as inconceivable that I have an experience of deciding and yet I’m not aware of whether/ what I decide. Pfänder argues that the motive of a decision is distinguished from any striving that precedes it (cf. Pfänder 1911: 138–141 [tr. 24–27]). It is true that a striving may have some influence on a decision of mine. However, he claims, the relation between the striving and my decision is not a motivation for two reasons. (1) I can decide for some reason against striving in me. (2) We can decide in the total absence of striving. Here, one might object that decision-making in the second case is explained by the victory of the striving of which I am not aware. But Pfänder refuses this objection as a prejudice (cf. Pfänder 1911: 139 [tr. 25]). His contention, it seems, is that the postulation of unconscious striving is justified only if one already presupposes that the motive of a decision is striving that causes the decision, but this is question-begging.

In this way, Pfänder gains support for his negative claim that motivation is not a causal relation. But it remains unexplained what is the positive characteristic of motivation. Thus he turns to the phenomenological analysis of motivation.

Phenomenology of motivation3 Pfänder begins his phenomenological analysis motivation with four examples: (1) A person enters a room, perceives the chill in it, and decides on the ground of the perceived chill to leave the room. (2) A person receives from another a piece of work he had ordered; he recognizes that it has been done with special care, and he decides, on the ground of the recognized fact that the other has done the work so carefully, to give him a special reward. (3) A man remembers that when he lived in a certain region, he was 33

Genki Uemura

always very well, and he decides on the ground of the remembered facts to revisit that region. (4) A fourth person decides to forego an action on the ground of the thought that another person could feel hurt by this action (Pfänder 1911: 141 [tr. 27], translation modified, my numberings). Obviously, they all are cases in which a subject makes a decision more or less deliberatively. Even though it is difficult to determine whether he admits the possibility of motivated decision without deliberation, but we can set aside this problem, for it is plausible that motivation is experientially most salient to us in deliberation. Of the four examples, Pfänder picks the first one for his detailed analysis of motivation. Since deliberation is a temporarily extended exercise, the present case must be understood as consisting of more than one temporal stage. Then the question is: at which stage does that person experience something as a motive and get motivated to leave the room? To deal with this question, Pfänder breaks his case into four stages and claims that the motivation is located in the last of them. His four-stage phenomenological analysis could be reformulated as follows (cf. Pfänder 1911: 142–144 [tr. 28–30]; to underline the fact that phenomenology of motivation is at issue, we restate Pfänder’s claims by using the first-person singular pronoun): 1 2 3

4

I enter the room and perceive the chill in it. The perceived chill causes not only attention to it, but also mental listening to it at the I-center. This mental listening involves a questioning stance toward the perceived chill. I receive a demand to leave this room that comes from the perceived chill and acknowledge it. I thereby obtain knowledge about what I ought to do by relying on the perceived chill. Relying on the demand from the perceived chill, I decide to leave the room.

The description of the first stage, which Pfänder does not deal with as a separated stage, does not seem to need any further explication. And, as already pointed out, it is implausible that the motivation in question takes place here. Quite often I perceive something and yet I do not make any decision. The second stage is analyzed as the initial point of my deliberation, in which I have not been motivated yet. At this stage, I come to notice that the room is too cold and try to figure out how to deal with the chill. To describe such a change in my experience, Pfänder talks about the phenomenal causation of attention and that of “mental listening [geistiges Hinhören].” While the former explains how I come to notice that the room is too cold, the latter is meant to pin down what is going on in my experience when I start to deliberate on the current situation. Toward the perceived chill, I take a certain stance, which can be expressed as: “What should I do (right now)?” At the third stage, I receive a response from the perceived chill to my question. Since here I ask about my ought in a broader, non-moral sense, the response to it takes the form of demand (Forderung) for me to do a certain type of action. In the present case, the perceived chill demands me to leave the room. Then I eventually acknowledge (anerkennen) the demand. But, according to Pfänder, the acknowledgment of the demand or ought to leave the room does not amount to the decision to do that. “For the decision of the will is no mere apprehension of ought [Sollen-Erkenntnis]” (Pfänder 1911: 143 [tr. 29], translation modified). It is at this point that Pfänder commits (3): a judgment about one’s own ought does not necessitate motivation. Since he defines motive as the ground of the will, such a judgment, 34

Alexander Pfänder

by means of which I acknowledge the relevant demand, does not by itself bring about motivation unless I simultaneously have a volitional experience in accordance with the demand. At the third stage, in which I do not decide to leave the room, Pfänder maintains, I do not have the experience of motivation. Later I will examine such an argument in detail. For Pfänder, therefore, it is at the fourth stage that I have motivation. As I decide to leave the room, he claims, the demand from the perceived chill becomes a motive for my decision. In other words, the demand is not a motive unless if it is actually connected to the decision (or some other volitional experience). To account for such an actual connection, Pfänder posits an experience called “relying-on [Stützung-auf ].” “This relying-on-something in the performance of an act of willing is a peculiar mental doing” (Pfänder 1911: 143 [tr. 29], translation modified, our italics). One might be suspicious about this move of Pfänder. If a decision is always made on some ground, why is it necessary to add an extra element that connects the decision and the ground? He keeps silent about this problem, but we can give a case that supports his idea. Let us suppose the following: when I enter the room, I feel the smell of smoke and took the stance of mental listening to the felt smoke too; then the felt smell demands me to leave the room; but my decision to leave the room is not based on this demand, because I know a window of the room is opened and that there is someone building a fire outside; I make my decision just because of the demand from the perceived chill. In this situation, I receive two demands, but I experience only one of them as the motive for my decision. To explain the experiential difference in this circumstance, one has to posit an experience that makes the demand from the perceived chill into the motive. It is to describe this experience that Pfänder calls it “relying-on.” Now, by generalizing the analysis of the fourth stage, we obtain the following claim about phenomenology of motivation. (PM) To be motivated for S is to decide to φ while relying on an object of S’s experience that demands S to φ. Note that an object that demands S to φ is not necessarily perceived. (PM) does give no restriction to the kind of that object. In the third of Pfänder’s four examples, it is something remembered that serves as the reason for decision. In his fourth example, the reason for the decision is something thought about. They are, needless to say, not perceived.

The nature of motives and motivation With the help of his phenomenological analysis of motivation, Pfänder then argues for his claim (1): motivation is not causation. Since he distinguishes real and phenomenal causation, he would have to discuss (1) in two different variations: (1-R) Motivation is not real causation. (1-P) Motivation is not phenomenal causation. In support of (1-R), Pf änder gives three claims. First, questions concerning the real cause of the will are irrelevant for phenomenology, because we do not experience the real cause of any experiences in the first place (cf. Pf änder 1911: 149 [tr. 34]). Second, while the real cause of a decision must be present and real at the moment in which I make the decision, the motive of the decision may not be real nor present (cf. Pf änder 1911: 150 [tr. 35]). 35

Genki Uemura

Third, while the real cause of a decision has no degree in its being a cause, the motive for decision may differ in degree (cf. Pfänder 1911: 153–154 [tr. 37–39]). The causation of decision must always involve sufficient cause. In contrast, we can decide on an insufficient reason. As for (1-P), Pfänder gives an argument, which could be reconstructed as follows (cf. Pfänder 1911: 148–149 [tr. 33–34]): i ii iii

I experience my decision as phenomenally caused by the I-center. A motive is mind-transcendent. (Namely (2) above) Therefore, the motive for my decision is not the phenomenal cause of my decision.

How could the two premises of this argument be defended? The first premise seems to rest on the idea that we experience our decision as an active and thus free act. Let us say that we then enjoy the experiential freedom. But some might wonder whether the experiential freedom is well accommodated by introducing the I-center as the phenomenal cause of the relevant experience. Against those skeptics, we can defend Pfänder in the following way. If Pfänder is wrong, three alternative phenomenological descriptions are possible for my experience in question, but none of them succeeds in accommodating the experiential freedom. First, it might seem to be the case that the decision is experienced as ex nihilo, namely without any slightest awareness of its cause. If this is correct, however, making a decision would be as free as, say, a sudden feeling of a pain in my head. Second, it might seem to be the case that the decision is experienced as caused by something outside of me. Such a description is also incompatible with the experiential freedom because, if it is correct, my decision would be experienced as being out of my control. Third, it might be the case that the decision is experienced as caused by some mental item such as a belief, a desire, or a pair of them. According to this description, the decision would be experienced as caused by a mental item just in the same way in which, say, an anxious feeling is experienced as caused by my conscious thought that I might not be able to finish a paper by the deadline. But this does not amount to the experiential freedom. In describing the decision as a free act, Pfänder seems to endorse the following idea of Chisholm: “I shall say that when an agent, as distinguished from an event, causes an event or state of affairs, then we have an instance of immanent causation” (Chisholm 1964: 7). Then one might think that Pfänder is subscribing to a position that is not easy to defend, for immanent causation, which is now more often called “agent causation,” seems to conflict with some plausible claims such as the causal closure of the physical world. This is, however, too hasty a conclusion. It is true that Chisholm himself conceives agent causation as part of reality. It is also true that the idea of agent causation, if it is taken in such a metaphysically loaded way, would face difficulties. At the same time, however, the idea of agent causation has a phenomenological dimension that could be admitted while remaining neutral, at least to some extent, about the metaphysics of agency (for a similar idea, see Horgan et al. 2003: 329). It is possible and perhaps even plausible to interpret Pfänder’s endorsement of agent causation as limited to such a phenomenological dimension alone. Nowhere in “Motive und Motivation” does he say anything about the extent, if any, to which his phenomenology of motivation has metaphysical implications. Also, back in Einführung in die Psychologie, he admits that mental reality would remain real (wirklich), no matter whether it may be superfluous (überflüssig) (cf. Pfänder 1904: 14). In other words, his position in 1904 is compatible with the claim that phenomenal causation is in fact epiphenomenal. Now, what about the second premise, namely the claim (2)? Here Pfänder seems to miss another possibility that is not excluded yet. There remains a room in his framework for the 36

Alexander Pfänder

claim that the motive of my decision is a mental item rather than something out there in the mind-transcendent world. According to Pfänder, my decision is phenomenally caused by the center of the I, but it seems implausible to say that the phenomenal cause, in this case, is the center of the I simpliciter. Instead, my decision seems to be caused by the center of the I qua a subject of the belief that the perceived chill demands me to do so, for, if I had not believed so, I might not make the same decision. Isn’t it my belief about the demand, rather than the demand itself, on which I actually rely when I make the decision? Unfortunately, we do not find any answer nor hint of an answer to these questions in “Motive und Motivation.” But his claim can be supported. As Gareth Evans famously points out, “in making a self-ascription of belief, one’s eyes are, so to speak, or occasionally literally, directed outward – upon the world” (1982: 225). For example, he continues, If someone asks me “Do you think there is going to be a third world war?”, I must attend, in answering him, to precisely the same outward phenomena as I would attend to if I were answering the question “Will there be a third world war?.” (Evans 1982: 225) In short, it is the world that tells us what we believe. Now, just like this case, if we are asked what we ought to do, we would very likely consider the situation we are in rather than looking inwardly into our mind. In this way, it seems, the world tells us not only what we believe but also what we should realize in a given situation. Therefore, against those who consider motives to be mental items, Pfänder could claim that such an idea leads us to a compromise on how we experience those motives when we are deliberating.4 Indeed, such an attempt of defense is far from decisive – the two premises may be still disputable. We hope to have shown, however, that Pfänder’s idea of motivation as a noncausal relation, according to which motives are in the mind-transcendent world, can be one of the viable options.

The connection between ought-judgment and motivation Pfänder’s argument for (3) – a judgment about one’s own ought does not necessitate motivation – could be understood as resting on the following observation: when I judge that I ought to do something, I do not always decide to do it. For instance, when I judge that I ought to help suffering people in front of me, I might consider whether that is feasible before deciding. Such a situation may be most straightforwardly described as an experience of ought-judgment without any volitional experience. Then, since for Pfänder a motive is the ground of a volitional experience, it would follow that I have no motive and thus no motivation yet at the moment of my judgment in question. This would serve as a counter-example to the claim that a judgment about one’s ought necessarily yields motivation. Hence we would obtain (3). Reconstructed in this way, however, Pfänder’s argument for (3) does not seem convincing as a phenomenological argument. Admittedly, there is a sense in which I do not decide anything at the moment I judge that, say, I ought to help the suffering people. If we are to understand the verb to decide as applicable only to cases in which we are pointedly or thematically aware of what we decide, the present case would be described as that of an ought-judgment without a decision and thus motivation. At the same time, however, there might be another sense in which I may decide something even in this case. If we are to understand the verb as applicable even where we are aware of our decision only at the margin of consciousness, the 37

Genki Uemura

present case might possibly be described as follows: I judge that I ought to help the suffering people and thereby decide to do so, but I am not thematically aware of my decision, which is immediately given up (or suspended) because of the consideration about the infeasibility of the decided action. To argue for (3) phenomenologically, therefore, one has to exclude such a possibility by means of the description of experiences. Let us restate the problem in a more general way. What is at stake here is the difference between the following two possible forms that an experience of ought-judgment might have. (OJ-1) I judge that I ought to φ and yet I do not decide to φ. (OJ-2) I judge that I ought to φ and thereby I have decided φ, which is immediately given up (or suspended). To argue for (3), it is at least necessary to show that an experience that instantiates (OJ-1) is not only logically but also phenomenologically possible. For any putative case of such an experience, however, one seems to be able to regard it as an instance of (OJ-2) rather than (OJ-1). Proponents of (3) have to deny this alternative and, to do so, they must be able to tell an instance of (OJ-1) from that of (OJ-2). There certainly is a sense in which the difference is obvious. Arguably, the schematic sentences express two distinct logical forms, as it were, of experience. But this does not seem to certify our ability to discern them phenomenologically, i.e. from a first-person point of view. In other words, it seems that the phenomenal difference, if any, between an instance of (OJ-1) and that of (OJ-2) would be too elusive for us to capture. This is fatal for the proponents of (3), since Pfänder holds this claim as a phenomenological one. Therefore, we have to conclude that we cannot decide whether a judgment about one’s ought has a necessary connection to motivation from a phenomenological point of view adopted by Pfänder. Further note that, contrary to an impression one might have, (3) does not in itself imply any position in the debate between motivational internalism and externalism in contemporary metaethics. It is true that the internalism in this debate is roughly formulated as the claim that there is no internal, a priori connection between moral judgment and motivation; if I judge that to φ is the right thing to do, I am, as the matter of certain necessity, thereby motivated to φ. If we refuse this claim, we will subscribe to a version of externalism in the debate. We should not forget, however, that the very notion of motivation is here construed differently. The contemporary notion of motivation could be characterized as follows: “someone who is motivated to perform an action is inclined to perform it” (Alvarez 2010: 55). To be motivated to φ in this weaker sense, we do not have to decide to φ. Rather, a want or a desire to φ would suffice. What is interesting in this context is that Pfänder’s position could possibly be interpreted as internalism in the contemporary debate. According to him, a received demand is not a motive in his stronger sense unless the agent makes a decision by relying on it. At the same time, however, he admits, “even when something makes a practical demand received by the I, it is not yet a real motive but still a possible one” (Pfänder 1911: 147 [tr. 33], translation modified, my italics). This seems to suggest that the demand, once it is received by the agent, brings about motivation in the weaker sense for her to perform the demanded action. Otherwise, it would be rather mysterious why Pfänder characterizes demands as demands in the first place. Even if this interpretation is correct, however, Pfänder’s position would face another problem. Internalism is typically challenged by the possibility of amoralists (see, for instance, Brink 1986). Amoralists are rational agents who make a moral judgment and yet remain unmotivated to act in accordance with their judgment. In other words, they are potential

38

Alexander Pfänder

subjects of an experience of the form: I judge that I ought to φ and yet I am not inclined to φ. To defend internalism from a phenomenological point of view, then, one has to show that such an amoralist experience is impossible. However, to any attempt to establish the impossibility of the amoralist experience by means of phenomenological description, one could object: “You cannot imagine what it is like to be an amoralist just because you are a moralist. The limit of your imagination is one thing, the limit of the possibility of experience is another.” How is it possible to respond to this objection without begging the question? The best available way would be to note that the objector also has no phenomenological support for the possibility of amoralist experience. For any allegedly amoralist experience, we could offer an alternative description: I judge that I ought to φ, and I am thereby inclined to φ, but my inclination is immediately overridden. Given those considerations, we seem to be reaching toward a limit of Pfänder’s phenomenology of motivation. Regardless of whether we are to define motivation in accordance with Pfänder or with the contemporary debate, his discussion would not help us much with deciding whether there is an internal connection between ought-judgments and motivation.

Concluding remarks – Pfänder’s position in the classical phenomenology of agency Let us close the chapter with some historical remarks. Pfänder’s phenomenology of motivation had considerable influence on philosophers from the phenomenological tradition. Having read Pfänder’s piece on motivation, Husserl left a series of manuscripts in the 1910s (cf. Uemura and Yaegashi 2012: 269–271). Geiger’s and Stein’s discussions of motivation can be understood as reactions to Pfänder’s conception of motives (cf. Uemura and Salice 2019). Pfänder’s idea of the I-center as the phenomenal cause of volition is echoed in Reinach’s characterization of spontaneous acts as experiences that have the I as “phenomenal originator [phänomenaler Urheber]” (Reinach 1913: 705). Stein subscribes to a similar idea in discussing free acts (Stein 1922: 46). Reiner provides a detailed analysis of volition in which Pfänder’s distinction between the I-center and the I-body plays an important role (Reiner 1927: 127–133). Last but not least, Ricœur draws on Pfänder in his phenomenological analysis of decision (1950/2009: pt. 1, ch. 1). As such a short glance already suggests, it is not an exaggeration to say that Pfänder is an originator (if not the originator) of the classical phenomenology of agency.5

Related topics Chapters 2 (on Pfänder and Husserl), 5 (on Reinach), 7 (on Reiner), 15 (on Ricœur), 19 (Horgan & Nida-Rümelin)

Notes 1 It must be noted that the analysis of motivation is a part of Pfänder’s phenomenology of volitional experiences. For an overview of the latter, see Uemura and Yaegashi (2012). 2 At this moment, Pfänder no longer uses the term “I-body” and holds that the I consists of the center and the “self (Selbst)” (cf. Pfänder 1916: 67–68). 3 Parts of this and the next section overlap with section 2 of Uemura and Salice (2019). 4 At the same time, Pfänder’s position has to square with the fact that sometimes we decide on reasons that cannot be identified with anything in the mind-transcendent world. For instance, we may decide to leave the room by following the demand from the hallucinated chill. On this issue, see Uemura and Salice (2019).

39

Genki Uemura



40

4 SCHELER’S PHENOMENOLOGY OF FREEDOM AND HIS THEORY OF ACTION Eugene Kelly

The platform: ethical personalism In the Second Part (§ 6) of Der Formalismus in der Ethik und die materiale Wertethik (cf. Scheler 1973) Max Scheler is concerned to distinguish his new platform for normative ethics both from the deontology of Kant and the teleological eudaemonism of Aristotle and the Utilitarians. First, his material value-ethics is not a formal ethics such as Kant’s, that is, an ethics founded in a single formal or a priori moral principle, the categorical imperative, which governs a consistent system of moral laws that determine, in ways specified by Kant, what is morally legitimate for an agent. These formal and rational moral laws are dictated by an agent to him or herself insofar as he or she is rational. Hence Kant’s ideal agent, like that of Scheler, acts autonomously. Kant teaches, according to Scheler, that the impulse to action emerges from the chaos of natural “inclinations” toward self-gratification, but that practical reason must override these out of respect for duty and allow the rational good will to direct action in conformity to the formal system. The phenomena that Scheler discloses speak against this Kantian account of moral action. Conformity to duty alone does not give an agent moral value, for duty, simply as a formal command to obey the moral law, bears no reference to values beyond itself. Duty emerges, on Scheler’s account, from the cognition of an “ought-to-be” directed at the agent’s will. The agent perceives an obligation to realize something of higher value than currently exists, or to destroy something of lower value than currently exists. Thus the duty to abstain from drinking before driving is founded upon the value-perception (or cognition of value) that the value of preserving human life is higher than the value of the pleasure of drinking. The notion of “duty for duty’s sake” is nonsense, for the adherence to duty must itself be justified by its aiming at a value beyond itself – that is, it must be itself governed by a value that is not simply the value of adherence to duty. Otherwise, we could not question the legitimacy of some purported duty. Kant’s notion of a good will concerned only with its duty and not also with the value of that for the sake of which it exercises its agency would have dignity, but no other material worth. Such a will, Scheler quotes Sigwart, would be a will that “does not will what it wills” (Scheler 1973: 121; all quotations are taken from this text). Now it would seem that the only possible alternative for ethics to a moral principle such as “duty for duty’s sake” would be a eudaemonistic moral theory. J.S. Mill, a utilitarian 41

Eugene Kelly

eudaemonist in ethics, was aghast at Kant’s failure, as he saw it, to establish any specific moral rules (cf. Mill 1998: ch. 1.)1; he had posited instead an empty moral rule that requires us to act according to moral rules. Utilitarians, in contrast, can give material content to moral rules by demonstrating their tendency to achieve desired states such as happiness, pleasure, or utility. Justice and virtue are satisfying to contemplate in themselves, but their value is derived from the consequences of just and virtuous acts, which usually cause, intensify, and distribute pleasure or happiness in the form of states or goods. Whether such utility-motivated actions and activities must or in fact always conform to moral criteria such as right, virtue, or goodness independent of the utilitarian Greatest Happiness Principle need not be at issue. Kant presupposes that eudaemonism must of necessity “place the ground of all ethical valuation in the instinctive egoism of man’s natural organization” (Scheler 1973: 7), that is, in our sentiments alone. However, Scheler believes that there is an alternative, an ethics founded upon phenomenological, a priori, and rational knowledge of material values. During his middle period, approximately from 1912 to 1921, Scheler, along with similar efforts by E. Husserl, N. Hartmann, and D. von Hildebrand, attempted to establish a basis for normative ethics that would achieve the best of both worlds: like Kant’s ethics, it would be non-teleological, for it would not aim at the realization of goods as things we desire, and yet not be “empty,” but have material content. It would also be non-deontological, like Mill’s ethics, and yet possess concepts of duty and right. Thus their efforts aimed at what came to be called the material ethics of value. The material content is supplied by what Scheler called the “values themselves,” that is, the content of essential value-structures or meanings – Husserlian essences, if you will, or ideal existences, as Hartmann put it 2 – that are intuitable in cognitive noetic acts of feeling. Values are “carried” by objects such as goods and events and persons, as Abraham Lincoln is perceived by many to have “carried” the value of honesty as a virtue. Pleasure or happiness is only one kind of good, and its achievement has no moral value in itself. But when a person wills to realize some state or object because it carries a value higher than those currently present, or wills some normative ought-to-be founded upon a higher value than any alternative, he acts morally. What, now, are some kinds of values, and how do we have cognitive access to them? For Scheler, the world given in perception is saturated with values. It is a task of phenomenology to make those values the objects of systematic reflection by re-executing carefully the cognitive feelings, including those of preference and rejection, in which they are given, and to distinguish among their objects. The acts of preference and rejection suggest that pure values possess an order of relative worth that can be made evident when the phenomenologist reenacts feelings that have been purified in phenomenological reflection (that is, those performed after the bracketing of specific goods and situations). He then directs noetic acts of feeling upon values as qualities in themselves, and not as they appear on the contingent objects that may carry them. Thus we may attempt to intuit the value of courage by first directing our attention toward the many objects to which persons tend to attribute this material value: toward certain human actions and the situations they respond to or toward a certain type of human character. Then we may abstract from these goods and focus on the value itself. By varying our examples, our intuitive sense of the value of courage becomes clearer. We may also seek to exhibit the ways in which some values are intrinsically higher than others.3 This emerging table of values extends from the values of physical states, in general the pleasant and the unpleasant, to the values of life, which occur on two levels, the values of the useful and the worthless and those of the noble and the common. The still higher values of the spirit are those of the true, the good, and the beautiful, whose relative disvalues are the false, the wicked, and the ugly. Finally there are the values of the sacred 42

Max Scheler

and the profane, which have their foundation in the value of the holy. On certain occasions a good pair of boots may have more value to the wearer than the works of Shakespeare, although the wearer of the boots would acknowledge that, viewed apart from his immediate situation, the value of great literature is preferred as a higher value than the value of comfort. The subtitle of Formalism in Ethics is “A New Attempt at an Ethical Personalism.” The human person bears the highest value, although the value of the personhood of an individual human being is not a material value that can be brought to givenness in a noetic act. An individual’s personality or character can be intended; it is intuitable as carried by her ego. But the personhood of the individual, although present in each of her acts, is not an intuitable content. Thus we can grasp the phenomenon of personhood as the highest value, but we cannot make some individual person who carries this and many other values an object of knowledge. For a person is forever incomplete; she exists verbally; she is, as it were, a being whose existence is executed in acting as a spirit, mind, and body. The subjective core of a person is what Scheler calls the order of her loves and hates, a scale of feelings directed at values themselves apart from goods. Persons can glimpse this order in others, and yet not exhaust the nature of their personhood. This crucial doctrine will occupy us again in the context of Scheler’s discussion of human freedom. With this doctrine of the “Ordo amoris,” (1912-14)4 Scheler affirms that the feelings and emotions of every individual are neither chaotic nor random, nor invaders of the rational core of the autonomous human person, as Kant believed. They have an internal integrity that refracts, clearly or darkly, sharply or skewed, in each person in a unique way, the objective order of values. Upon the basis of an individual’s Ordo amoris emerge three additional and intuitable elements of human personhood: first, the Gesinnung or the basic moral tenor of a person. Second, her sense of a calling in life (Bestimmung), and third, her sense of her fate (Schicksal), which latter term Scheler uses for the sense we usually have of a continuous meaning of our lives. A normal person’s activities and purposes are coherent in the present and as she moves toward the future; they are not the products of caprice, for they trace their ancestry to the stable Ordo amoris. Even one’s acts of will, as we shall see, are posterior to an agent’s apparently inborn sense of values and how they are prioritized.

The phenomenology of freedom It is not possible to understand Scheler’s phenomenology of action and the relation he discovers between the pure content of will and the process of its execution except on the background of a far more complex picture of the human person than those of Hume, Kant, Bentham, or even of Aristotle. This is clear from an essay that Scheler’s wife and later editor, Maria Scheler, included in a posthumous collection of brief essays from the time of Formalism in Ethics (1913-1916) for the second edition of the Schriften aus dem Nachlass I under the title “Zur Phänomenologie und Metaphysik der Freiheit” (cf. Scheler 1957). This phenomenology of freedom will introduce our discussion of moral action in Formalism in Ethics, and we will close with an assessment of the consequences of both for Scheler’s inchoate early metaphysics of freedom. First, he notes, the sense of being free rests upon the presence of two phenomena intuitively visible in the agent. The first is capability (Können), the sense that I can do something or other. Without it, there can be no freedom; there would be no room for an action’s being mine. Second, there must be a sense of my own power to will, the sense not just that I can do something abstractly (visit Mongolia, e.g.) but that in these circumstances I can proceed to action. The sense of being able to act in some way, that is, to encounter my own freedom, is 43

Eugene Kelly

founded in awareness of my ability to determine spontaneously my choice of some course of action. This spontaneity does not imply indetermination or a lack of a sufficient reason to act, which is merely negative; freedom is a positive sense of having a capacity to act or forbear at this moment, and eventually to determine the will to act or forbear for the sake of some end. The determinist will now object that this awareness of capacity may be false; an agent’s sense of power over his actions may be the result of causal conditions over which he has no consciousness or control: typically, the shaping of the agent’s behavior by his environment and its determination by evolutionary and genetic factors. But this hypothesis, when it is applied to all cases of human action, and considers them as events in a causal chain, misses the origin and complexity of the phenomenon of action. Scheler writes, Determinism knows nothing about the making (Schließen) of decisions (Entschlüsse), of the swelling up of life, from which [decisions] intuitively arose – rather it presupposes them as done, and takes notice of them. It binds them together afterwards, as a process of effect and cause – but determinism does not observe how each these processes, which it simply ties together in this way [as cause to effect], come to pass out of the inner workshop of effective life itself. (Scheler 1957: 158) This brief critical phenomenology of the process by which human action comes to pass will be exemplified in the following account of Scheler’s phenomenology of action.5 Afterward, we will return to the metaphysical question of whether some human actions are free because they are founded only upon the spontaneity of the person, and whether that freedom confers moral responsibility upon the agent. Further we must ask whether Scheler’s theory of the person limits the range of the human capacity in a way that compromises any metaphysical claim that humankind is free and morally responsible for its actions.

Action theory Human action is a process whereby a person is initially goaded to action by a subliminal urge (Streben) or conation that becomes conscious and who then passes, reflectively, emotionally, and physically, through a unified process of responses, to action itself, whose aim is achieving some state of affairs in which values or disvalues present in the conation are realized or destroyed. The phenomenological description of this process is of course central to ethics, which, as aiming at the understanding and moral evaluation of action, does not require reference to processes and events outside of the awareness of the acting persons themselves such as to brain events, to learned associations, or to processes in a purported unconscious mind. It will, however, presuppose and be an application of the phenomenology of the human person and her world that was sketched earlier. All action begins with conation. “Conation,” Scheler writes in Formalism (Scheler 1973: 30 fn 24), “here designates the most general basis of experiences that are distinct from all having of objects (representation, sensation, perception), as well as from all feeling (states of feeling), etc.” Conation is the restless urge in the ego that awakens our attention to itself even before we become aware of a specific object toward which it may be directed. It may be entirely undirected (a “dumb urge:” something’s got to change!), or it may move toward or away from an unobjectified external or internal state, as with an inchoate desire for, or fear of, something it knows not what; it “rises up within us.” The conation can arise from a

44

Max Scheler

stimulus within, as when we feel hunger or lust even before we become thematically aware of it and desire something to relieve it, or from one without, as when we sense the somber growth of darkness in the late afternoon even before we think of turning on the light or seeking to change our mood. In sum, conation first moves toward values of which the agent is antecedently, although at the moment only subliminally, aware, and not toward representations of objects. That by which conations are initially differentiated is always a value of some kind that is felt as positive or negative – again even before we give it an object, or represent to ourselves a state that ought-to-be. “The values of things are given to us prior to and independent of pictorial representations.” Thus, the conation of hunger has a clear direction “toward” the value of satisfying hunger, even if it may not yet possess a representation of a goal (Zweck) such as a meal (Scheler 1973: 33) Conations are the foundations of goals – if no conation, no goal – but the conation does not determine the goal; one may refuse to respond to the value toward which the conation moves, and move instead toward some other value. Conation is different from desire, which always “pictures” its object. The value given in the conation raises, as it were, the very question of its goal: given the conation, what next? Imagine a physical weariness out of which wells a conation that is directed toward relief, at “taking a vacation from labor.” An unspecific urge toward the value of relief from weariness floats in front of the mind as a goal, and only later imagination “fills in” the desire with activities possessing specific value-states that could make up the relief from work. Having such a general “goal” – a day off, a long lunch with friends, a good night’s sleep – is not yet the performance of an act of will, for no content that could function as a purpose of action has yet been posited. But in the conation toward the value of relief a representation or picture of such relief begins to form, i.e. some imagined state of affairs that would provide relief. An essential law of action states that the former founds the latter in that the choice among possible states of affairs is contemplated as appropriate or inappropriate to the values toward or away from which the conation moves (cf. Scheler 1973: 33-34). Scheler summarizes: “The conations themselves are determined and differentiated by (1) their direction, (2) the value-component of their ‘goals,’ and (3) the picture- or meaning-content arising from this value-content” (Scheler 1973: 39). Purposes are fundamentally different from goal-directed conations. Purposes involve the representation of a state of affairs as possible, but thinking about or visualizing a thing does not necessarily involve a conation toward it, or the willing of its (possible) existence as a goal. A detective, examining a bank, may ask himself, “How might a thief obtain access to this bank?” He imagines a thief positing such a purpose as a response to a conation toward money, but the detective’s own purpose is not to obtain access to the bank. Scheler puts the key issue as follows. What distinguishes “purpose” from a mere “goal” [e.g., of satisfying hunger], which is already given “in” conation itself and in its direction, is the fact that in a purpose a goal-content (i.e., content already given as a goal in conation) is represented in a special act. It is only in the phenomenon of “withdrawing” from conative consciousness toward representational consciousness, as well as toward representational comprehension of the goal-content given in conation, that the consciousness of purpose comes to a realization. Anything that is called a purpose of the will therefore presupposes the representation of a goal! (Scheler 1973: 39–40)

45

Eugene Kelly

Conation is not an analogue of purposeful willing; it has its origin both in the biological urges and the ego of persons, which are structured differently from the level of mind that grasps purposes and consciously wills them to be. The having of purposes, like the having of goals, does not necessarily lead to acts of will. Representations of purposes remain the mere objects of wishes (“I’d like to marry that girl.”) unless the representation is also given as to-be-realized by the agent (“She’d never accept.”). The genuine purpose has a value that the agent apprehends as demanding to be responded to and brought into existence by him. Thus only when both the representation of the goal and the demand are given can there be a purpose leading to action. Scheler marks again his distance from Kant: values, he argues, are not dependent upon or arise out of our conations or ego-driven goals; rather they found the differentiated goals of conation, and are, as goals to be realized, also the foundation of purposes. In the psychic life of persons, knowledge of values and our individual attunements to them are antecedent to goals and purposes. This is a key to understanding Scheler: values are native to persons, ab initio as it were; they are given in the acts of love and hate that open and close infants to the world before they govern goods that speak to their developing ego and become objects of craving. When we judge a person to be acting freely, the events leading to action possess a unity of meaning or sense. The process flows in a person from an initial bodily and psychic state to conation, and then, via a representation of the values before his mind, to the facts present in the situation in which the person is to act, then to the person’s sense that the values before him demand a response from him, and terminate in the performance or “deed” (Handlung). When the process lacks such unity, or appears to be driven by factors having no realistic connection to the situation, as with a maniac, an observer would fail to understand the process and the deed that emerges from it. Scheler notes that in such cases we cease to understand the agent’s behavior, but instead attempt to explain it. “A human being is that much more comprehensible, the more free he is” (Scheler 1957:159). We may take the following situation as typical of a unified action that has moral content.6 A runaway van overturns on the street, the gas-tank has cracked, and gasoline drips to the street. The driver lies unconscious in the van. A woman passing by experiences an emotional arousal, a conation: Danger! That conation is not initially aimed at any specific end. It is a directing of attention toward an external event in her milieu; as yet, it is arousal without specific purpose. The passer-by then grasps the danger to the driver, assesses the danger to herself, senses herself called upon by the value of the driver’s life, rushes to the van, flings the door open, adjusts her movements to the vicissitudes of the situation, and with others carries the driver to safety. The moral assessment of the act of the passer-by requires that her action be viewed as a unified process, and should not, as in Kant, be seen either as a synthesis of drive, intellection, will, movement, and coordination, or as a causal series of behavior and laws covering them. On the phenomenal level, the Kantian causal models assimilate human behavior to that of a machine, whose mathematical laws render its operation lucid; but for human behavior such thinking submerges and makes invisible the organic unity of the action, or so Scheler believed. For we cannot locate the moral value of the person, conceived as an isolated “force” in a causal process, solely in her will to respect and obey the moral law. Kant’s purposes in affirming that thesis is noble, for it insulates the moral value of the agent from the contingencies of fate or luck, and it establishes a kind of moral equality among persons: each is entirely responsible for what she wills; no one can be responsible for the factitious outcome of acts of will: the success or failure of the rescue attempt, even the degree of risk involved, has no

46

Max Scheler

bearing on the agent’s moral worth. But Kant’s moral focus on will loses sight of the disparate values realized in the process of the moral action. The elements in an action or deed are as follows. We follow Scheler and apply each element to the narrative of the rescue-scenario. 1

2

3

4

The presence of the situation and the object of the deed. A conation is directed toward some “practical value-objects” that have appeared to the agent within the range of values available to her, given the structure of her milieu and her basic moral tenor. In our situation, a conation directed toward the value of the imperiled life of the driver may be prominent. The agent senses an incipient demand for action: “Something must be done.” The practical object in the situation is given as resisting or “with-standing” the will, which aims at saving the driver’s life; if she could simply “wish” the driver’s safety and it would be done, no action would be needed, and a “willing to do” could not be formed. This resistance to the will need not be physical; the mere will of another can resist her (the command of a policeman), or, in some cases, the qualities of the situation itself. The phenomenon of resistance is normally found in the following order: it is placed (1) in the object beyond the ego and body of the agent (a barrier of some kind – a wall, distance – stands in the way of action); (2) in the body itself (one cannot run fast or far enough); (3) in the psychic sphere (the person may be conflicted; she wishes, after all, to preserve her own life, which a spark to the gasoline could end). The content of the efforts required by the deed (What is to be done?). Here, along with their concomitant values, “pictures” arise in the mind in which possible courses of action appear: “running away from,” or “running towards” the stricken driver. The willing of one set of compossible value-contents. The path of the decision leads from the moral tenor through intentions, deliberation, and resolution. The moral tenor of the agent may incline her to accept the level of risk to her own life in an attempt to rescue the driver. On the level of intention, she experiences the rescue of the driver’s life as an ought-to-be that founds her purpose. She then deliberates as to how best to effect the rescue, and resolves to take that course of action. Of course she cannot propose to act counter to her moral tenor, and her deliberation takes account of her own capacity for action, that is, her knowledge of what she is capable of doing (Scheler 1957: 128). Thus if this passer-by is disposed to assist the trapped driver, she could only imagine herself performing a rescue that is within the range of her power (her sense of the capacities of her body). She might imagine herself running to the rescue, but not flying there. Scheler notes, admirably, that this experience of capacity or incapacity is an immediate sense, and is not dependent upon one’s past successes or failures in conducting a rescue in situations similar to this one. It is simple and unique, like the sense of being alive, which is neither simply a piece of knowledge, nor built up from such states as vitality or weariness. The sense of “being able” cannot be improved by exercise and practice; it determines rather what activities we will practice and develop, and how we go about doing so (cf. Scheler 1973: 232–237, “Ability and Ought” [Können und Sollen]). The psychic activities directed toward the lived body leading to movement of the members (the “willing-to-do”). Here the agent moves beyond mere intention (a wish that the driver be saved, that he “ought to be” saved) to a decision to engage herself in the values carried by the situation (“I will run to his assistance”). The will-to-do is a willing to do something – not the willing of its outcome (I will save him), which may be beyond

47

Eugene Kelly

5

6

7

the agent’s control. Of course, if she is wise, she “calculates” her chances of achieving a given outcome, but she wills the action in the hope that it will have that outcome. The states of sensation and feeling present during the execution of the selected action. The transition from a willing-to-do to the action itself takes place across feelings and sensations, such as the bodily movements of the agent, or her sense of strength or weakness as she moves. She then immediately (i.e. non-mediately, unvermittelt) senses the effectiveness of her resolution upon her body. “There is an efficacy of willing that acts on our lived body and issues forth into movement” (Scheler 1973: 130). The kinematic sensations and visceral events direct or specify the conations that propel the action; they may indicate a need for more effort, or a change of direction, and they may change as the action proceeds. Since the efficacy of will is variable, it has moral relevance: she may initially desire to come to the rescue, may even imagine a course of action and proceed to act, but then feel reluctant and cease; her conation evaporates. Alternatively, she may be overly fearful, foolishly triumphant, or selfless, and these states modify the moral value of her action. The experienced realization of the content (the “performance”). The object for the sake of which the action was initiated – the saving of the driver – is joined with the content of the will-to-do (i.e. to do the action thought needed to save him). The trapped driver is or is not reached, is or is not pulled free. The agent grasps the success or failure of her performance. The states and feelings responding to the ends that have been realized. This will be joy or sadness, satisfaction or regret. Note that the outcome of the action – the safety or death of the driver, or of the passer-by – does not belong to it as part of its moral content and certainly not to its unity of meaning as an action. For the results are always contingent, that is, dependent upon luck or circumstances and not upon the agent’s will. Yet if the agent survives, her feeling of success or failure in what she tried to accomplish brings additional moral value to her action. We tend quite rightly to admire an agent who belittles her role in the success of her action, or is aggrieved at her failure. Such evaluations are directed at the moral values manifested upon the agent, not upon her success or failure itself, which outside observers would simply welcome or regret. Those observers would of course rejoice at the sight of the saved driver, but that is not to evaluate the rescuer.

The teleologist in morals, who proclaims that the action was good simply and only because the man was saved or because other positive outcomes arose from it, such as the joy of the rescued person and his family, his ability to return to work, and so forth, is simply untrue to the values we feel and the judgments we make when we perceive the action. The deontologist who says that the action was good simply and only because the agent’s will was good – that is, because it chose to do its duty “for the sake of ” the duty – offers an inadequate account of the agent’s moral worth; he leaves out morally relevant value-material such as the virtue of her boldness, and the value of the life of the driver. If this phenomenology of action is correct, material value-ethics is able to argue that the moral worth of the action is not located in the formal intention of the agent alone (respect for the moral law, the sense of obligation to help a person in distress), though her intention has moral worth, but it is also not dependent upon the success or failure of the action. Its moral worth is found in the person – thus ethical Personalism – who, whether or not the intended action is brought to completeness, acts throughout its execution. Of course, the central moral feature of the unified act is described in point 3, in which the value of the content possibly present in the situation (the life of the driver as a value) is affirmed by the passer-by. It originates in the basic moral tenor of the agent, from which springs a conation urging her toward this specific material 48

Max Scheler

value. Point 4 elevates action above a mere wish that things be a certain way: in wishes a material value-content is no doubt given (I wish there were no hungry children), but there is no resolution to do something to realize that wish; it does not terminate in action. What are the implications of Scheler’s characterization of an action as a unity? Of course there are phenomenally distinct parts within the action that can be brought to givenness. In the complex action we have taken as a model, there may be other factors contained within it, indeed random and incomplete ones that do not dissolve the unity of the act, such as the agent’s initial hesitation, her fear, her self-questioning (“Why get involved?”), insofar as the agent “pays these intruders no mind.” We must keep in mind that the unity Scheler is speaking of is a unity of sense or meaning that runs through the entire action. This unity of meaning is what the “disinterested rational observer” understands when he judges the act as the moral whole it is. Scheler’s insistence upon this point derives from his concern that if we divide the action into a chain of events, where the only genuinely moral event of the chain is the act of will that “causes” the action, then the action could not bear any moral value, for the will alone would bear it, all else would be a causal outcome of the act of will and the circumstances in which it was exerted. In that case, the moral philosopher is forced to choose between a deontological and a teleological ethics: the act is good either because the will was good or because the outcome was good. This was the problem we faced at the outset. But for material value-ethics, for as long as the rescuer is engaged in the act, she bears moral merit or demerit of different kinds throughout, until events pass “out of her hands” to where she no longer has effective control. In our moral assessment of a person we normally ask ourselves: what material values are present in her action? What situation and the values it will bear is she attempting to bring about? Is her action free and spontaneous, on the one hand, or either coerced or the product of caprice? Could determination, skill, and fortitude be perceived upon the agent’s person and upon her character, factors which made it possible for her respond in acts of feeling to the positive values she felt within her situation and upon the ends she willed to realize? We sense the moral worthiness of this rescuer as she carries out her deed; we admire her for what she attempted to do by means of the deed. Moral evaluation must consider and prioritize these meanings if it is to be complete. In sum then, the agent seeking to rescue the trapped driver does not will that he be saved, for that is impossible. Rather she wills to perform an action aimed at saving him, whose values, purpose, and the technique it demands she grasps and visualizes beforehand and as she moves toward the driver. Her moral merit or demerit lies in the material values functional in her person that condition her moral attitudes, her fortitude of will, and even her response to the outcome of her action.

The metaphysics of freedom The metaphysical reflections upon freedom and determinism that follow upon the phenomenology of freedom in the essay, Zur Phänomenologie und Metaphysik der Freiheit (manuscripts written between 1912-14), are founded upon the conviction that the phenomenon of freedom and moral responsibility must be affirmed as real, given the phenomenology of action in Formalism in Ethics that was just presented. He writes in this posthumous essay, “I or another may in any given concrete case be under an illusion about my own or he about his freedom; that there is freedom, is a phenomenon found even in illusion” (Scheler 1957: 162). He is harsh in his rejection of past thinkers who denied the phenomenon itself as illusory. In Hume and Bentham, according to Scheler, moral agency is seen as an application of a mechanistic metaphysical theory derived from the sciences of nature, where human action 49

Eugene Kelly

is assumed to be determined by a struggle within the ego or the sentiments among motives of varying strength. Scheler notes that the idea of an ego of constant character that works upon motives from a position outside of them, or chooses “freely” among them, implies a pure indeterminism that makes no sense of action (Scheler 1957: 175). Kant postulates a noumenal freedom of the human person that renders freedom humanly incomprehensible, and Bentham, at one point, suggests without further discussion that we can only do what our desires for pleasure and fear of pain can move us to do, and hence humans are unfree.7 Fundamentally, the error that is common to the Enlightenment thinkers is “the separation of the ego from the so-called ‘motives.’ The human being varies itself as a whole in each of its experiences [Erlebnisse]. Experiences occur in pure duration as a unique concrete flow, in which nothing is separated” (Scheler 1957: 162). “Ego” in this essay seems to be conflated with the “person,” although an appendix to Scheler’s essay attempts to clarify the distinction. In Formalism, Scheler distinguishes the functionality of the ego, which is presumably the sources of conations, and which corresponds to the lived body (Leib), from that of the psyche, which is the emotional center of a person, and that again from the “unobjectifiable” individual human person. “Experience” refers to noetic acts, including body-awareness, the feeling of values, and intellectual judgments. But there is one self-developing human person. The separation of the ego from the “motives” or conations results in viewing the human being as a complex mechanism whose “states” are measured by clock time rather than by duration. Scheler argues further against Enlightenment thinkers that epiphenomenalism, which developed in the wake of the Enlightenment, is a “construction” that maintains that ideas are emanations of brain activity and can exert no influence upon it. It has an absurd consequence: “Thus the denial of psychic causality by the epiphenomenalists leads to a conception according to which the processes of psychic events are purely arbitrary” (Scheler 1957: 163). Some state of mind A m that is the epiphenomenon of brain event Ab could have no link in meaning to the mental state Bm that follows it, for Bm is an effusion of a different brain state Bb, and yet two brain states cannot be linked by meaning nor caused by A m or Bm. In a word, how could we account for a meaningful and coherent succession of ideas if they are caused by a series of uncomprehending brain events? Rather he holds it to be phenomenologically evident that mental acts, selecting from and interpreting the phenomena present to them, emerge from what Scheler calls the lawfulness of life, one of the highest levels upon which freedom appears. These levels form a “step-like construction of growing freedom, whose essential content we can describe upon the phenomenal level without reference to the concrete order of nature” (Scheler 1957: 162). No doubt the phenomenon of the oasis is given in the mirage; but the question remains, is it a mirage? Scheler is well aware that behind the scenes of our consciousness causal events in our brains, unknown in their details, could be the source of some or all of the conscious events whose running-off they determine. He cites the claim of determinists that the consciousness of freedom “could be an illusion; that it could be illusory also in that the determining motives of our action are hidden from us, and we place other [motives] or none at all, in their place” (Scheler 1957: 162). Powerful considerations today indeed speak for such a possibility despite Scheler’s skepticism,8 such as the principles of sufficient reason and the conservation of mass and energy, considerations that weigh especially heavily when, as today, the nature of conscious mind in its relation to the brain and body is still quite unclear. The complex phenomenology of human action, where human conations are shaped by the feeling of values, where ordered emotions cognize goals, purposes, and actions while the mind rationally assesses the actual situation, makes the case for a linear mechanistic determinism of human action unpersuasive. The question of moral freedom, however, is not 50

Max Scheler

solved until the question of the origins of our moral character is solved. Simply put, to what extent are the Ordo amoris, Gesinnung, moral milieu, calling, and fate of a person, around which a person’s ontic life takes on meaning, God-given, inborn, or freely chosen by him or her? Scheler has no clear answer. Scheler admonishes the person to “become what she (normatively) is,” but he leaves the source of what a person is, her unique order of values, quite obscure, even more so by insisting that its origin and content are not subject to the will of the individual. Nicolai Hartmann saw clearly Scheler’s further error: by denying that the person, who is constituted morally by these phenomena, can become an object, Scheler also denied to the person the status of a subject given in intuition that is the object of moral judgment.9 It seems that Scheler is in the end judging not the personhood of the intrepid rescuer of the driver, which lies beyond all direct knowing, but the character she manifests externally. Only God knows her personhood.

Related topics Chapters 21 (Hanna), 23 (Strawson).

Notes 1 Mill refers to Kant’s failure to deduce actual duties of morality from the categorical imperative as “almost grotesque.” 2 Cf. Hartmann (1932: 184): “In their mode of being, values are Platonic ideas.” 3 Formalism in Ethics, Part I, Par. 2, exhibits phenomenologically this ladder of values. 4 The term “ordo amoris” appears in a posthumous essay written, according to Scheler’s editor and wife, at about the time of Formalism in Ethics (1913–16); it may have been intended for inclusion in the second half of that work, but Scheler never released it for publication, and rarely refers to the term in his published work, though he speaks in Formalism of Pascal’s ordre du coeur or elsewhere of the Ordnung des Menschenherzens. He did not attempt to clarify the origins of the Ordo amoris or its development over time in individuals. 5 In Scheler’s late work, especially in his The Human Place in the Cosmos (cf. Scheler 1928/2009), this account of action is radically transformed. Scheler’s inchoate metaphysics and philosophical anthropology posit the impotence of the human will to realize values without the connivance of the blind “drives” (Drang). The current writer judges that the phenomenology of action that Scheler produced in an earlier phase of this career in philosophy is quite compelling, whereas the picture of humankind and its “place” in the cosmos are deeply flawed and unconvincing. 6 The following example was taken from and develops an earlier analysis of the process of action according to Scheler in Kelly (2011). 7 Cf. Bentham (1948: chapter I, 1): [“Pleasure and pain] govern us in all we do, in all we say, in all we think.” Bentham, I believe, was not always consistent in his adherence to this empirical claim. 8 He argues that even in illusion (Täuschung) the phenomenon of freedom is given. Certainly that is so; I may intend the phenomenon of freedom of the will, even when I am subject to manipulations unknown to me. But Scheler is not in a position to hold that phenomena have metaphysical force. 9 Cf. Hartmann (1932: 319–321): “Ethics takes as its object [a person’s ‘actively transcendent acts (disposition, will, conduct)’] what Scheler says is incapable of becoming an object.”

References Bentham, J. (1789/1948) An Introduction to the Principles of Morals and Legislation, New York: Haffner. Hartmann, N. (1926/1932) Ethics. Volume I: Moral Phenomena, London: George Allen & Unwin. Kelly, E. (2011) Material Ethics of Value: Max Scheler and Nicolai Hartmann, Berlin: Springer. Mill, J. S. (1863/1998) Utilitarianism, Oxford: Oxford University Press. Scheler, M. (1913–16/1973) Formalism in Ethics and Non-Formal Ethics of Values, trans. by M. S. Frings and R. Funk, Evanston, IL: Northwestern University Press. German edition: Formalismus in der Ethik und die materiale Wertethik (Gesammelte Werke: Band 2), Bonn: Bouvier, 2000.

51

Eugene Kelly Scheler, M. (1928/2009) The Human Place in the Cosmos, Evanston, IL: Northwestern University Press. Scheler, M. (1957) Schriften aus dem Nachlaß. Band 1: Zur Ethik und Erkenntnistheorie (Gesammelte Werke: Band 10), Bern, Munich: Francke.

Further reading P. Blosser, Scheler’s Critique of Kant’s Ethics (Athens: Ohio University Press, 1995). Blosser notes some limitations of Scheler’s understanding of Kant. M. S. Frings, The Mind of Max Scheler (Milwaukee: Marquette University Press, 1997). R. Funk (1974), “Thought, Values, and Action,” in M. S. Frings (ed.), Max Scheler (1874–1928). Centennial Essays (The Hague: Nijhoff, 1974) is one of the few essays in English on the topic of Scheler’s action theory. P. H. Spader, Max Scheler’s Ethical Personalism (New York: Fordham University Press, 2002) is the only full-length study of Scheler’s personalism in English. K. Wojtyla (1979), The Acting Person (Dordrecht: Reidel, 1979). The later Pope John-Paul II was a careful student of Scheler.

52

5 THE INTENTIONALITY AND POSITIONALITY OF SPONTANEOUS ACTS Adolf Reinach’s account of agency Francesca De Vecchi Introduction The aim of this chapter is to discuss Adolf Reinach’s account of agency and his phenomenological approach to agency. I will show that Reinach’s account of agency concerns intentional acts, rather than intentional bodily actions, and explain how it is grounded in a phenomenological account of intentional lived experiences (Erlebnisse) that are characterized by different levels of positionality. More specifically, I will explain how in his work The Apriori Foundations of the Civil Law (Reinach 1913a) Reinach outlines an account of intentional acts as spontaneous, free acts and, in doing so, presents an attractive phenomenological account of agency. I will focus on his concept of spontaneous acts (spontane Akte) understood as intentional acts distinct from both intentional states and intentional bodily actions and show in what sense Reinach’s account of spontaneous acts is a phenomenological account of free acts characterized by a sense of agency. In particular, I will discuss his claim that, unlike other intentional lived experiences, e.g. passive or active (aktiv) experiences, spontaneous acts are specifically operative (tätig) experiences characterized by an “inner doing” of the self, where the self appears as the “author” of acts. Accordingly, I will argue that the sense of agency captured by Reinach’s account of spontaneous acts is a sense of “authorship” that ought to be sharply distinguished from a sense of ownership. Finally, I will dwell on social acts as a peculiar kind of spontaneous act and highlight its distinctive features. Unlike spontaneous acts, social acts are acts in which the “inner doing” of the self is grasped by another self, and, like spontaneous acts, social acts are acts that are not intentional bodily actions. Moreover, I will show that (i) the agency of social acts requires the involvement of at least two individuals and that (ii) they are position-takings of a second level, as spontaneous acts are, or even of a higher level.

Passive, active, and spontaneous intentional experiences In the opening of the section on “social acts” of his work on the legal a priori, Reinach identifies three kinds of intentional lived experiences, namely, those characterized by (i) passivity, (ii) activity, and (iii) spontaneity. I claim that, on the basis of that distinction, 53

Francesca De Vecchi

Reinach provides a phenomenological account of intentional experiences that are marked by (i) a sense of ownership, (ii) a sense of agency, and (iii) a sense of authorship, respectively.1 Let’s examine the key passage where Reinach describes this landscape of various forms of intentional experience: Out of the infinite sphere of possible kinds of intentional experiences (Erlebnisse) let us select out a certain kind: the experiences which not only belong to a self but in which the self shows itself as operative (tätig). We turn our attention to a thing, or we make a resolution: these are experiences which are not only opposed to the ones in which something like a sound or a pain imposes itself on us, but also to the ones where we cannot speak of a real passivity of the self, as when we are happy or sad, enthusiastic or indignant, or when we have some wish or resolution and we bear them in ourselves (wenn wir einen Wunsch oder Vorsatz haben und in uns tragen).2 We want to call the experiences in question spontaneous acts; this spontaneity refers to the inner doing (das innere Tun) of the subject. It would be quite a mistake to want to find the distinguishing mark of these experiences in their intentionality. The regret which rises up in me, or the hatred which asserts itself in me, are also intentional in that both refer to some object (irgendetwas Gegenständliches). Spontaneous acts have in addition to their intentionality also their spontaneity, which lies in this, that in them the self shows itself to be the phenomenal originator (Urheber) of the act. Spontaneity also has to be definitely distinguished from activity in its many possible meanings. Thus I can call indignation active since it issues from me and forms a contrast to the sadness which creeps over me or assaults me suddenly (die Betrübnis, die mich beschleicht oder plötzlich überfällt).3 Or I call having of a resolution active in that I am the one who has the resolution. But we distinguish the having of a resolution, whether actually or inactually, from the making of the resolution, we distinguish what exists in us as a state (zuständlich) from the punctual experience, which precedes or at least can precede it; and only here in the making of a resolution do we have an example of what we mean: a doing (Tun) of the self and thereby a spontaneous act. We right away think of all kinds of examples of such acts: deciding, preferring, forgiving, praising, blaming, asserting, questioning, commanding, etc. (Reinach 1913a/1989: § 3: 158; eng. trans. 2012: § 3, 18) Reinach thus claims that spontaneous acts are a specific kind of intentional lived experience. Experiences, as intentional, are directed toward “something objective” (irgendetwas Gegenständliches). Thus, passive, active, and spontaneous experiences all count as forms of intentional experience. However, unlike passive and active experiences, spontaneous acts are marked by a spontaneity (Spontaneität) that is not reducible to any kind of “activity” (Aktivität). According to Reinach, the fact that intentional experiences can be either passive, active, or spontaneous is explained by two different factors: the role played by the self in each of them (i.e. a passive, active, or operative role) and the temporality of the experiences, whether they are enduring states or punctual acts. These factors, we will see, are grounded in a further factor that Reinach does not make explicit here, but that he had focused on in his work On the Theory of the Negative Judgment (1911), where he first hints at the phenomenon of “spontaneous acts.” This third factor emerges on the basis of the phenomenological claim that intentional lived experiences exhibit positionality in differing degrees. I suggest that while, unlike passive experiences, both spontaneous experiences and some active experiences are position-takings, spontaneous experiences alone exceed the first degree of positionality and count as free acts. 54

Adolf Reinach

In developing these three points, I will focus on (i) Reinach’s description of each of these kinds of experience, (ii) Reinach’s conception of “activity,” and (iii) the positionality of intentional experiences revealed in phenomenological analysis.

Passive intentional experiences Examples of passive intentional experiences include cases such as a sound or a pain imposing itself on one, sadness creeping over or suddenly assaulting one, and hatred fomenting within one. In these examples, passive experiences are either sensations (sound and pain) or emotions (sadness and hatred), which, as it were, occupy the self from the outside and exist independently of the self ’s initiative. The self is here a passive subject at the mercy of its experiences. It is not able to avoid or to manage them. Passive experiences belong to the self only in the sense that they happen in it or that it “lives in” them. The relation between the self and passive experiences, I argue, exhibits a sense of ownership, but no sense of agency or authorship.

Active intentional experiences According to Reinach, active intentional experiences are experiences “where we cannot speak of a real passivity of the self, like when we are happy or sad, enthusiastic or indignant, or when we have some wish or resolution and we bear them in ourselves.” These examples of active intentional experiences mentioned by Reinach are, respectively, emotions and desiderative or willing states. Reinach claims that both of these are active because the self is active in them. However, it seems to me that in these two cases the grounds on which Reinach argues that the self is active are quite different. Further clarification is needed beyond what Reinach is explicitly affirming here. Emotions can be active experiences, as with emotions that involve position-taking, i.e. the self ’s response to some experienced state of affairs. Thus, indignation is an active emotion since it issues from the self (in contrast with sadness which creeps over the self or assaults it suddenly) in response to something it encounters. Desiderative or willing states (e.g. having and bearing in ourselves a wish or resolution), on the other hand, are active experiences, since the self, Reinach says, is the one who has the resolution or the wish, and is the experience’s bearer (Träger). What exactly it means to call the self active as the bearer of such states is not clear at all and needs further explication. We will return to this below in the section on Senses of activity of the self. Setting that issue aside for the moment, let’s say that the distinctive feature of active experiences like emotions is that they do not merely belong to the self (like passive experiences do), but rather issue from the self. They are its responses to situations in which it finds itself. This feature corresponds, further, to a sense of agency had by the self in such experience. A further distinctive feature of active experiences is that they can be states in which the self persists, like in cases where the self is the subject of a wish or resolution.

Spontaneous intentional experiences: spontaneous acts Reinach claims that spontaneous experiences share four essential and interconnected features that make spontaneous experiences spontaneous acts: i ii

The “doing of the self ” in which “the self shows itself as operative (tätig)” The “punctuality” of the experience, i.e. its instantaneous temporal character 55

Francesca De Vecchi

i ii iv

The authorship of the self insofar as it “shows itself to be the phenomenal originator (Urheber) of the act” The “inner doing of the subject,” which, qua internal, need not involve intentional bodily action (see the section below on Intentional Acts and Intentional Bodily Actions).

Reinach presents as examples of spontaneous acts experiences like turning our attention to a thing, making a resolution, deciding, forgiving, praising, asserting, questioning, commanding, preferring, and blaming. It is worth noting that some of these examples are spontaneous social acts (praising, blaming, commanding questioning – see the section on Spontaneous Acts as Social Acts). Spontaneous acts are characterized both by the self appearing as the author of its inner doing and by the instantaneity of that doing, whereas active experiences are merely borne and temporally extended. Spontaneous acts existentially depend on the inner doing of the self, which is both a necessary and sufficient condition for their existence – by contrast, I would say that in active experiences the self is just a necessary and not also a sufficient condition for their existence. In spontaneous acts the self is not only an agent but also an author, the creator of its experience. The crucial point Reinach makes here is that generic activity is not enough for an intentional experience to count as a spontaneous act. Rather, what is needed in addition is that the self manifests itself as the author of the act, and, in so doing, signs off the act, so to speak. I suggest that “spontaneous act,” as intended here, involves a sense of autonomy and self-determination. The fact that I am the source and originator of the act means that what I do – making a resolution, asserting, forgiving, preferring – depends entirely on me. In other words, spontaneous acts embody the highest degree of agency. In them the self is the agent of free and voluntary acts, which are free insofar as the performance of the act depends solely upon the self (see the section on Positionality and The Phenomenological Account of “Free Acts”). It nevertheless seems to me that the grounds for Reinach’s distinction between active and spontaneous acts deserve further clarification. Why does he speak of active experiences as states, on the one hand, and of spontaneous experiences as acts, on the other hand? In what sense, on Reinach’s view, do desiderative or willing states qualify as active? How does being a “bearer of the will” make the self active? To tackle these problems, I will focus on the conception of “activity of the self,” “spontaneous act,” and “positionality” hinted at in Reinach’s work On the Theory of the Negative Judgment (Reinach 1911).

Senses of activity of the self In what sense is the self active in certain intentional lived experiences? Reinach does not provide a single, univocal definition for “activity.” He does, however, admit that “activity” is an ambiguous term and uses it to refer to an intentional lived experience that either: i ii

“issues from the self,” as, for instance, with emotions such as indignation; or has the self as its bearer (Träger), as in cases of “having of a resolution and bearing it in myself ” (which is not to be confused with making a resolution, something Reinach considers to be spontaneous and not merely active).

Concerning the second understanding of activity, it is not clear at all in what sense being the bearer of the experience makes the self active. It must be a specific sense of “bearer,” distinct

56

Adolf Reinach

from the “standard” sense (i.e. within classical phenomenology) in which the subject of intentional experiences is always their bearer.4 In what further sense is the self active, then, in being the bearer of intentional lived experiences? I suggest that Reinach uses “bearer” to refer to the self as the bearer of an intentional experience that is a state (Zustand), a quid that is zuständlich, as in the examples already mentioned, i.e. “having of a resolution and bearing it in myself,” “having of a wish and bearing it in myself.” In these cases, the self is the subject of an intentional experience that is a state. What is peculiar to this sort of case becomes apparent when it is contrasted with being the bearer of a different kind of experience such as a punctual, instantaneous act or a spontaneous act that is “a doing of the self.” Reinach thus maintains that we must distinguish between states where the self is the bearer of, say, a resolution or a wish, on the one hand, and acts in which the self is the author of, say, a resolution or a wish. One may still wonder why bearing a state should be characterized as active. He does not provide an explicit answer, but nevertheless gestures in the direction of an answer. The state of having a resolution is a state that is grounded in the punctual act of making a resolution, which “precedes it.” Consequently, the self of the state of having a resolution is an active self, not a passive self, because the self here is resolute and intent. The self finds itself in the state of having a resolution and is the bearer of this state because it is, based on its earlier act, the author of the resolution.

Positionality of spontaneous acts: position-takings on position-takings According to Reinach, intentional lived experiences are not to be distinguished solely on the basis of (i) the role of the self (as active, passive, or operative) and (ii) the temporality of the experience (as punctual acts or durable states), but, crucially, also on the basis of (iii) the positionality of the experience and on the layering and modality of the experiences. I claim that this third element of positionality is intertwined with the previous two criteria for distinguishing among intentional experiences and a deeper understanding of them. Whether the self is passive, active, or spontaneous in an intentional experience depends on the modality and degree of its positionality, which can also determine whether the experience is a state or an act. In On the Theory of Negative Judgment (1911), two years before his account of spontaneous acts in The A Priori Foundations of the Civil Law, Reinach had already introduced the concept of “spontaneous act.” There he had also focused on the problem of how to distinguish between states and acts as forms of intentional lived experience and hinted at the idea that positionality involves a layering of intentional experiences. In his 1911 work, he clarifies the concept of “ judgment” (Urteil) and differentiates two kinds of judgment, namely, judgment as conviction (Überzeugung), which is a state, and judgment as assertion (Behauptung), which is an act, even a “spontaneous act.” Let us focus on this distinction, since it nicely illustrates the crucial place of positionality in Reinach’s account of spontaneous acts: We can [. . .] see with indubitable clarity that conviction or belief on the one hand, that which emerges in us in the presence of a particular object (angesichts der Gegenstände erwächst), always involves something that people describe sometimes as a feeling and sometimes possibly even as a situation of consciousness, but which is in any event a state

57

Francesca De Vecchi

of consciousness (etwas was man mitunter als Gefühl, mitunter wohl auch als Bewusstseinlage bezeichnet hat, jedenfalls eine Zuständlichkeit des Bewusstsein); but that assertion, on the other hand, which does not “emerge” within us but is rather “made” by us (von uns “gefällt” wird), is totally different from any feeling, from every state, and is much rather to be characterized as a spontaneous act. (Reinach 1911/1989: 99; eng. trans. 1982: 320; translation modified) Conviction and assertion, as Reinach sees things, are both kinds of judgment. What sets the two apart is that conviction is a state, one that may or may not involve a sensory component, which wells up in us when faced with objects we have apprehended, and whose emergence takes time. Hence, conviction is here identified by Reinach as a position-taking of the first degree, grounded in apprehension, and as such it qualifies as an active experience. Assertion, on the other hand, is a judgment that does not emerge or develop, but is “delivered and expressed by us” (gefällt von uns) spontaneously. It is, therefore, not a state at all, but rather an act, and a spontaneous one, Moreover, Reinach carefully articulates in his description the layering of apprehension, conviction, and assertion as different kinds of experience with different corresponding degrees of positionality. It is worth noting that apprehension has a zero degree of positionality: it is the basic cognitive intentional experience, the most fundamental one, on which position-takings such as conviction and assertion can be grounded:5 In apprehension the state of affairs is presented to me, and on the basis of the apprehension there develops in me the conviction on, or belief in, that state of affairs. Conviction is in this case founded in apprehension; the former is the position which I take up, my receipt, so to speak, for that which apprehension offers to me. (Reinach 1911/1989: 120; eng. trans. 1982: 344) Conviction, as Reinach explains here, is a position-taking of the first degree, grounded in an apprehension of some state of affairs. As a position-taking with respect to that, apprehended, state of affairs, the conviction is, as it were, a receipt for the apprehension, a lasting testament to what is experienced in apprehension: On the other hand [. . .] a state of affairs of which we remain convinced can become re-posited in an act of assertion. We have already seen that at the basis of every assertion there lies a conviction. (Reinach 1911/1989: 125; eng. trans. 1982: 355) Assertion is a position-taking of the second degree that is grounded in conviction. In the act of asserting, the state of affairs initially apprehended and subsequently consolidated in conviction is then posited again and confirmed. Therefore, assertion is a position-taking about a position-taking, which is what makes it a spontaneous act, i.e. an inner doing of the self that is freely performed by the self. In other words, nothing external compels the self by necessity to perform the act of assertion. The grounding relation between assertion and conviction is a relation of motivation, where the self can – but need not necessarily – make conviction into a motive for performing the assertion. It is in just this sense that the self is a spontaneous, free author of its act of asserting. This is also why I said earlier that spontaneous acts are marked by a sense of agency more narrowly conceived as a sense of authorship. 58

Adolf Reinach

Positionality and the phenomenological account of “Free Acts” I shall now turn to the question how Reinach’s account of spontaneous acts is in keeping with and, indeed, is part of the theory of intentionality developed by other figures within classical phenomenology like the fellow Munich scholar Alexander Pfänder, Edith Stein, who worked under Reinach, and Edmund Husserl himself, the founder of the “phenomenological movement” and under whom Reinach worked in Gottingen.6 The general thrust of this approach is not only that intentional experiences have different levels of positionality but also that the layering of intentional experiences of various degrees of positionality constitutes the intentional life of persons.7 In other words, the key to fully grasping the full theoretical import of Reinach’s account of spontaneous acts is to recognize that it is part of the broader phenomenological account of intentionality as positionality. On that view, spontaneous acts represent the highest level of position-taking, that of the free acts of persons, i.e. the level of personhood.8 Reinach’s treatment of spontaneous acts, Pfänder’s treatment of the act of willing, and Stein’s treatment of free acts all make the same point, namely, that there is a subset of intentional lived experiences comprised specifically of the self ’s acts and position-takings of the second degree and beyond this. Doubtless, in developing his account of spontaneous acts, Reinach had in mind Pfänder’s account of willing. Like spontaneous acts as understood by Reinach, Pfänder considers acts of willing to be acts originally performed by the self and not merely issued from the self. In the act of willing, he says, “the ego-center is not only the subject and origin but also the original performer of the act (der originäre Vollzieher des Aktes)” (Pfänder 1911/1963: 133; eng. trans. 1967: 20). Moreover, like spontaneous acts, acts of willing also existentially depend only upon the self who is their initiator. In his words: “phenomenally, the act of willing appears precisely not as an occurrence caused by a different agent but as an initial act of the ego-center itself ” (Pfänder 1911/1963: 133; eng. trans. 1967: 20). Finally, just as spontaneous acts are characterized by the inner doing of the self in which the self shows up as the author of the act, acts of willing, too, are marked by “a peculiar doing in which the ego-center centrifugally, from inside itself, performs a mental stroke (geistiges Schlag)” (Pfänder 1911/1963: 135; eng. trans. 1967: 22; see also Stein 1922/1970: 53). Turning now to Stein, she refers explicitly to both Pfänder and Reinach in her work on the motivational life of persons, in which she takes the highest level of positionality to be that of “free acts” (Stein 1922/1970: 46). Stein characterizes free acts as “acts in which the self not only experiences, but emerges as the master of its experience (Akte in denen das Ich nicht nur erlebt, sondern als Herr seines Erlebens aufritt).” Regarding free acts, she writes that “I must bring them forth from within myself, as if I were mentally hitting out (Ich muss sie aus mir heraus erzeugen, gleichsam geistig einen Streich führen)” (Stein 1922/1970: 46; my translation). Only those acts that “the self carries out from within itself (die das Ich aus sich heraus vollzieht)” are free. To better clarify the nature of spontaneous, free acts it will help to follow Stein’s analysis and then compare it to Reinach’s. Like Reinach, Stein also contrasts spontaneous, free acts, on the one hand, and lived experiences that are not spontaneous, on the other hand. For instance, sensations and emotions are not spontaneous acts since one cannot help but undergo them. One cannot avoid feeling a pain when knocking one’s leg against a table or avoid feeling sad when receiving bad news. Sensations and certain kinds of emotions are precisely “passive experiences,” as Reinach calls them, which stand out as such only by the sense of ownership accompanying them.9 However, the crucial point that Stein makes is that one can take a position concerning these experiences, one can manage them. That is, one can accept 59

Francesca De Vecchi

or endorse (annehmen) them or withhold attention from them, even repressing them (ablehnen), if one likes (see Stein 1922/1970: III, § 3: 42–46). This constitutes a first level of position-taking corresponding to what Reinach’s calls active experiences, I suggest. Further, one can take a position with respect to what one has endorsed or ignored in a prior position-taking and can turn it into a ground or motive for further acts. This is the level of spontaneous, free acts. They are second-order (or higher) position-takings. They make up the sphere of experience requiring a self and depending on nothing beyond it. Again, it must be observed that the motives on which spontaneous, free acts are grounded do not force the self to perform them; as Stein points out, “the self can also abstain from them (das Ich kann die Motive haben [. . .] und kann die Akte ebensogut unterlassen)” (Stein 1922/1970: 48–49; my translation). That is why the “founding” (Fundierung) relation of spontaneous, free acts as resting on lower-order position-takings is a relation of motivation and neither a necessitating nor a causal relation.10 In the light of this layering of intentional experiences and position-takings through which the life of persons as subjects of spontaneous-free acts becomes intelligible, it is now possible to more adequately understand the sense of agency as authorship distinctive of specifically spontaneous acts, as I have endeavored to show from the outset. Position-takings of the first level are, of course, characterized by a sense of agency and not merely by a sense of ownership, which is the mark of lower-order intentional experiences like sensations and certain emotions. However, only position-takings of the second level (or higher), that is, only spontaneous, free acts exhibit the strongest sense of agency. Indeed, the latter involves at the same time a sense of authorship. The existence of the act depends upon the self freely and willingly playing the role of the act’s author.11

Intentional acts and intentional bodily actions What I have examined so far is Reinach’s account of spontaneous acts as free acts of persons, acts representing the highest level of positionality and marked by a sense of agency as authorship. I would like to dwell now on the fact that, according to Reinach, spontaneous acts are not bodily actions. He speaks of spontaneous acts simply as intentional lived experiences involving an internal doing of the self, without mentioning any accompanying bodily action. I suggest that Reinach’s distinction between spontaneous acts and intentional bodily action is consistent with the phenomenological account of intentional lived experiences provided by Husserl (1901) in his Fifth Logical Investigation (§§10, 12–13) according to which intentional lived experiences (Erlebnisse) are not actions in the sense of intentional bodily actions. Husserl explains that his concept of “act” excludes the common and original meaning of “act” (actus) as a form of “activity” (Betätigung) performed in the world.12 However, this fundamental connection does not exclude an important difference between Husserl and Reinach regarding the way of intending “act” and “intentionality” as “positionality.” In contrast to Husserl, Reinach holds that acts are only those intentional experiences marked by a spontaneous doing of the self and by position-takings of a second degree, at least. As we have seen, Reinach sharply distinguishes spontaneous, free acts from any other kind of intentional lived experience: from active lived experiences that are position-taking of first degree, on the one hand, and from passive intentional live experiences as well as from basic cognitive lived experiences (like apprehension or perception) that are both position-takings of zero degree, i.e. no position-takings, on the other hand.

60

Adolf Reinach

For Husserl, on the contrary, any intentional lived experience is per se an act. Husserl uses “act” interchangeably with “intentional lived experience” because in any such experience the self is directed toward and faced with something: thus, according to Husserl, any intentional lived experience – even basic cognitive experiences like perceptions – involves position-taking since the self responds in different ways, i.e. by occupying different levels of positionality, to the experienced object. The response may be an intentional lived experience that is a free act of the person like a decision and that is marked by the highest degree of positionality: it is a position-taking on position-takings; but the response may also be an intentional lived experience that is characterized by the lowest degree of positionality like a perception that does not presuppose other position-takings; nevertheless, Husserl’s point is that, even in such intentional lived experiences that have the lowest degree of positionality, the intentional relation between the self and the experienced object is still characterized by positionality. For instance, in the case of perception the self still takes a position in so far as it takes the experienced object as actually existing – and not as a non-existent object as in hallucination. This is why Husserl claims that “all life is position-taking” (Alles Leben ist Stellungnehmen) (Husserl 1911/1984: 56; eng. trans. 2002: 290).13 It is worth noticing that this way of equating “intentional lived experience,” “act,” and “positionality” implies that when Husserl and other phenomenologists like Stein refer to second-degree (or higher) position-takings, they must explain that they are using the term “act” stricto sensu, and not merely as a synonym for “intentional lived experience.” Stein, for instance, calls them “free acts” (see Stein 1922/1970: 46). Anyway, apart from that divergence, it seems to me that the crucial point is that both Reinach and Husserl agree on the view that there is a specific kind of intentional lived experience consisting of (at least) second-order position-takings, i.e. free acts of persons. Now, what about the relation between spontaneous, free acts (i.e. acts stricto sensu, in the sense just explained), on the one hand, and intentional bodily actions, on the other hand? Actions are goal-directed intentional bodily movements. Such bodily movements are intended with the aim of satisfying a goal specified in a spontaneous, free act. In other words, what are commonly called “intentional actions” can arise from intentional acts and fulfill their content, as long as both have the same content. For instance, the action of doing X fulfills various intentional acts like deciding to X, preferring to X, or being commanded to X. Intentional actions involving bodily movement are characterized by a sense of agency absent in mere bodily behavior, which are nevertheless accompanied by a sense of ownership.14 Of course, the relation between intentional acts and intentional bodily actions is a motivational one that is neither necessitating nor causal. As a relation of motivation, it is a moment of the fundamental phenomenological idea of intentionality as positionality.

Spontaneous acts as social acts As I have just shown, in Reinach’s account of spontaneous acts, the doing of the self characteristic of spontaneous acts is an internal doing, and spontaneous acts are not bodily actions. Nevertheless, not all spontaneous acts are solely internal acts. Reinach introduces a crucial distinction among spontaneous acts and affirms that spontaneous acts include a subset of acts that are not internal, but external. Such acts are external in the sense that they have to be turned toward, communicated to, and grasped by another person: they are social acts. Therefore, spontaneous acts can also be social acts. Indeed, in Reinach’s work spontaneous social acts play a more significant role than simple spontaneous acts. He focuses on spontaneous acts

61

Francesca De Vecchi

precisely in order to highlight the kind of intentional lived experience to which social acts belong. In the following passage he describes the nature of such social acts: In looking more closely at these cases [of spontaneous acts (deciding, preferring, forgiving, praising, blaming, asserting, questioning, commanding, etc.)], we right away notice an essential difference, and this is the difference which is important for us here. The act of deciding is an internal act. It can be performed without being announced (verlautbar) or needing to be announced. [. . .] It can unfold entirely within, it can rest in itself and not receive an expression in any sense. One sees right away that it is otherwise with certain other spontaneous acts. Commanding or requesting, for instance, clearly cannot be performed entirely within. [. . .] Commanding [. . .] announces itself in the act of turning to the other (sich wenden an), it penetrates the other (dringt in den anderen ein), and has by its very nature a tendency to be heard (vernommen) by the other. [. . .] The command is according to its essence in need of being heard (vernehmungsbedürftig). [. . .] We designate the spontaneous acts which are in need of being heard, social acts. [. . .] Our concept of social acts centers only on the need of being grasped. (Reinach 1913a/1989: § 3; eng. trans. 2012: 18–19). Acts such as praising, blaming, questioning, commanding, as well as promising, informing, promulgating a law (acts not mentioned by Reinach in this passage) are spontaneous social acts, and not merely spontaneous acts. In spontaneous social acts the doing of the self is not merely internal. It is turned toward another self, the addressee of the act, and needs to be expressed in order to be grasped by the addressee. In describing the role of the addressee, Reinach uses the German expressions vernehmen and Vernehmungsbedürftigkeit. The idea is that the addressee must perceive the spontaneous doing of the agent as a necessary condition for the performance of a social act. We can also translate vernehmen by “taken up” and Vernehmungsbedürftigkeit by the “necessity of being taken up.” Social acts must be taken up by their addressee just as “speech acts” do. Indeed, Reinach’s treatment of social acts and J. L. Austin’s treatment of speech acts about fifty years later target the same phenomena.15 It should also be observed that, like spontaneous acts, spontaneous social acts need not involve intentional bodily actions (for instance, think of social acts such as praying, asking, informing, promising: they do not need any bodily action at all), but, unlike spontaneous acts, they are acts that emerge from an agent’s mind in order to reach another individual, the addressee of the act, by whom the act is to be grasped (see Reinach 1913a/1989: § 3; eng. trans. 2012: 20).16 For considerations of space, I cannot dwell here on Reinach’s account of the essential features of social acts and their role in the theory of the a priori foundations of law. I nevertheless want to make just two points about the nature of social acts as spontaneous.17 Both concern the agency of spontaneous social acts. One regards how agency is “modified” in social acts and the other concerns social acts as (at least) second-order position-takings. First, in spontaneous social acts, the inner doing of the self is turned toward another person, the addressee of the act, who must perceive it for the act to be performed. Indeed, social acts are “heterotropic.”18 The addressee of the social act is a counterpart to the act’s author and must validate, as it were, the author’s spontaneous doing. There is no such requirement for simple (i.e. non-social) spontaneous acts. That is why social acts are not internal acts as spontaneous acts are. They are, rather, acts whose completion involves an external moment,

62

Adolf Reinach

i.e. the intentionality of the addressee. Social acts existentially depend on the addressee’s uptake. Importantly, that uptake, I suggest, need not itself be a spontaneous act. It need only be an intentional lived experience in which the addressee apprehends the act. On Reinach’s account, this uptake is the sort of experience we referred to earlier as a basic cognitive experience that has a zero degree of positionality. In it the addressee apprehends the author’s act. The addressee can then take a position on that act, endorsing or rejecting it, and can, moreover, take yet another position concerning it, for instance, deciding to respond to a question, if the social act was one of posing a question, to obey a command, if it was one of issuing a commanding.19 In other words, the addressee’s uptake is not per se a positiontaking, because it might be merely an act of apprehension (see the section above Positionality of Spontaneous Acts). Second, social acts are essentially (at least) second-order position-takings. Hence, they exhibit the highest level of positionality that spontaneous acts can occupy and enrich our understanding of that level. As Reinach emphasizes, social acts are necessarily grounded in other intentional experiences, ones that have the same content as the corresponding social acts yet are not themselves social acts: We now have to take note of the remarkable fact that all social acts presuppose [. . .] internal experiences. As a matter of apriori necessity (wesensgesetzlich) every social act presupposes as its foundation some internally complete experience whose intentional object coincides with the intentional object of the social act or is at least somehow related to it. Informing presupposes being convinced (Überzeugung) about what I inform someone of. [. . .]. Commanding presupposes as its foundation not only the wish but the will that the one who is commanded carry out my command; [. . .] Like all social acts, promising presupposes an inner experience which has the content of the promise as its intentional object. As with commanding, this inner experience is that of intending (Wille) that something occurs, not of course through the addressee but trough the promisor himself. (Reinach 1913a/1989: 162, 164–165; eng. trans. 2012: 21–22, 25–26) It is worth noting that, in Reinach’s analysis, social acts like commanding and promising are third-order, not second-order, position-takings. Both commanding and promising are grounded in the will that something should occur through the addressee’s (in commanding) or agent’s (in promising) subsequent action. The will that something should occur is already a second-order position-taking with respect to a first-order position-taking such as a belief about some previously apprehended state of affairs. Finally, it is also crucial to highlight that, as Reinach explicitly affirms in speaking of “a matter of a priori necessity” or more precisely “a matter of essential lawfulness” (a more literal translation of his expression, wesensgesetzlich), the relation between intentional experiences of different levels of positionality – that is, the layering of intentional experiences – is not a contingent relation, but an essential one. It is a distinctively motivational relation, one that must be carefully distinguished from contingent empirical relations. Reinach’s entire approach to understanding social acts and the social world is marked by the distinctively phenomenological claim that there is a “material a priori,” understood as essential constraints that define any kind of entity as such.20 The specific contribution of Reinach’s work is the extension of that idea to social reality and the discovery of such material a priori in the social world.21

63

Francesca De Vecchi

Related topics Chapters 2 (on Pfänder & Husserl), 3 (on Pfänder), 9 (on Stein), 24 (De Monticelli).

Notes

64

Adolf Reinach of at least two persons to occur. On social acts as heterotropic acts, see De Vecchi (2014) and De Vecchi and Passerini (2012). 19 On the zero degree of positionality of the addressee’s uptake in Reinach, see De Vecchi (2020). 20 For the sake of brevity, I cannot develop the crucial phenomenological point concerning the “material a priori,” and I just limit myself to refer to the most fundamental source, namely Husserl’s Third Logical Investigation (1901). 21 Reinach’s A Priori Foundations of Civil Law is certainly a work of social ontology. See Mulligan (1987, 2016), Benoist (2005), De Vecchi (2012, 2013, 2014, 2016a, 2016b), and two conferences devoted to Reinach in 2017: “The Philosophy of Adolf Reinach,” 30 November–1 December 2017, Universities of Lugano and Milan; “Reinach centennial conference,” 14–16 December 2017, University of Munich.

References Austin, J. L. (1962) How to Do Things with Words, Oxford: Oxford University Press. Benoist, J. (ed.) (2005) “Adolf Reinach: philosophie du langage, droit, ontologie,” Les études philosophiques 2005/1. Bratman, M. E. (1987) Intention, Plans and Practical Reason, Stanford: CSLI Publications. De Monticelli, R (2011) “Alles Leben ist Stellungnehmen,” in V. Meyer, C. Erhard and M. Scherini (eds.), Die Aktualität Husserls, Freiburg: Alber, pp. 39–55. De Monticelli, R. (2015) “Outline of a Theory of Embodied Rationality,” Les Cahiers philosophique de Strasbourg II, 175–193. De Monticelli, R. (2020) “The Phenomenology of Rational Agency,” this volume. De Vecchi, F. (ed.) (2012) Eidetica del diritto e ontologia sociale. Il realismo di Adolf Reinach/Eidetics of Law and Social Ontology. Adolf Reinach, the Realist, Italian and English Edition, Milano-Udine: Mimesis. De Vecchi, F. (2013) “Ontological Dependence and Essential Laws of Social Reality. The Case of Promising,” in M. Schmitz et al. (eds.), The Background of Social Reality, Dordrecht: Springer, pp. 233–251. De Vecchi, F. (2014) “Three Types of Heterotropic Intentionality. A Taxonomy in Social Ontology,” in A. Konzelman Ziv and H. B. Schmid (eds.), Institutions, Emotions and Group Agents. Contribution to Social Ontology, Dordrecht: Springer Verlag, pp. 117–137. De Vecchi, F. (2016a) “A priori of the Law and Values in the Social Ontology of Wilhelm Schapp and Adolf Reinach,” in H. B. Schmid and A. Salice (eds.), The Phenomenological Approach to Social Reality. History, Concepts, Problems, Dordrecht: Springer-Verlag, pp. 279–316. De Vecchi, F. (2016b) “The Existential Quality Issue in Social Ontology: Eidetics and Modifications of Essential Connections,” Humana.Mente 31, 187–204. De Vecchi, F. (2020) “Proxy Social-speech Acts: A Particular Case of Plural Agency,” Language and Communication 71, 149–158. De Vecchi, F. and Passerini, L. (2012) “La tipologia degli Erlebnisse e degli atti spontanei in Adolf Reinach,” in F. De Vecchi (ed.), Eidetica del diritto e ontologia sociale. Il realismo di Adolf Reinach/Eidetics of Law and Social Ontology. Adolf Reinach, the Realist, Italian and English Edition, Milano-Udine: Mimesis, pp. 261–280. Husserl E. (1901/1984) Logische Untersuchungen, Zweiter Teil. Untersuchungen zur Phänomenologie und Theorie der Erkenntnis, ed. by U. Panzer, The Hague, Netherlands: Martinus Nijhoff (=Hua XIX, Vol. 2); eng. transl. by. J. N. Findlay (2001), Logical Investigation. Second Part. Investigations Concerning Phenomenology and the Theory of Knowledge, London and New York: Routledge. Husserl E. (1911/1987) “Philosophie als strenge Wissenschaft,” in E. Husserl (1987), Th. Nenon und H. R. Sepp (eds.), Aufsätze und Vorträge 1911–1921, The Hague, Netherlands: Martinus Nijhoff (=Hua XXV); eng. transl. by M. Brainard (2002), “Philosophy as a Rigorous Science,” The New Yearbook for Phenomenology and Phenomenological Philosophy 2, 249–295. Husserl E. (1912–28/1952) Ideen zu einer reinen Phänomenologie und phänomenologischen Philosophie. Zweites Buch: Phänomenologische Untersuchungen zur Konstitution, ed. by M. Biemel, The Hague: Martinus Nijhoff (=Hua IV); eng. transl. by R. Rojcewicz and A. Schuwer (1989), Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Book 2: Studies in Phenomenology of Constitution, The Hague: Martinus Nijhoff. Gallagher, S. (2000) “Self-reference and Schizophrenia: A Cognitive Model of Immunity to Error through Misidentification,” in D. Zahavi (ed.), Exploring the Self: Philosophical and Psychopathological Perspectives on Self-Experience, Amsterdam: John Benjamins, pp. 203–239.

65

Francesca De Vecchi Gallagher S. and Zahavi, D. (2008) The Phenomenological Mind: An Introduction to Philosophy of Mind and Cognitive Science, New York: Routledge. Mulligan, K. (ed.) (1987) Speech Act and Sachverhalt. Reinach and the Foundation of Realist Phenomenology, Dordrecht: Kluwer. Mulligan, K. (2014) “Knowledge First – A German Folly?,” in J. Dutant, D. Fassio and A. Meylan (eds.), Liber amicorum Pascal Engel, Geneva: University of Geneva, pp. 380–400. Mulligan, K. (2016) “Persons and Acts – Collective and Social. From Ontology to Politics,” in A. Salice, H.-B. Schmid (eds.), The Phenomenological Approach to Social Reality. History, Concepts, Problems, Dordrecht: Springer, pp. 17–45. Pfänder, A. (1911/1963) “Motive und Motivation,” in A. Pfänder (ed.), Münchener Philosophische Abhandlungen, Leipzig: Barth, pp. 163–195. Eng. trans. “Motives and Motivation,” in H. Spiegelberg (ed.) (1967), Phenomenology of Willing and Motivation and Other Phaenomenologica, Evanston: Northwestern University Press, pp. 12–40. Reinach, A. (1911/1989) “Zur Theorie des negativen Urteils,” in Adolf Reinach (1989), Sämtliche Werke. Textkritische Ausgabe in 2 Bänden: Band I, ed. by K. Schuhmann and B. Smith, Munich: Philosophia Verlag, 95–140. Eng. trans. by B. Smith in B. Smith (1982) (ed.), Parts and Moments: Studies in Logic and Formal Ontology, Munich: Philosophia, pp. 289–313. Reinach, A. (1913a/1989), “Die apriorischen Grundlagen des bürgerlichen Rechts,” in A. Reinach (1989), Sämtliche Werke, ed. by K. Schuhmann and B. Smith, Munich: Philosophia Verlag, pp. 141–278. Eng. trans. The A priori Foundations of the Civil Law, in J. Crosby (2012) (ed.) The A priori Foundations of the Civil Law. Along with the Lecture “Concerning Phenomenology” with an introduction by A. McIntyre, Berlin: De Gruyter, pp. 1–142. Reinach, A. (1913b/1989) “Einleitung in die Philosophie,” in A. Reinach (1989) K. Schuhmann and B. Smith (eds.), Sämtliche Werke, Munich: Philosophia, pp. 369–214. Schuhmann, K. (1993) “Adolf Reinach und Edith Stein,” in R. Luzius Fetz (ed.), Studien zur Philosophie von Edith Stein, Freiburg: Alber, pp. 53–88. Searle, J. R. (2001) Rationality in Action, Oxford: Oxford University Press. Searle, J. R. (2010) Making the Social World: The Structure of Human Civilization, Oxford: Oxford University Press. Smith B. (1982) “Introduction to Adolf Reinach ‘On the Theory of the Negative Judgment’,” in B. Smith (ed.), Parts and Moments. Studies in Logic and Formal Ontology, Munich: Philosophia, pp. 289–313. Smith B. (1990) “Towards a History of Speech Acts,” in A. Burkhardt (ed.), Speech Acts, Meanings and Intentions. Critical Approaches to the Philosophy of John R. Searle, Berlin, New York: De Gruyter, pp. 29–61. Stein, E. (1922/1970) “Beiträge zur philosophischen Begründung der Psychologie und der Geisteswissenschaften,” in E. Stein (1922), Beiträge zur philosophischen Begründung der Psychologie und der Geisteswissenschaften – Eine Untersuchung über den Staat, Tübingen: Niemeyer, pp. 2–284.

Further reading The special issue of the journal Philosophie devoted to “Adolf Reinach” (ed. D. Pradelle 2016, n. 128) contains important contributions to Reinach’s account of intentionality and proxy-social acts. A. Reinach (1913–14/1989), “Die Überlegung: ihre ethische und rechtliche Bedeutung”/“Reflection: Its Ethical and Legal Significance,” in A. Reinach, Three Texts in Ethics – A German-English Parallel Edition, transl. J. Smith and M. Lebech, intro J. Smith, foreword by A. Salice (Munich: Philosophia, 2017). This essay is a fundamental source for Reinach’s account of position-takings. E. Stein, Eine Untersuchung über den Staat (Edith Stein Gesamtausgabe, vol. VII), ed. I. Riedel-Spangenberger (Freiburg: Herder, 1925/2006); English transl. M. Sawicki, An Investigation Concerning the State (Washington, DC: ICS Publications, 2006). This treatise represents an impressive development of Reinach’s account of spontaneous social acts from the perspective of an account of law-making acts as proxy and collective social acts.

66

6 DIETRICH VON HILDEBRAND ON THE WILL AND INTENTIONAL AGENCY* Alessandro Salice

Introduction The window of your office is open and you look outside: it is a sunny and, as usual, busy day. A lot of cars are parked on the street, the wind moves the branches of the trees, people are walking and talking to each other. One of them stumbles, loses her balance, and falls. But she rapidly stands up again and proceeds on her way. Luckily, she doesn’t seem hurt. The scent of the blossoming trees reminds you that spring has finally arrived. Even this brief glance outside the window allows you to perceive a rich multitude of different things. Some of them are objects with their properties: the shining sun, the red car parked near the blue one. Others are actions. So, for instance, the guy with the black handbag, who walks down the street, is acting. The two people talking to each other are acting, too. Interestingly, perceiving actions differs from perceiving other kinds of events. One of the elements that make the perception of actions somewhat peculiar is that when you perceive actions, you apparently ascribe mentality to their agents. In fact, you won’t describe the movements of the branches as actions: not only the branches are merely moved by the wind but also trees, in all likelihood, do not belong to the sorts of creatures that have the resources required for exercising intentional agency. But humans generally do have those resources. You know that because you have performed actions yourself: you have made decisions in the past and acted on them – more or less successfully. So, you know the difference between things that happen to you and actions that you perform. For instance, that person’s stumbling does not qualify as an action – at least at first glance. True, the bodily movements that you perceive do not univocally reveal to you whether they constitute an action or not. In this particular scenario, it may be legitimate of you to assume that the person has fallen unintentionally, but you could not exclude that, at least in principle, those movements do constitute an intentional action after all: e.g. the person could have fallen on purpose. But in order to make this assumption, you again would have to ascribe a certain mental state to her. And this raises the question of what mental state turns a set of bodily movements into an intentional action. That is, what kind of mental states should agents be in for their movements to constitute an action? Ordinary language offers a straightforward answer to that question: a person φs intentionally, if that person intends or wants to φ. But what do we mean when we say of a person that she wants to φ? 67

Alessandro Salice

On the one hand, it seems clear that, by using the verb “to want,” we ascribe a certain mental state to a person. Just as when we say “Pam is sad,” we do ascribe an emotional state to Pam, namely, sadness. On the other hand, it is noteworthy that the English language has no noun that corresponds to the predicate “to want.” As Harry Frankfurt puts it: “It is perhaps acceptable, albeit graceless, to speak in the plural of someone’s ‘wants.’ But to speak in the singular of someone’s ‘want’ would be an abomination” (Frankfurt 1971: 7). In personal conversation, Kevin Mulligan suggested that precisely this linguistic peculiarity of the English language may represent one of the reasons why large parts of English-speaking philosophy of action assume that the pivotal notion in the analysis of conation is the one of desire. As a paradigmatic example of this tendency, one can precisely adduce Frankfurt, who in his seminal article of 1971 on Freedom of the Will and the Concept of a Person writes: To identify an agent’s will is either to identify the desire (or desires) by which he is motivated in some action he performs or to identify the desire (or desires) by which he will or would be motivated when or if he acts. An agent’s will, then, is identical with one or more of his first-order desires [. . .] the notion of the will [. . .] is the notion of an effective desire. (Frankfurt 1971: 8, my emphasis) Sure, on Frankfurt’s view, we also have the capacity to form second-order volitions, which is a capacity that distinguishes persons from mere “wantons.”1 But these volitions of higher order are also desires. Put another way, Frankfurt and many others deny that willing is a psychological category on its own right, for willing is or can be traced back to desire. Mulligan’s suggestion gains traction the moment one realizes that philosophers of action trained in languages which do have a noun corresponding to the verb “to want” not only carefully distinguish between desire and will but also put the will – and not desire – in the central position of conation. This is most visible in the German philosophical tradition. The German language ordinarily employs nouns such as “der Wille” or “das Wollen” to refer to the mental state one ascribes to a person when one says “she wants to φ.” It should hence not come as a surprise that, in German philosophy, entire books have been devoted to the will (think of Schopenhauer’s The World as Will and Representation) and even well-known philosophical mottos mention the will (think of Nietzsche’s The Will to Power).2 Arguably, one of the most detailed treatments of the will and of its role in intentional action can be found in Austro-German Phenomenology: questions about the will, its main features, and its relations with other experiences have occupied members of the phenomenological movement from its very beginning in the first decades of the 20th century (for a reconstruction of this debate, see Salice 2018). One of the first phenomenologists who have explicitly tackled that bundle of questions and, more generally, the nature of intentional action is Dietrich von Hildebrand. Some of his most interesting ideas on this topic can be found in his dissertation, which the young phenomenologist submitted in 1912 at the University of Göttingen under the supervision of Edmund Husserl. A revised version of this dissertation was published four years later in the third volume of the Jahrbuch für Philosophie und phänomenologische Forschung with the slightly revised title Die Idee der sittlichen Handlung (The Idea of Moral Action). This is a complex piece of written work, both from a systematic and a historic point of view, and a highly representative one of the short but extraordinarily prolific Göttingen period within the history of phenomenology.3 This chapter aims at reconstructing Hildebrand’s contributions to the philosophy of action and, in particular, his view about the will as presented in his dissertation. What makes 68

Dietrich von Hildebrand

this work unique is the attempt to develop a theory of the will within the broader framework of an original theory of intentionality. Against this background, the chapter is organized in two parts. In the first, I lay out Hildebrand’s theory of intentionality by mainly focusing on his distinction between cognitive experiences and attitudes. In the second part, I will discuss his phenomenology of the will and its relevance for a definition of the notion of intentional action. In the conclusion, I shall come back to Frankfurt and show that, based on Hildebrand’s understanding of the will, many of his considerations about ‘A wants to X’ do not apply if the predicate ‘to want’ expresses a state akin to the ‘Wille’ in Hildebrand’s sense.

Hildebrand’s theory of intentionality4 Hildebrand’s theory of intentionality is characterized by a sharp distinction between what he calls Kenntnisnahmen (cognitions) and Stellungnahmen (attitudes).5, 6 Let’s start with the first concept – what are cognitions? Back to the example: by looking outside the window, you are immediately acquainted with a multitude of objects.7 You see colors and shapes with a certain dimension. You hear sounds. You smell the perfume of the blossoming flowers, but you also feel the delicacy of the scent (that is, you feel a certain value). Interestingly, all these experiences enter specific correlative relations with their intentional objects: seeing can only be directed at colors and shapes, hearing can only be directed at sounds and noises, smelling can only be directed at olfactory phenomena like perfumes, and feeling (Fühlen) – sometimes also called “value-ception (Wertnehmen)” – can only be directed at values (see also Reinach 1989a for these correlations). All these experiences secure an “objectual having (gegenständliches Haben)” of the object in the sense that their objects are merely “had” by the subject. The same can be said for experiences that target more complex kinds of objects, like states of affairs. In fact, while watching from the window, you also perceive that one car is blue, that flowers are blossoming, or that this car is parked next to the red one. To put this differently, your visual field is not only made up of qualitative objects (nuances of colors, shapes, sounds, values), but it also includes states of affairs such as the being-blue of the car or the spatial relation between one car and the other. And you can be acquainted with these states of affairs, too, by ‘having’ them in corresponding experiences. Hildebrand contends that, despite their differences, all these experiences have something in common (and therefore all qualify as “cognitions”), which is that they are entirely passive. The subject, when living through them, is not active in any sense (Hildebrand 1969: 11). Note that, by emphasizing the passivity of the subject, Hildebrand draws the attention to two phenomenological features that qualify cognitions. The first feature concerns the kind of intentional relation that these experiences enter with their objects: cognitions, Hildebrand argues (but this aspect will become fully clear only after the discussion about attitudes), do not establish an ‘ideal’ relation with their objects (Hildebrand 1969: 10). In fact, they are merely a “consciousness of (Bewußtsein von)” objects. By contrast, attitudes do enter such ideal relations with their objects – e.g. you are enthused over the spring’s onset, you are angry at the guy honking at the other car, you are afraid for the person that fell. I will come back to this feature at the end of this section, but for the moment it suffices to pinpoint that all these prepositions (“over,” “at,” “for”) express a particular way in which the subject is ‘ideally’ related to their objects, which is different from the mere ‘of-ness’ of cognitions (on this point, see Müller 2018). The second feature about the phenomenology of cognitions is that, when the subject cognizes something (in the sense of having cognitions), she is not “doing” anything. It may 69

Alessandro Salice

well be that the subject’s cognitive system is “active” (in a broad sense), for instance, it may well be that the subject’s cognitive systems are computing all sorts of inferences and predictions, but these activities are sub-personal and thus not phenomenologically detectable by the subject. The subject is not first-personally aware of them and even cannot be first-personally aware of them. Hildebrand goes as far as stating that, even if one can correctly describe the perceived states of affairs as “formed (Geformtsein)” or “organized (gegliedert),” this should not mislead one into postulating a “forming” or “organizing” activity on the side of the subject: “all this [talking] refers to purely objective distinctions (rein gegenständliche Unterschiede)” that affect states of affairs (Hildebrand 1969: 12). Before moving over to attitudes, it may be important to touch upon another experience of the cognitive kind that plays a crucial role in Hildebrand’s philosophy of mind and epistemology. When you come to know something, you live through an experience of a particular kind, which Hildebrand calls “cognizing” or “knowledge (Erkennen)” (see Mulligan 2014; as Hildebrand acknowledges, the very idea of Erkennen as an experience of its own kind stems from Reinach; see again his 1989a). Albeit related to Kenntnisnahmen, cognizing does not belong to that psychological category. Why not? Hildebrand provides four interrelated answers to that question. First, whereas Kenntnisnahmen can be directed at both, objects and states of affairs, cognizing is always and only directed at states of affairs. Second, cognizing is an act, whereas cognitions – even those directed at states of affairs – are states, which can endure through time. For example, you can observe the spatial relation between two cars for a longer period of time, but the act of securing knowledge (or of coming to know), insofar as it is an act, is episodic: it makes no sense to ask how long your coming to know lasted. (This characteristic also distinguishes knowledge in the sense of Erkennen from the state of knowledge, which Hildebrand calls “Wissen,” cf. Hildebrand 1969: 21ff.) Third, if you come to know a given state of affairs, that state of affairs is given to you as evident or actual – to put it differently, its subsistence must be immediately manifested to you. This is Hildebrand’s example: “I can certainly ‘cognize’ the imagined visit of my friend, for instance in order to savor everything joyful about it in advance, but in this case I cannot [come to] know that my friend is visiting” (Hildebrand 1969: 19). Fourth, in order for a state of affairs to reveal itself as subsisting to a subject, it does not necessarily have to be given to that subject in an intuitive way. To put this differently, Erkennen can be founded by a cognition: if I see that two cars are beside each other, I can come to know that they are beside each other. But Erkennen does not have to be founded by cognitions: if I see smoke, I can come to know that there is fire, without seeing that there is fire. It is now time to move our attention to the notion of attitudes (Stellungnahmen). Attitudes come into three basic kinds: doxastic, affective, or conative. Volitions will be the specific topic of the next section, but the basic idea about attitudes also applies to volitions and can be stated presently: as the German term of Stellungnahme already suggests, once the subject ‘has’ the object, that is, once the subject is acquainted with an object or state of affairs (regardless of whether this acquaintance is secured by cognitions, acts of knowledge, or states of knowledge), the subject can adopt an attitude (a mental position or a stance) toward it. If we come back to our initial example, one can easily see that you are not only a passive recipient of information present in your visual field, for you can – and generally do – take a particular position vis-à-vis that information: you not only perceive that the branches are moved by the wind, you also believe that. You not only perceive a person falling on the street, you are also afraid for her. And, perhaps, the noises coming from the street motivate you to want to close

70

Dietrich von Hildebrand

the window. Believing, being afraid, and wanting are then attitudes (of obviously different kinds) that you have adopted toward those objects and states of affairs. Another way of formulating the idea that the subject adopts an attitude is by saying that the subject responds to objects and states of affairs. Hildebrand’s notion of a response is complex as it collates several different ideas together. The first one is that attitudes have specific phenomenal properties. You always respond in a certain way to states of affairs: you are afraid or happy, convinced or doubtful, motivated or disinclined (to act). And this very phenomenal character of the experience is something you are immediately aware of by simply living through the experience (that is, you don’t need to activate a numerically distinct state directed at the attitude in order for you to be aware of its phenomenal character). This is why Hildebrand claims that attitudes do not only have an intentional object (or “content (Inhalt),” see fn. 5), but also a “content (Gehalt).” By contrast, cognitions do not have a content (in the sense of Gehalt); they just ‘have’ their objects.8 The second idea captured by the notion of a “response” is that the contents (Gehalte) of attitudes (and therefore the attitudes themselves) are intentionally directed toward objects and states of affairs: “every attitude is an attitude toward [zu] an object and thus possesses an intention [Intention] towards something objectual” (1969: 13). To have an intention is to enter an ideal relation toward the object, which is linguistically captured by the peculiar prepositions that, as we saw earlier, are required when linguistically reporting (the contents of ) attitudes. One interesting consequence of this position is that it amounts to a dismissal of the idea that cognitions are intentional experiences. It goes beyond the scope of this chapter to exhaustively elucidate this point, but it may be important to highlight that Hildebrand aligns with many other phenomenologists (though, crucially, not with Husserl in his Logical Investigations) in distinguishing different kinds of intentional directedness, which do not have an overarching genus (see Salice 2012, 2015a). To put this another way, the sense in which attitudes are about objects and facts in the world is merely analogous, but ultimately different from the sense in which cognitions and cognitive experiences more in general (incl. Erkennen and Wissen) are about objects and facts. Finally, attitudes – precisely in virtue of their being responses – secure some knowledge about their objects. This claim, however, needs specification. The first specification is that attitudes stand on (or: are funded by) cognitive experiences. For an attitude to occur, there must be a cognitive experience, which is presupposed by the attitude. Any cognitive experience can here play the founding role: so, an emotive attitude can be funded, say, by feeling the value of an object, by the knowledge (qua act) that a given object has a certain value, or by the knowledge (qua state) that a given object has a certain value (1969: 25–26). For example, it is based on your feeling of the danger (the disvalue) instantiated by that person’s falling, that you respond to that disvalue by becoming anxious. But your anxiety could be founded by the act (or state) of knowledge that that person is in a dangerous situation. This first specification has two interesting consequences. The first, which can be mentioned here only in passing, is that the different way in which an attitude can be founded has moral significance: Hildebrand contends that the more intuitive the acquaintance with the value grounding the attitude is, the more profound is its moral significance (see Salice 2015b). The second is that there are two different senses in which an attitude may be assessed as fitting or not. An attitude may be unfitting because the founding cognitive experience is unfitting. To make an extreme case: if you hallucinate a barking dog and you elicit an emotion of fear, the emotion is unfitting simply because there is no barking dog. However, the emotion could be assessed as unfitting at a different level, too: suppose you do see a barking dog and you elicit an emotion of intense

71

Alessandro Salice

fear, but the dog is, say, a Chihuahua – in this case, the intensity of the emotion does not fit the actual danger displayed by the dog.9 The second specification about the relation between attitudes and knowledge concerns the following: attitudes secure knowledge about their subjects. In fact, knowing the attitudes of a subject is to know something about that very subject in a non-trivial sense. The very person and her nature is first and foremost revealed in her attitudes: we call a person sensitive, brave, wise, unsecure, malign, smart, thoughtful, emotive, precisely on the basis of the affective, volitive, or doxastic attitudes of this person.

Willing and intentional action We have seen that attitudes can also be of a conative nature, but conation is a variegated category, which is populated by a large family of experiences. In particular, Hildebrand distinguishes between different concepts of volitions or willing experiences (Wollen), although only one of these concepts points to a psychological kind that is core to intentional agency. The first concept of will refers to an experience that is entirely inert and ineffective. Typical cases of this volitive experience are directed at other experiences: your neighbor did you some wrong and you want to forgive him, but you can’t. Or: you want to overcome your feeling of envy toward your neighbor, but again you can’t. Two facts make clear that this concept of the will is unique and has to be distinguished from other concepts. The first is that this experience can be directed at the two other forms of will – without being able to elicit them. The second, on which I will elaborate more below, is that this experience does not entail a disposition to form a decision or intention (Vorsatz). But first, what are these other two forms of will that can be aimed at by inert will? The second form of the will is exclusively directed at what Hildebrand calls “activities (Tätigkeiten),” example of which are: to eat, to walk, to hit, to fight . . . Activities, Hildebrand claims, consist in a “unity of movements and behavioral patterns (Verhaltungsweisen) of the body” (1969: 27) and do not constitute intentional actions (Handlungen). Although Hildebrand’s idea may be perhaps conducive to the German term of Handlung, it admittedly is at odds with the colloquial use of the English term “action,” according to which eating and walking do certainly qualify as actions. Linguistic considerations apart, what matters is the way in which Hildebrand characterizes the volitive experience as related to the activities it steers. In fact, the main reason Hildebrand adduces to support his claim is that the will, in this sense, is already and entirely satisfied when the activity has taken place. To put this differently, activities are not performed to produce states of affairs other than themselves: they are “basic actions” in the sense that, once the activity is realized, the conative experience is fulfilled (see also Chisholm 1964: 616; Danto 1965). Now, this is precisely what characterizes the third and last concept of the will: when you want something in this sense, you want to bring about a state of affairs as something “external” from you and from your consciousness, as it were. For instance, at this moment you may just want to (simply) eat, but you also may want your paper to be finally submitted. Your will, in this case, won’t be satisfied by a mere bodily activity (as in the case of wanting to eat) – that state of affairs in the world aimed at by the will must subsist in order for the experience to be satisfied. One way to distinguish these two concepts at the linguistic level is by arguing that the willing directed at activities requires infinitive sentences that necessarily include an agentive predicate: “I want to φ” (or “my will to φ”), where “φ” is such an agentive predicate. By contrast, the third concept of the will is captured by using the more

72

Dietrich von Hildebrand

convoluted expression “I want to bring about p,” (or “my will to bring about p”) where “p” points to the state of affairs the subject aims at realizing. Despite the important differences between these two concepts, it is crucial to highlight a feature they have in common: these two conative experiences entail the disposition to lead to a decision. If the decision is made, it initiates the realization (Realisation) of the activity or of the state of affairs: “[. . .] it belongs to the sense of this will, that it can inaugurate the realization of the state of affairs” (1969: 32).10 It is precisely this disposition that distinguishes these two forms of will from the inert one. Note that the possibility of the will to conduce to a decision requires the awareness that the state of affairs (or activity) can be brought about (or performed) by the subject. This lends further evidence to the thesis that the first concept of the will ought to be isolated from the other two: it is not possible for the subject to (willingly or voluntarily) elicit an attitude. And this is something that makes inert will akin to the concept of desire or wish (1969: 35 fn. 1).11 Let me take stock. We have seen that when it comes to, especially, the third notion of the will, Hildebrand establishes an important link between the volitive attitude, the decision, and the realization of the state of affairs. It is now time to zoom in on these relations as they open up the possibility to define the concept of intentional agency. Let us return to the example: your window is open and it is loud outside. You cannot focus on your paper, which is due the end of the month. You are aware that a closed window would help you focus. The closed window (that state of affairs) gains importance to you. You want to close the window. You decide to close the window. You act on that decision to the effect that the window is closed. Deliberative processes such as the one just described obviously consist of different experiences, which are merged together and often go unnoticed by their subjects. But it is precisely (one of the) task(s) of phenomenology to disentangle experiential knots, and to distinguish and describe its main constituents. In fact, based on Hildebrand’s theory of intentionality, one is now in a position to appreciate the several elements at the basis of deliberation. Start with cognitions: at some point you come to know that a state of affairs – like the window’s being closed – has importance to you.12 Once that state of affairs is “had” by you in the sense specified earlier, you can take a position toward it: you now want that to be the case. Note that you do not have to adopt an attitude toward that state of affairs. It may well be that, although you know of the importance that the closed window may have for your work, you remain neutral toward it. Perhaps it is the first day of spring and you prefer to enjoy the sweetly fragrant breeze at cost of being distracted by all the noises. But if you want the window to be closed, you have adopted a conative attitude, which – as such – is “ideally” directed toward that state of affairs. This intentional direction is enabled by the willing’s content (again, in the sense of Gehalt), which the subject is immediately aware of by simply living through that attitude. But then, what is the phenomenal character of the attitude? Hildebrand claims that, phenomenologically, the attitude presents itself to the subject as a commitment: if you want the window to be closed, you should close the window (1969: 31).13 The will, to put this another way, is accompanied by a sense of commitment, which makes up its qualitative character. This clearly distinguishes a mere desire – which does not elicit any sense of commitment – from the will (on this, see also Bratman 1987; Salice 2018). Desire may wax and wane, whereas willing makes one’s conduct stable through time and psychologically costly to revise – precisely because willing generates a sense of “I should” that infuses our mental life. However, as we have seen, there are forms of will that, as such, may generate that sense of commitment, but which are completely inert. So, what is distinctive about the will directed

73

Alessandro Salice

at states of affairs? This is a peculiar modal property: to want something in this sense can lead to the intention to realize the state of affairs, where the term “intention” (Vorsatz) points to an experience of a specific kind, as we will see. Again, it could be that not all tokens of willing (of this type) lead to an intention, but at least in principle they could. An attitude that qualifies as willing (in this sense), but which does not give raise to an intention, would still have to be distinguished from an experience of inert willing, which as such can never lead to an intention. One of the reasons for this claim is that, when you want to bring about p and you thus feel that you should bring about p, you are already oriented toward the realization of p. Thanks to this orientation the subject is open to two possible options: either the subject thinks that p can be realized or she thinks that p cannot be realized. It is only in the first case that the subject is disposed to form an intention to act toward p. It is to this effect that she will indeed decide to act toward p, barring all circumstances that would prevent her from doing that. But then, what is a decision or an intention (I am using these two terms interchangeably) in contradistinction to willing? On the one hand, the decision is a mental action of the subject and thus it is not an attitude (Hildebrand 1969: 36)14 – more precisely, the decision is a “spontaneous” experience in the sense that it is up to the subject if and when to form an intention, whereas we have seen that attitudes cannot be elicited at the subject’s discretion. Interestingly, intentions – other than attitudes – do not have polarity: although decisions could be directed at negative states of affairs (e.g. “I intend p not to be brought about”), decisions do not have a negative counterpart (see also Stein 1970: 311). On the other hand, the intentional object of an intention is different from the intentional object of willing. Intentions are still characterized by the sense of commitment that accompanies willing, but they determine the future conduct of their very subject. This is because they are not (only) directed at the (wanted) states of affairs, but rather at projects that aim at the realization of those states of affairs – that is, they are directed at states of affairs with an instrumental structure. Hildebrand is brief about the notion of a project (Hildebrand 1969: 43; Reinach 1989b: 291), but one could think of projects as states of affairs that have a X-by-means-of-Y structure, where X refers to the goal and the Y-variable ranges over actions of the subject that are means for the achievement of the goal (see Salice 2015c). To put this another way, if you want to bring about X and therefore you feel you should do that, then you are disposed to form an intention, the occurrence of which commits you to a certain course of actions (Y), by means of which X is supposed to be brought about. Note that one important consequence of this idea is that you can form decisions that are not supported by willing (Hildebrand 1969: 33ff.). For instance, it can be that the specification of Y compels you to an action, which you do not want to perform. Since forming an intention is under your control, you can intend to Y even though you do not have a corresponding willing toward Y. For instance, you might want to keep the window open to enjoy the fresh air, but because it is your intention to submit the paper by the end of the week, you nevertheless decide to close the window. Here, your intention to close the window is motivated by wanting the paper to be submitted, not by wanting to close the window (which is something that, by contrast, you do not want). All this indicates that intentions, insofar as they are about a project that is partly constituted by a subject’s action, lead to the realization of the wanted state of affairs. The realization, together with its psychological pendants, constitutes what Hildebrand calls Handlung or intentional agency. The realization itself requires a further act; this is an incipit act, as it were, which sets the conduct in motion (Hildebrand’s term for this further act is Inangriffnahme or “commencement” [1969: 36] – Husserl, following William James, will prefer the Latin fiat!; see Husserl 1988: 110). The existence of this act has been contested (see Reiner 1927: 76ff.), 74

Dietrich von Hildebrand

but what is more important is that the realization it triggers is not simply a set of bodily movements (a “mere doing”), but rather an experience: every doing . . . which is lived through (erlebt) as a realization, follows an intention. Insofar as it is a conscious realization of a state of affairs, there is a relation to a state of affairs, which a mere doing as such can never deliver. (1969: 37) It follows from this that the agent possesses a peculiar awareness about her action, which she does not gain through observation, but from her inner perspective (see Anscombe 2000): a set of movements is not described by the agent as an action because this is observed from a viewpoint external of the action. Rather, the set of movements is first-personally experienced as an intentional action from within the action itself precisely because it realizes the subject’s corresponding intention. To return to our initial example: all things considered, it may be legitimate of you to think that the person did not act when she fell on the street. However, this is and remains an assumption for it is her prerogative to describe the agentive situation: she (but not you) has first-personal epistemic authority in describing what she does. Sure, she may be mistaken in the description of the state of affairs resulting from her actions as an external party may be in a better position to observe the result of her action. Suppose she wanted to make somebody believe that she fell (this being the state of affairs she wanted to bring about) – to ascertain whether this state of affairs subsists she would have to rely on observation just as anybody else (and somebody may be better equipped than her to find out whether that state of affairs does indeed subsist). But any description of that state of affairs as the result of an intentional (or non-intentional) action cannot prescind from her personal perspective. Eventually, the only reliable way you have to find out what has happened on the street is to ask what her intention was (or, at the very least, whatever way one has to find out what has happened cannot ignore her intention).

Conclusion We can now come back to Frankfurt’s idea that the notion of the will can be reduced to that of desire and, thus, does not point to a genuine psychological category. Frankfurt writes: A statement of the form “A wants to X” – taken by itself, apart from a context that serves to amplify or to specify its meaning – conveys remarkably little information. Such a statement may be consistent, for example, with each of the following statements: (a) the prospect of doing X elicits no sensation or introspectible emotional response in A; (b) A is unaware that he wants to X; (c) A believes that he does not want to X; (d) A wants to refrain from X-ing; (e) A wants to Y and believes that it is impossible for him both to Y and to X; (f ) A does not “really” want to X; (g) A would rather die than X; and so on. (Frankfurt 1971: 7) How to assess this quote in the light of Hildebrand’s theory? To begin with, it is doubtful whether Hildebrand (and early phenomenologists more in general) would concede that there is a form of conation which can be squared with (a) for granting that point comes very close to denying that conation is an experience at all. Furthermore, the notion of awareness in (b) 75

Alessandro Salice

is too underdetermined to draw any sensible conclusion (but again if (b) dismisses any form of awareness, including the very possibility to become aware of one’s willing, then that again would be equal to denying that conation is an experience). However, Frankfurt is certainly right that some of those statements are consistent with the sentence “A wants to X.” For instance, if one looks at the desire-end of the volitive spectrum, which partly encompasses Hildebrand’s first notion of the will, then “A wants to X” can be compatible with (e) (precisely because wishes can be about states of affairs that are impossible to attain – like the conjunction of Y and X). In contrast, if “A wants to X” is supposed to express a decision or an intention, then certainly the statement is consistent with (c), (d), (f ), and (g) because decisions do not have to be supported by the will. Interestingly, based on what has been said in the previous section, none of those statements is compatible with “A wants to X” if the concept of “to want” is Hildebrand’s Wille. Against this backdrop, one conclusion to draw is that those claims about compatibilities hold or do not hold depending on the concept of “to want” at stake. But what does all this show? It seems to me that this just shows that the concept of “to want” is ambiguous and thus in need of an analysis. But then, which analysis should be preferred? There are reasons to believe that an analysis able to capture the fine-grained (phenomenological) differences between our conative experiences is preferable to any analysis that waters down all those differences. In fact, it is only on the basis of a phenomenologically sound theory of desires, (different forms of ) willing, intentions that one can ascertain which of those sentences is compatible with which conative notion. Only this way can disambiguation be acquired. But if this is correct, it strongly suggests that the best way to achieve disambiguation of our psychological concepts is by means of phenomenology and, crucially, by its core methodological idea that everything is what it is (and not something else).

Related topics Chapters 2 (on Pfänder & Husserl), 4 (on Scheler), 5 (on Reinach), 7 (on Reiner), 9 (on Stein), 24 (De Monticelli), 25 (Drummond).

Notes * I have presented the material of this chapter at the conference Phenomenology of Action and Volition (Prague, May 2018), where I have received much appreciated feedback. I am also thankful to Christopher Erhard and Danny Forde, who read and commented upon a previous version of this chapter. This chapter has been drafted during a research visit at the University of Memphis in April 2018. This visit has been enabled by a Research Fellowships granted to me by Prof. Shaun Gallagher and Prof. Albert Newen, for which I am very grateful. 1 A wanton is a creature the essential characteristic of which is “that he does not care about his will. His desires move him to do certain things, without its being true of him either that he wants to be moved by those desires or that he prefers to be moved by other desires” (Frankfurt 1971: 11). 2 In this context one, however, should not leave unmentioned O’Shaughnessy’s seminal work “The Will” (2008), where the author also sharply distinguishes desiring from willing. I am thankful to Christopher Erhard for drawing my attention to this notable exception. 3 Hildebrand’s philosophy has been profoundly influenced by his fellow phenomenologists – mainly by Max Scheler and Adolf Reinach –, and it stands in a complex relation with Husserl’s phenomenology. On the relations between Reinach and Hildebrand specifically on philosophy of action and its moral dimension, see Salice (2015a). Convergences and divergences between Scheler and Hildebrand have been discussed by Crosby in his 2002 work. Schuhmann has investigated Husserl’s reception of Hildebrand’s dissertation in his 1992 work. Finally, for an overview on the phenomenology of the Munich and Göttingen circles, see Salice (2015b).

76

Dietrich von Hildebrand

77

Alessandro Salice Brentano, F. (1995) Psychology from Empirical Standpoint, trans. by A.C. Rancurello, D.B. Terrell and L.L. McAlister, London: Routledge. Chisholm, R. (1964) “The Descriptive Element in the Concept of Action,” Journal of Philosophy 61/20, 613–625. Crosby, J. F. (2002) “Dietrich von Hildebrand: Master of Phenomenological Value-Ethics,” in: J.J. Drummond and L. Embree (eds.), Phenomenological Approaches to Moral Philosophy. A Handbook, Dordrecht: Kluwer, pp. 475–496. Danto, A. (1965) “Basic Actions,” American Philosophical Quarterly 2/2, 141–148. Frankfurt, H. (1971) “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68/1, 5–20. Husserl, E. (1988) Vorlesungen über Ethik und Wertlehre 1908–1914 (=Husserliana XXVIII), ed. by U. Melle, Dordrecht: Kluwer. Müller, J. M. (2018) “Emotion as Position-Taking,” Philosophia 46/3, 525–540. Mulligan, K. (2012) Wittgenstein et la Philosophie Austro-Allemande, Paris: Vrin. Mulligan, K. (2014) “Knowledge First – A German Folly?,” in J. Dutant, D. Fassio and A. Meylan (eds.), Liber Amicorum Pascal Engel, Genève: Université de Genève, pp. 380–400. O’Shaughnessy, B. (2008) The Will. A Dual Aspect Theory, 2 vols. Cambridge: Cambridge University Press. Reinach, A. (1989a) “Zur Theorie des negativen Urteils [1911],” in K. Schuhmann and B. Smith (eds.), Sämtliche Werke. Textkritische Ausgabe, 2 vols., Munich: Philosophia, 95–140. Eng. trans. by B. Smith (1982) “On the Theory of the Negative Judgment,” in B. Smith (ed.) Parts and Moments. Studies in Logic and Formal Ontology, Munich: Philosophia, pp. 315–346, 351–377. Reinach, A. (1989b) “Die Überlegung: ihre ethische und rechtliche Bedeutung [1912/13],” in K. Schuhmann and B. Smith (eds.), Sämtliche Werke. Textkritische Ausgabe, 2 vols., Munich: Philosophia, pp. 279–311. Reiner, H. (1927) Freiheit, Wollen und Aktivität. Phänomenologische Untersuchungen in Richtung auf das Problem der Willensfreiheit, Halle: Niemeyer. Salice, A. (2012) “Phänomenologische Variationen. Intention and Fulfillment in Early Phenomenology,” in A. Salice (ed.), Intentionality. Historical and Systematic Perspectives, München: Philosophia, pp. 203–242. Salice, A. (2015a) “The Phenomenology of the Munich and Göttingen Circles,” in E. Zalta (ed.), Stanford Encyclopedia of Philosophy (Winter 2015 Edition), http://plato.stanford.edu/archives/win2015/ entries/phenomenology-mg/ Salice, A. (2015b) “Actions, Values, and States of Affairs in Hildebrand and Reinach,” Studia Phaenomenologica XV, 259–280. Salice, A. (2015c) “There Are No Primitive We-Intentions,” Review of Philosophy and Psychology 6, 695–715. Salice, A. (2018) “Practical Intentionality. From Brentano to the Phenomenology of the Munich and Göttingen Circles,” in D. Zahavi (ed.), The Oxford Handbook of the History of Phenomenology, Oxford: Oxford University Press, pp. 604–622. Scheler, M. (1973) Formalism in the Ethics and Non-Formal Ethics of Values. A New Attempt Toward the Foundation of an Ethical Personalism [1913/16], Eng. trans. by M. S. Frings and R. L. Funk, Evanston, IL: Northwestern University Press. Schuhmann, K (1992) “Husserl und Hildebrand,” Aletheia V, 6–33. Stein, E. (1970) “Beiträge zur philosophischen Begründung der Psychologie und der Geisteswissenschaften,” in E. Stein (ed.), Beiträge zur philosophischen Begründung der Psychologie und der Geisteswissenschaften – Eine Untersuchung über den Staat, Tübingen: Niemeyer, pp. 2–284. von Hildebrand, D. (1969) “Die Idee von der sittlichen Handlung [1916],” in K. Mertens (ed.), Die Idee von der sittlichen Handlung. Sittlichkeit und ethische Werterkenntnis, Darmstadt: Wissenschaftliche Buchgesellschaft, pp. 1–126. Wittgenstein, L. (1984) Bemerkungen über die Philosophie der Psychologie, 2 vols., Frankfurt am Main: Suhrkamp.

78

7 THE VARIETIES OF ACTIVITY Hans Reiner’s contribution Christopher Erhard Introduction It is often said that finite conscious animals like us are essentially active and passive beings – both agents and patients. Just think of Kant’s famous thesis in the Critique of Pure Reason according to which human cognition results from the receptivity/passivity of sensibility and the spontaneity/activity of understanding (cf. e.g. A51/B75). This Kantian entanglement thesis, as we might call it, seems easily generalizable to almost all human experiences. At this very moment, for instance, while I’m actively typing words, I am also passively affected by the smooth hardness of my notebook’s keys that affects my writing activity. Indeed, it seems hard to imagine episodes of either “pure” activity or “pure” passivity. As long as we are conscious and alive, we seem to be inextricably both. Given this peculiar entanglement of activity and passivity, what exactly does it mean to be active on the one hand, and passive on the other hand? How are we to draw the line between these two experiential modalities? What is the role of activity and passivity in the mental lives of (rational) human animals like us? How are activity and passivity related to other central ideas in philosophy of mind and action theory such as rational and intentional agency, intentionality in general, and free will? Questions like these set the agenda for a heavily neglected work written within the phenomenological tradition,1 namely Hans Reiner’s 172 pages dissertation Freedom, Willing, and Activity. Phenomenological Investigations Towards the Problem of Free Will2 (1927). Reiner (1896–1991) studied philosophy with Husserl and Heidegger at Freiburg, where he also submitted his doctoral thesis in 1926, which was supervised by Husserl (see Husserl’s (1994: 464–466) assessment). After the war, Reiner became a professor of ethics at Freiburg, focusing on moral issues and the development of a phenomenologically grounded Wertethik (Reiner 1974, 1975). While Reiner’s later work is more often referred to, his early dissertation is mostly forgotten, although it serves as a foundation for his later writings. In this chapter, my primary goal is expository in nature, thereby also pleading for reconsidering Reiner’s 1927 book, both for historical and systematic reasons. To begin with, Reiner’s book is not only the first systematic phenomenological treatment of the classical problem of free will but also extraordinarily rich in content and full of concrete analyses and distinctions, all of which gravitate around our list of opening questions. 3 79

Christopher Erhard

In addition to that, as I will argue, Reiner’s work can be read as a bridge between (early) Husserlian phenomenology and the (allegedly) more action-oriented (“enactivist”) existentialist phenomenologies developed by Heidegger, Sartre, and Merleau-Ponty. Call this the bridge thesis with respect to Reiner’s early work. On my interpretation, the bridge thesis manifests Reiner’s historical importance within the phenomenological movement with respect to human agency. This can be seen from the fact that he does not only cover self-reflective deliberated intentional actions but also more subtle and spontaneous activities that need not be guided by antecedent decisions or “prior intentions” (Searle 1980). This broad focus also makes him interesting within the context of contemporary Analytic philosophy of action. According to the standard picture within this tradition, the “problem of action” consists in finding the “mark” of the rational- cumintentional “deeds” of persons which are contrasted with things “merely happening” to them (Davidson 1971). An underlying assumption of this view is that all “events” a human animal is involved in can be exclusively and exhaustively divided into rational-cumintentional action on the one hand, and passive occurrences on the other hand.4 In contrast to such a dichotomous view, Reiner draws a more nuanced picture that allows for numerous intermediate “stages” (156) between full-fledged rationalizable action and more basic levels of activity. Thus, an important systematic upshot of his phenomenological analyses is what we might call Reiner’s variety thesis of activity, according to which human agency is not restricted to rational-intentional “deeds,” but covers various forms or ways of activity, all of which are experientially manifest to us in at least a minimal way, or, if one prefers, for which there is something it is like to do them: (Variety Thesis) There is a variety of mutually irreducible kinds of phenomenally manifest activity. Reiner’s variety thesis can be understood as a systematic complement to the historical bridge thesis. In particular, the former is to be understood in the sense that besides the (standard) cases of intentional action, there are also non-intentional activities which are nonetheless more than “mere happenings.” In contrast to the still widespread classical Davidsonian dichotomy (intentional actions vs. mere happenings) Reiner thus pleads (at least) for a trichotomy (activity: intentional or non-intentional vs. passivity). The examples Reiner offers for such nonintentional, or better pre- or sub-intentional, activities include laughter and crying, quick and habitual movements, spontaneous shifts of attention, but also breathing and yawning. By extending the notion of activity, Reiner opens up the possibility of enlarging the notion of responsibility and personhood beyond those “deeds” that result from a person’s reflectively conscious beliefs, desires, and intentions. Although Reiner’s primary goal is to offer a solution to the free will problem (cf. § 1; Reiner 1931), his book mainly deals with preparatory analyses of activity, action, and willing. In fact, only §§ 27–29 and the conclusion concentrate on human freedom. In light of this, and given the focus of this volume on the phenomenology of agency, I will also concentrate on Reiner’s analyses of activity and intentional action, only in passing commenting on the theory of freedom. In doing so, the variety thesis will serve as my reference point. Accordingly, the chapter is structured as follows: In part one, I outline elements of Reiner’s phenomenological method. In the next section, I present Reiner’s final picture of human activity in some detail by means of a four-field schema. This schema represents Reiner’s version of the variety thesis. In the next section, I introduce Reiner’s novel notion of the “inner

80

Hans Reiner

will,” and shall argue that he uses it in order to include forms of non-deliberative actions he dubs “given activity” (§ 26). Before concluding the chapter, I will shortly highlight interesting parallels between Reiner’s and Harry Frankfurt’s accounts of will and agency.

Reiner as a phenomenologist From the very beginning Reiner makes clear that he wants to approach the problem of free will and action from a phenomenological perspective (cf. §§ 1, 30; Reiner 1931). What does that mean? First, a phenomenological account should never start with premature definitions, but rather develop the relevant notions gradually by carefully describing and “evincing (aufweisen)” the concrete phenomena we always-already “designate and experience (erfahren) as free in our prescientific daily life” (1).5 Most of Reiner reflections thus consist in “analyses of the conscious experiences (Bewußtseinserlebnisse) in which freedom manifests itself ” (3). In this sense, freedom, activity, and action are conceived as phenomenally manifest dimensions that can be reflectively pointed out by means of “self-observation (Selbstbeobachtung)” (67) from the first-person point of view. Note that Reiner uses the Husserlian notion of Erlebnis in a very broad sense such that it encompasses both active and passive mental acts, processes, states, and attitudes that are, however minimally and pre-reflectively, phenomenally conscious (cf. § 10). As Reiner points out in § 21, “consciousness” can be used both in a transitive (relational) and intransitive manner, and he argues that an experience can be intransitively conscious without thereby being a transitive and “higher-order” intentional object for that subject. This entails that activity and freedom can be experienced by the agent without eo ipso being intentional objects. On such a view, even subtle and marginal activities such as absent-mindedly, automatically, and even post-hypnotically induced actions (101, fn. 2) can be characterized as “conscious.” Depending on whether Erlebnisse are “performed (vollzogen) actively or passively,” Reiner calls them “activities (Tätigkeiten)” or “experiences (Erleben),” respectively. In light of this terminology, it seems worth mentioning that Reiner takes for granted that human activity and passivity have their own peculiar phenomenal dimensions. In other words, there is what-it-is-likeness both for being a patient (as in feeling one’s arm rising in virtue of external forces) and for being a doer (as in raising one’s arm spontaneously). Activity amounts to a “specific form of consciousness (eigenartige[] Bewußtseinsform)” (18). Given that the standard cases for what-it-is-likeness or qualia in contemporary philosophy of mind are restricted to passive sensory experiences such as feeling pain or sensing redness, 6 it should be emphasized that classical Phenomenology has a much more flexible notion of consciousness including both the active and passive side of mentality. Moreover, the phenomenal dimension of activity and passivity does not consist in “raw feels,” i.e. non-intentional sensations; quite the contrary, it also includes their intentional dimensions since (most) Erlebnisse are directed toward something distinct from themselves. As such, they have meaningful intentional objects, and the job of the phenomenologist consists in elucidating the “sense” (29) which is involved in our mental episodes. As I will show, Reiner offers a very rich intentional analysis especially of volitional actions. Second, in addition to the descriptive and reflective exploration of the agentive phenomenal-cum-intentional dimension of activity, Reiner aims at discovering the essence of the relevant phenomena as they occur in human beings like us, i.e. in “embodied (leibbegabte) finite persons” (2). In contrast to an empirical investigation based on “induction,” Reiner seeks an “essential consideration (Wesensbetrachtung) of human freedom” (2).

81

Christopher Erhard

Reiner’s intentional descriptions are therefore to be understood as essentially and necessarily true “synthetic a priori” statements. However, besides applying these well-known Husserlian tools, Reiner also introduces a couple of novel methodological concepts. One important notion is the so-called subjective “human sphere of power (menschlicher Machtbereich)” (38).7 A (type of ) mental episode M is a member of this sphere if and only if it belongs to the essence of finite embodied beings like us that we can do or realize M. Moving at least some of our limbs, for example, belongs to the subjective sphere of power. Within this sphere Reiner introduces several sub-distinctions, two of which are especially important for our purposes. First, there is the distinction between the “absolute” and the “profective” or “conditional” (39) sphere of power. Episodes within the absolute sphere can be performed immediately or directly by trying to do so (e.g. raising one’s arm).8 In contrast to this, activities belonging to the “profective sphere” (from the Latin proficere = doing something for . . .) can only be performed as a result of a previous “absolute” performance, but not directly at will. This is because a profective act essentially relies on a passive occurrence which must be “given” (29) to us. As examples for a profective activity Reiner mentions finding an adequate expression for a thought, or finding the solution for a problem (27–8). These are a profective processes because we cannot find a suitable expression directly by willing to do so – finding, like discovery, cannot be “forced.” This is not within our absolute sphere of power, in contrast to simply saying something or uttering a sentence. All we can immediately do is endeavoring, striving, or making an effort as a result of which we might eventually find a solution or a suitable expression (cf. Reiner 1931: 58). If we are successful, Reiner claims, the resulting “entire act” (28) of finding the solution exhibits a “peculiar mixture of activity and passivity” (27). On the one hand, it contains a passive moment, namely an “idea (Einfall)” (28), a kind of Aha! moment. On the other hand, such an Aha! experience is not a purely passive occurrence, but rather presupposes active striving and attentional resources. Profective experiences thus illustrate a kind of process that cannot be easily captured by the standard dichotomy between intentional actions and “mere happenings” since they essentially incorporate both an active and a passive dimension. Generally speaking, for Reiner, the contrast between the active and the passive is not to be regarded as a contradictory opposition, but rather as a spectrum (cf. § 30) the poles of which can be blended into “mixtures” (29), “hybrid character[s]” (103), and various “complications” (47). Although the event of finding the solution for a puzzle is in some sense happening to me, it differs sharply from a passive sensation such as pain or hunger that can occur to me out of the blue. This is because the act of finding is essentially experienced as an accomplishment or fulfillment (Erfüllung) of the previous “absolute” activity of looking for a solution.9 Besides the distinction between the absolute and the profective sphere, Reiner introduces a distinction between the “creative” and the “deletive” sphere of power. Processes within the creative (or instead deletive) sphere are those (types of ) episodes whose initiation (or instead termination) lies within our subjective (absolute or profective) sphere of power. Perceptual experiences, e.g. belong neither to the creative nor to the deletive absolute sphere of power. In fact, they are a subset of the profective sphere. This is because I cannot bring, say, a visual perception of a chair about, or stop it from existing, just by trying to do so. This is not so much because perceiving a chair is a factive episode that requires the chair’s actual existence, but rather because perceptual experiences are au fond passive episodes whose occurrence requires that they are ultimately given to me. In contrast to perceiving a chair, visually imagining a

82

Hans Reiner

chair belongs both to the creative and the deletive absolute sphere of power since I can (in principle) initiate and terminate such an imaginative episode at will.10 For Reiner, these are not mere empirical truths, but rather essential truths about the nature of perception and imagination in relation to human powers in general. By means of these two distinctions within the overall subjective sphere of power, Reiner develops a detailed “material investigation” (38 f., 151), which is supposed to figure out which kinds of experiences are in principle within our power (cf. §§ 11–15, 18–24). In this spirit, Reiner explores how sensations, perceptions, judgments, volitions, feelings and emotions, position takings, imaginings, intentional actions, decisions, intentions, and other kinds of episodes are related to the “human sphere of power.”

Reiner’s variety thesis: a four-field schema Summarizing the main ideas of Reiner’s book is not an easy task. This is mainly because Reiner draws from various authors, offers a huge number of examples and distinctions, thereby tending to proceed rather casuistically and eclectically. Accordingly, in his assessment of Reiner’s dissertation, Husserl (1994: 464-466) remarks that the reader is likely to get confused by the uncountable distinctions that do not seem to crystallize into a final unified view. This impression is justified to a certain extent. However, in Section § 21 of the published version of his dissertation (p. 98), Reiner draws a (basically) “four-field” schema that encapsulates the core distinctions of his account. (The original schema is reprinted at the end of this chapter.) My main goal in this section is to elucidate this schema, in order to substantiate Reiner’s variety thesis of activity. The gist of Reiner’s view is constituted by two essentially different though closely related and non-exclusive ways of being active, namely bodily “motor activity (Bewegungsaktivität)” or “activity of moving oneself (Aktivität des Sich-Bewegens)” on the one hand, and rational “activity of the will (Willensaktivität)” on the other hand.11 While Willensaktivität is typically self-reflectively (or at least saliently) conscious, Bewegungsaktivität can also occur in a relative non-conscious, “sub-liminal” manner. To each of these two forms of agency corresponds a distinctive contrary phenomenon, namely passivity (“mere happenings”) in the case of motor activity, and “pre-free (vorfreie)” experiences in the case of volitional activity.12 All in all, a four-field schema emerges: Activity of the will (rationalcum-intentional action or merely “innerly willed” activity)

Inactivity of the will (pre-free experiences)

Bodily motor activity

Field I Examples: deliberately tidying up your desk, laughing, crying

Bodily passivity

Field II Examples: mentally calculating 36, refraining from eating chocolate

Field III Examples: absentmindedly wiggling your toes while reading, breathing, yawning Field IV Examples: pain, remembering X on the occasion of seeing Y (association)

Reiner’s Four-Field Schema

83

Christopher Erhard

In what follows, I will explain each field of this schema. I shall begin by elucidating the notion of willed activity (Fields I and II), which is the most elaborated one in Reiner’s book. Field I: Imagine that you are (deliberately) cleaning up your chaotic desk. This is an example of what Reiner calls an “action (Handlung)” (§§ 15, 18–19). According to Reiner, this is also a paradigmatic instance of a “full activity (volle Aktivität)” which is based on “selfdetermination of one’s own future life (Selbstbestimmung des eigenen kommenden Lebens).” This idea of self-determination serves as Reiner’s paradigm for all volitional activity (Fields I and II). It is an essentially forward-looking activity in which the agent tries to realize a certain goal or purpose which is consciously represented and whose “realization” she positively evaluates. In doing so, the agent “controls (lenken)” (17) her life which she experiences as being within her own power, and which she actively brings about: she is a “self-actor (Selbsttäter),” (17) and not a passive “spectator” of her experiential life. Elaborating on these initial ideas, Reiner draws from Husserl’s Third Logical Investigation, and argues that intentional actions are certain mereologically structured units that consist of distinguishable “partial acts (Teilakte)” (90) or “moments” each of which plays a well-defined role within the action as a whole. In addition to that, Reiner defends a volitionalist view, according to which the nature of intentional action is spelled out in terms of the multifaceted phenomenon of the human will. In sum, Reiner distinguishes four main aspects of intentional actions:13 (I) The irreducibly modal “awareness of one’s own power (Bewusstsein des eigenen Könnens)” – in brief: the “I can (Ich kann).” A twofold manifestation of the will in terms of a (II) “volitional stance (Willensstellungnahme)” on the one hand, and (III) an “intent(ion),” “resolve,” or “decision” (Vorsatz, Entschluss, Entscheidung) on the other hand, and finally, (IV) the “performance” or “execution (Ausführungstätigkeit).” Let me comment on these in turn. Ad (I). Reiner claims that intentionally performing (or trying to perform) an action A requires an occurrent awareness of one’s own power to perform A: I cannot intentionally try to A unless I take myself to be able to A. This “taking” must not be understood as an explicit judgment or thought, but rather as a non-inferential “immediate awareness (Innesein)” (27, fn. 1) of being up to the task, so to speak. Thus, it is more like a savoir-faire (basically knowing how to move or halt my mind or body), rather than an explicit savoir-être (knowing that I can do certain things). The awareness of power thus captures the phenomenal moment of up-to-me-ness that is contained in voluntary action. Reiner mainly uses perceptual vocabulary to describe the I can: it is a sort of “sight (Sicht)” (22f., 160f.) of my practical possibilities in a certain situation; but it could also be described as a feeling.14 Importantly, this feeling of power (“I can do A.”) can occur apart from the corresponding cognitive and self-conscious belief or judgment (“I belief that I can do A.”), and also vice versa: It is possible to believe that one is able to A without actually feeling capable. The I can thus exhibits a certain cognitive impenetrability – a phenomenon known from various emotive experiences. Furthermore, Reiner argues that although the I can is necessary for the motivational structure of intentional action, it is not in itself an action or activity, but rather a receptive awareness that must be “given” to the subject. To be more precise, the I can results from so-called “motivated passivity” (§ 3) or association. In our example of cleaning up the desk, my action is motivated by the feeling of being able to clean it up, which, in turn, might be non-inferentially motivated or “afforded” by the untidy look of the desk. The I can also plays a crucial role for Reiner’s theory of freedom since it opens up our sphere of power in the first place. Without this experience, our freedom would be phenomenologically mute, so to speak. Thus, the awareness of power can be characterized as a necessary (though insufficient) condition of free will.15 In this sense, freedom requires the “sight of one’s own ability (Sicht eigenen Könnens)” (160).

84

Hans Reiner

Ad (II). In addition to the awareness of power, “volitional stances (Willensstellungnahmen)” constitute a further indispensable part of intentional action. Using Brentanian vocabulary, Reiner describes these stances as emotion-like experiences (Gemütsbewegungen) (53). Like standard cases of emotions such as joy or grief volitional stances are characterized by a positively or negatively attuned intentional orientation in virtue of which they are inherently “interested” in their object (cf. 110ff., 132) thereby caring about their goals. A volitional stance can also be a mere “wish” (72) without agential commitment, or a “(dis)approval” of some already obtaining state of affairs (cf. Reiner 1974: § 19). Besides, these stances can be either patent experiences or latent “attitudes (Einstellungen)” (cf. §§ 28–29) that need not be occurrently phenomenally conscious all the time. Noematically speaking, stances are essentially directed toward “transcendent” (69) state of affairs.16 In this respect, they resemble propositionally structured “pro-attitudes” (see Davidson 1971). They want the world to be in a certain way. In our example, I want that my desk is tidied up again. Reiner further claims that volitional stances have a normative dimension and are essentially responsive experiences. This is because they are founded on evaluative feelings or, to use Husserl’s phrase, Wertnehmungen that immediately grasp axiological features of the intentional correlate. Volitional stances thus “answer” to an apprehended value through a kind of felt (not necessarily moral) “oughtness (Sein-Sollen)” (74, 106) that is dependent upon my personal interests: that the desk is tidied up again appears to me a valuable state of the world that is demanded by its actual untidiness. Thus, my volitional stance is a responsive attitude toward this felt demand, and makes the corresponding state of affairs appear as something that ought to be the case. Besides their constitutive role for intentional agency, volitional stances also make up the core of Reiner’s theory of freedom and personhood. Although they are not essentially active episodes, but typically passively motivated and unreflectively adopted stances that belong to the profective sphere of power, Reiner eventually argues that we can become reflectively aware of them, which, in turn, opens up the possibility of critically evaluating and thereby endorsing or rejecting them (cf. §§ 28–29). Thus, there is a sense in which we are free with regard to our own will (cf. § 27).17 A person, for Reiner, is essentially constituted by the whole (hierarchical) system of latent and patent volitional stances that guide and inform her actions and evaluative attitudes (cf. 153). Ad (III). According to Reiner, states of affairs do not exhaust the things I want to be realized when I act intentionally since I also want to be the one who realizes them. In other words, in intentionally acting, I not only want the world to be in a certain way, I also want myself to act in a certain manner. This points to a fundamental ambiguity of the German term Wille/Wollen, which Reiner refers to in terms of the “double intentionality of the will” (69). In particular, Reiner claims that there a two essentially different kinds of volitional experiences involved in intentional actions. These two kinds of willings have different noetic features and noematic correlates, to use Husserl’s terms. Besides the volitional stance, there is the “act of resolve (Entschluß),” “intent (Vorsatz),” or “forming of an intent (Vorsatzfassen),” all of which are ultimately rooted in an act of decision. But what does it mean, phenomenologically, to take a decision? In contrast to volitional stances, decisions are primarily directed toward the agent’s “executional performance (Ausführungstätigkeit),” not toward the transcendent states of affairs she thereby wants to realize in the world (cf. 71, 90).18 In this sense decisions contain an implicit self-reference. I must be aware of the activity I decide upon as being within my sphere of power: deciding to clean up my desk requires taking myself to be able to do this. My will thus has, to use Searle’s terminology, a causally self-reflexive “world-to-mind direction of fit” (Searle 1980). Moreover, decisions, again in contrast to

85

Christopher Erhard

stances, are intrinsically active phenomena that must be performed by the subject in order to exist (cf. § 25). Reiner also claims that decisions are not emotion-like experiences like volitional stances, but more akin to spontaneous “speech acts” such as assertions or promises (cf. 75). As such, decisions can be taken in a rather “cool” intellectual manner, which entails that the value of the performance decided upon can be purely instrumental. In the case of the chaotic desk, it is both possible to love or hate to do the tidying up; I can also be emotionally indifferent or relatively neutral, and just do the job because it serves as means to an end I appreciate. Note also that Reiner carefully distinguishes between decisions and choices (cf. §§ 8–9), where only the latter necessarily involve an awareness of at least two practical possibilities. By contrast, deciding can be a “simple activity” (22). Despite these essential differences, Reiner holds that decisions and volitional stances are intimately related. To be more precise, decisions are one-sidedly founded on volitional stances: one cannot take a decision without relying on a volitional stance. On the other hand, it is possible to adopt a volitional stance without eo ipso taking a decision. I might, for instance, strongly want that my desk looks orderly again without doing anything about it – because of akrasia, laziness, procrastination, or any other inner or outer obstacles.19 With regard to Reiner’s initial characterization of intentional action in terms of “self-determination of one’s own future life,” decisions play a crucial role. This is because decisions are essentially future-directed acts that rely on a reflexive self-awareness, which, in turn, depends on the I can. In deciding to A, I “posit” the performance of A as to be “put into play by me (von mir ins Spiel zu setzen)” (71).20 Ad (IV). Finally, the decision brings about the performance or execution, the active “core part” (79) of action (cf. 57–62, 80ff.). In virtue of the performance, the will (volitional stance + decision) becomes a causally efficacious and perceivable part of reality – at least in the case of bodily constituted actions. But performances are not movements of our bodies and/or minds simpliciter, but rather founded processes having several bodily and mental “strata (Schichten)” (81). A (normal) bodily performance, for instance, is essentially founded on a bodily movement which is both kinesthetically experienced “from within” and perceived “from without” (cf. 58). This does not mean that the performance is an overt bodily movement caused from within and monitored by means of perception, because the relation between willing and performance is not only a causal but also a motivational one. The performance is experienced as fulfilment and “realization” of the will. It is (experienced as) the coming into being of the content of the will, and it brings – if all goes well – the willed state of affairs about. This motivational causation is not mechanical or ballistic in nature since neither the volitional stance nor the decision vanish as soon as the performance sets in; both remain “actual” during the performance into which they are, as it were, absorbed (79ff., 143). Thus, for Reiner, the structure of action is not only phenomenally, intentionally, mereologically, and motivationally complex but also cumulative in nature: all mental progenitors of the performance are still “alive (lebendig)” (80) when the realization (= whole action) takes place. Let me sum up Reiner’s extremely detailed phenomenological description of voluntary (bodily) action (=Field I). A rational voluntary action A is a unified act or experience consisting of four not necessarily successive partial acts fused together into one highly structured unitary whole: an occurrent feeling of power (“I can do A”) + a volitional stance directed toward a valuable state of affairs p + a decision to A in order to realize p + an act of execution. Field II: Up to now, I have described Reiner’s account of volitional bodily action in some detail. He also touches upon “purely ‘mental’ action (rein ‘geistige’ Handlung)” (81), and claims that intentional bodily and mental actions are structurally “analogous” (81). This entails that the fourfold mereological structure from above (I can + volitional stance + decision + execution) can also be found in mental actions. Take the example of mentally calculating 36 (=729). 86

Hans Reiner

According to Reiner, this action is made up by a volitional stance directed toward the state of affairs of knowing the result of 36 together with the intention to perform the required calculatory activity. Besides, I must also take myself to be able (or at least to be able to try) to calculate the result. Field II also covers “hybrid” (156) forms of willed activity, namely “volitional acts of forbearance (willentliche Unterlassung)” (98, 156). Take the action of consciously not eating a delicious piece of chocolate. For Reiner, this is an instance of genuine mental activity that is accompanied by bodily passivity. In the remaining part of this section, I shall comment on the so-called “pre-free sphere” (98) which comprises both non-volitional bodily movements (Field III) and utterly passive experiences (Field IV). This sphere is called “pre-free” since the notion of freedom cannot be meaningfully applied to it (cf. § 31). Freedom essentially requires an awareness of power which, in turn, relates to our volitional capacities. According to Reiner, this awareness of will-power is missing in Fields III and IV. Field III: The “motor activities” pertaining to Field III emerge relatively late in Reiner’s book (cf. §§ 21, 25–26, 30), and are only discussed briefly. Reiner’s examples include breathing, blinking, yawning, and stretching your limbs absent-mindedly while lying in bed. Given that most of these motions seem to have a rather “sub-personal” or reflex-like nature, one might be surprised that Reiner includes them among his phenomenological descriptions. According to Reiner, indeed, these movements are phenomenologically characterized by a felt active (bodily) movement, which he describes as an “instinctive (triebmäßig, triebhaft)” (58, 102) motion of “going out of oneself (aus sich Herausgehen)” (99). In contrast to bodily movements guided by the will, these “vital movements” (Reiner 1931: 57) do not exhibit intentionality in the object-directed sense of Brentano-Husserl.21 Nevertheless, they are not supposed to be mere mechanical, reflex-like, and passive happenings initiated from without. In contrast to utterly unconscious bodily processes such as (undisturbed) digestion or blood flow, these vital activities can be directly influenced and modulated by the will, although they are not in themselves willed movements (cf. 97). Just think of the various breathing techniques we can use to influence our respiratory circles. Unfortunately, Reiner does not engage in a more elaborate analysis of these primordial activities. Field IV: The last field includes passive experiences that are neither accompanied by active bodily movements nor are initiated, guided, or sustained by the will. Positively speaking, this “sphere of passivity” is characterized by a certain absorbedness of the ego. The subject is “captured (Ergriffenheit)” (16) by the presence of some “content (Inhalt)” (16), which can be immediately present, but also be remembered or expected. In this sense, being phenomenologically passive involves adhering to the present contents of the mind, the dimension of the future and the possible remaining temporarily out of sight. Passivity thus involves a sense of “givenness.” Moreover, a passive experience centripetally affects the subject, whereas in being active we experience ourselves as centrifugally moving toward the world. As an example, think of a salient bodily pain that imposes on you and absorbs all of your active resources. Starting with these initial characterizations, Reiner distinguishes two forms of experiential passivity, namely purely affective and motivated passivity (cf. § 3). When we are passively affected by a “stimulus (Reiz)” (10) such as a sudden noise or pain, our consciousness is, so to say, frozen, and completely occupied by these contents. Motivated passivity, by contrast, involves an “internal meaningful connection from one act to another” (9). It occurs in all cases of association, in which one object reminds us of another. But this movement from one act to another happens independently from and sometimes also against the subject’s self-conscious will. In being passively motivated, Reiner writes, consciousness “jumps by itself (springt von selbst)” (6) from one object to another. 87

Christopher Erhard

With regard to the general relation between activity and passivity, Reiner holds that activity must always be motivated by passivity (24–26, 62–64, 89, 114, 166): “there is no activity without a . . . concretely enabling passivity” (157). This is supposed to be a non-empirical, a priori “eidetic law (Wesensgesetz)” (62, 114). To mention just one example, Reiner argues that in order to act intentionally the agent must become aware of being able to perform the relevant action. But this awareness or feeling of power, as already mentioned, must ultimately be “given” to the agent: which practical possibilities are open to me must dawn on me, so to speak. Intentional actions are thus essentially founded on a passively motivated component. This can be seen as a phenomenological confirmation of the Kantian entanglement thesis mentioned at the beginning of this chapter, at least in one direction: all three forms of human agency (Fields I–III) essentially comprise passive or receptive components.

The extended (inner) will: why volitional agency includes more than rational-cum-intentional action Up to now, Reiner’s description of volitional activity (Fields I-II) looks rather traditional. Ignoring the details for a second, he follows a venerable tradition dating back to thinkers such as Aristotle, Aquinas, and Kant who take the will to be a rational, autonomous, and causally efficacious power of human beings that can be exercised with regard to cognitively presented goals. In this section, however, I will try to show that Reiner introduces a further notion of willing that does not fit smoothly into this rationalistic tradition. To be more precise, Reiner argues that besides volitional stances and decisions/intentions, there is a third mani festation of the human will, which he calls “inner will” (inneres Wollen). With respect to the four-field schema, the inner will is supposed to extend the field of volitional activities beyond the sphere of rational and decision-based actions in the narrow sense. In other words, there are will-based activities that do not involve planning, (forming of an) intention, “prior intention” or even “intention in action” (Searle 1980), or taking a decision. Paradigmatic cases of such activities are unreflective emotional reactions like anger, smiling, frowning, laughter, and crying, but also spontaneous perceptual or cognitive shifts of attention. Although these experiences need not be grounded on a previous or simultaneous decision, they still involve the will because they are to some extent “innerly willed,” i.e. endorsed, admitted, or sustained by the agent. These activities also show that the standard Davidsonian dichotomy between rational-cum-intentional actions and “mere happenings” is not fine-grained enough. But what is this prima facie mysteriously sounding “inner will” supposed to be? Let’s take the case of spontaneous laughter as an example. Although laughter is typically not initiated by a decision or intention to laugh (apart from, let’s say, actors pretending to laugh on stage), it does not seem to be something that “merely happens” to us – like a pain or itch. Unless laughter turns into hysterical laughter or is triggered physiologically (through tickle or otherwise), it seems to be a personal process that we are living-through. (The way someone laughs or smiles can tell us a lot about the kind of person s/he is.) Furthermore, laughter usually does not occur out of the blue, but rather emerges as a specific response to a certain situation we are faced with. It is difficult to specify the precise patterns which make us laugh, but often a certain disharmony or incongruence is involved to which we spontaneously respond, although we are fully aware that our laughter won’t effectively change the situation. To put it into Sartrean terms, one might characterize laughter as a “magical,” i.e. non-instrumental,

88

Hans Reiner

“transformation of the world” (Sartre 1939). According to Reiner, in normal cases of laughter, we “devote (Hingabe)” (132) ourselves to the situation “without any inhibition (Hemmung) and reflection” (132). It is this sense of “devotion,” endorsement, or identification that is captured by the notion of the inner will. The laughter may involuntarily begin, but for it to unfold, we must to some extent actively give in, let it happen, or go along with it. So, what Hanna and Maiese have recently written about anger can equally be said about laughter: Again, consider normal anger. . . . it seems that we can always choose either to continue to work ourselves up into an absolute fury (say, by trying to shout even louder) or to begin to calm ourselves down (say, by trying to lower one’s voice). Thus it seems that, even if it begins involuntarily, the process of normal anger can always eventually be voluntarily escalated or suppressed to some extent. To be sure, for people with pathological anger management problems, it is a very different story. But that is abnormal anger, and not what we are talking about. (Hanna and Maiese 2009: 243–244) On my interpretation of Reiner, the role of the inner will is precisely to organize or modulate such spontaneous ways of behavior either by letting them happen (or even escalating) or by resisting or suppressing them. Importantly, this phenomenon of “letting it happen” or “resisting” does not require an extra additional higher-order reflective act. It can rather happen “seamlessly (bruchlos)” and “without further ado (ohne weiteres)” (77, 130, 134). Thus, the inner will is an inbuilt aspect of the very activity it accompanies and modulates. In this context, Reiner argues that his notion of inner will is supposed to be a volitional variety of Brentano’s “inner awareness (inneres Bewußtsein)” (§ 26). As is well-known, Brentano holds that all conscious mental phenomena involve an inbuilt reflexivity or self-awareness. In conjunction with Brentano’s view that consciousness is also characterized by “intentional inexistence” (Brentano 1874/1995: 68), it follows that every act has two objects, namely a “primary” and a “secondary” one. 22 While an “outer” object is intended primarily, the act serves as its own secondary object. On such a one-level account neither “reflective grasping (reflexives Erfassen)” nor “noticing (Bemerken)” (129) is necessary for an act to be phenomenally manifest. Now, just like Brentano’s inner consciousness, Reiner’s inner will is not a second- order volitional stance, but rather a pre-reflective and non-attentive “consent (Zustimmung)” (128) to one’s own experiences on the basis of which the agent “devotes” (124, 134) herself to the primary object. In contrast to decisions or “prior intentions,” the inner will essentially unfolds synchronously with the willed activity, accompanying and molding it from beginning to end. Therefore, the inner will can be understood as an internal structuring and re-organization of the agent’s mental life. In this way, Reiner claims, the inner will accounts for the difference between an experience that just occurs to or happens in us, and an active episode we go along with or try to resist. In virtue of the inner will, the passively occurring urge to laugh spontaneously and “without further ado” grows into the activity of laughter; the subject “simply follows a tendency (einfaches dem Zuge Folgen)” (134). This minimal form of volitional activity in terms of the inner will is also called “given activity” (§ 26) since it requires neither priming or premeditating acts nor even an additional effort or intervention through the agent. A further dimension of the inner will consists in its “ego-centrality (Ich-Zentralität)” – a term Reiner borrows from Pfänder (1911). The core idea is that activities shaped by the inner

89

Christopher Erhard

will stand in a special relation to the active subject. Willed actions are essentially “performed by me” (von mir vollzogen); they do simply not occur to or in me – like “eccentrial” or “peripheral” experiences (125). This “by-me-ness,” as we might call it, contributes to the fact that the agent identifies herself only with her willed activities; she owns her voluntary doings, as a result of which she can also be made morally responsible for them.23 (Interestingly enough, we often blame people for laughing or the way they laugh. Some people, for example, laugh in a vulgar manner, and we criticize them for doing so.) On Reiner’s view, inner will and centrality are co-extensional notions such that a mental episode is innerly willed if and only if it is also performed centrally: To begin with, in the case of volitional activity (Willensaktivität) the act of resolve, which essentially constitutes the character of activity, is always necessarily performed by the ego-center. And the same goes for the execution of the action, which must also emerge from the central ego in order to be and remain a genuinely volitional execution and not degenerate into a mere “mechanical” and pre-free doing. Thus, centrality is an essential feature of activity in the sense of voluntariness (Willentlichkeit). But this also holds conversely, since voluntariness is an essential constituent of centrality! That is, no act performed by the ego-center could be an act of pure intuition (reine Anschauung); and there is no volitional stance which exhausts itself in being directed towards the intentional object. On the contrary, every central act of the ego, even if it is not brought about by another act of the will, incorporates a certain inner consent (innere Zustimmung) of the will to the act itself! (128) Since all will-based activity – in contrast to mere motor activity – is a form of selfdetermination, the inner will, which is at the root of all higher-forms of willed activity (volitional stances, decisions/intentions, executions), also amounts to a form of minimal self-determination. This is because the inner will is reflexively built into an ongoing and present process of activity.

Reiner and Frankfurt Before closing this chapter, a brief comparison between Reiner’s rather novel notion of inner will-cum-centrality and the more familiar account of volitions, decisions, identification, internality, and “wholeheartedness” developed by Harry Frankfurt (1998) might be helpful. To begin with, Frankfurt often describes situations in which the desires of an agent stand in conflict with one another. In order to move on, a decision is needed. According to Frankfurt (1998c: 170), 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.” This can be understood in terms of Reiner’s claim that the inner will involves centrality, which, in turn, is constitutive for the things that “really” matter to the person. Reiner even holds that the essence of the agent qua (moral) person consists in her hierarchically structured volitional life. For the “essence of the person,” Reiner writes, “action and doing are . . . ultimately irrelevant; above all, her positional stances constitute her essence” (153). A further similarity concerns the status of the “passions.” Frankfurt holds that “there is a useful distinction to be made, however awkward its expression, between passions with respect to which we are active and those with respect to

90

Hans Reiner

which we are passive” (1998b: 60). He spells this difference out in terms of internal (=active) and external (=passive) emotions, which, in turn, depend on the way the respective passions are embedded within the overall will-structure of the person. In a similar vein, Reiner refers to the inner will in order to account for the difference between emotions merely occurring “in us” and emotions essentially performed “by us.” (cf. §§ 26–27). Reiner’s notions of inner will and centrality thus play a similar role like Frankfurt’s idea of internality. In doing so, both authors accept the idea that “passions” are not unequivocally passive occurrences – a view that also goes against the standard Davidsonian dichotomy between things we intentionally do and things “merely happening to us.” A difference between both accounts is that Frankfurt frequently appeals to higher-order desires and decisions in order to account for the internal/external distinction, whereas Reiner’s inner will is basically a first-order phenomenon that need not but can be fostered by higher-order attitudes or decisions. 24 The parallels between Reiner and Frankfurt even run deeper since in his theory of freedom Reiner heavily relies on the idea that personal freedom consists in the particular way in which we actively relate to our volitional stances. That is, genuine freedom of the will (in contrast to mere freedom of action) concerns our “ability (Können)” to actively endorse or reject lower-level volitional stances by means of higher-order stances, thereby molding the structure of our volitional lives (cf. §§ 6, 27–29; Reiner 1931: 60–66). For Reiner, thus, freedom of the will is the (non-uniquely determined) higher-order capacity to actively identify with some of our lower-level volitional stances. However, on a lower level, freedom is already at play when we, by virtue of the inner will, pre-reflectively endorse and go along with certain tendencies (cf. 134). So, for Reiner, spontaneous laughter and smiling are manifestations of human freedom. Frankfurt defends a similar hierarchical picture of freedom according to which it “is these acts of ordering and of rejection – integration and separation – that create a self out of the raw materials of inner life” (1998c: 170). Decisions thus create an “orderly arrangement” (1998c: 173) of my “inner life” by transforming it into an “integrated whole” (1998c: 174). As is well-known, this active and ultimately harmonious hierarchical organization of one’s “inner life” is central for Frankfurt’s notion of personal freedom (Frankfurt 1998a). The parallels between Reiner and Frankfurt could be spelled out in more detail, but this would go beyond the purpose of this chapter. I shall therefore close with some final remarks.

Closing remarks Reiner’s early work is a good and rich illustration of what the Munich phenomenologist Moritz Geiger’s calls the “passion of phenomenology,” namely “seeing the differences” (Geiger 1933: 4). Thus, my primary goal of this chapter was to summarize the essential “differences” Reiner claims to have discovered within the realms of human passivity and activity. Apart from the detailed analyses especially of rational-cum-intentional actions, I think we can draw three important lessons from Reiner’s phenomenology of agency: (1) there is a sense of willing, namely the “inner will,” which is operative even when explicit decisions or self-reflective intentions are not involved. In terms of this inner will, emotional reactions and “passions” like anger or laughter, but also spontaneous and non-deliberative movements of the body or the mind can be genuinely active. The inner will can thus be understood as a minimal form of activity in terms of which the agent structures or organizes her mental

91

Christopher Erhard

life, even if she does not deliberatively take decisions or plans anything. In this way, a great variety of human agency comes to the fore (see Reiner’s variety thesis). (2) The standard picture in Analytic philosophy of action suggests that human agency can be neatly separated into fully “rationalizable” acts we intentionally and deliberately do on the one hand, and things “merely happening” to us on the other hand. From a Reinerian perspective, this view should be replaced by a triadic picture. On such a view, innerly willed but not full-blown intentional doings lie in between the two poles of passivity and full-fledged rational-cumintentional action. In this regard, Reiner’s work can also be understood as a bridge between a more “intellectualistic” phenomenological tradition (represented, arguably, by early Husserl) and later phenomenologically oriented philosophers (such as e.g. Buytendijk and Plessner 1925, Heidegger, Sartre, and Merleau-Ponty), the latter highlighting the importance of unreflective spontaneous actions. (3) On Reiner’s nuanced view of human experience, pure activity or spontaneity is a myth. All activity, be it grounded in the “inner will” or self-reflectively planned, is essentially re-active in nature and entangled with passivity from the very start since it responds to something “given” that matters to us, thereby calling on us to take a stance.25

Related topics Chapters 2 (on Pfänder & Husserl), 3 (on Pfänder), 6 (on von Hildebrand), 21 (Hanna), 24 (De Monticelli), 25 (Figal).

Appendix: Reiner’s schema of human activity (quoted from p. 98 of his 1927 book)

Figure 7.1 Reiner’s schema of human activity

92

Hans Reiner

Notes

93

Christopher Erhard

19 20

21

22 23

24 25

states of affairs.” The desires or “pro-attitudes” in contemporary action theory roughly correspond to Reiner’s notion of Willensstellungnahme. Reiner describes various “modalities” of volitional stances (§ 19). I can, for instance, “decisively (entschieden)” or hesitatingly want that p. These modifications resemble Husserl’s doxic modalities in Ideas I. As already mentioned, Reiner also acknowledges cases of “pre-free action (vorfreie Handlung)” (101; see § 21); i.e. intentional actions directed towards the realization of a state of affairs that are not guided by an actual act of resolve, examples of which include habitually and absent-mindedly performed actions. However, Reiner claims that such actions must have been performed previously on the basis of an explicit decision. This is why he also calls them “post-free (nachfrei) actions” (102). Or is there a sense in which breathing, for instance, has an “intentional object”? Probably not, if “object” is taken in a narrow sense. More broadly conceived, however, it seems plausible to think of breathing as a future-directed striving that inherently though blindly aims at something. At what? Maybe just at relief from a felt bodily tension or pressure. Conceived as a form of striving that can be fulfilled, breathing thus has a “correlate” though not an “intentional object.” Cf. Brentano (1874/1995: 119): “Every mental act is conscious; it includes within it a consciousness of itself. Therefore, every mental act, no matter how simple, has a double object, a primary and a secondary object.” Reiner’s notion of centrality resembles the contemporary notion of agent causation or “self as source.” On such a view, not only events or states of affairs but also substances or subjects as such can be genuine causes (see e.g. Horgan, Tienson, and Graham 2003, and the chapter written by Horgan and Nida-Rümelin in this volume). Several passages in Reiner’s book point in this direction (cf. 140, 152). He claims, for instance, that central acts are “caused from the I itself (aus dem Ich heraus Gewirkte[s])” (165). This is of course possible and often the case. The more an experience is backed up harmoniously by higher-order stances, the higher is the resulting degree of freedom or, as Reiner puts it, the degree of “wakefulness (Wachheit)” of the person (cf. §§ 28–29). Research on this chapter was supported by a scholarship granted by the German Research Council (ER 819/2-1 and ER 819/2-2). I am grateful to Dan Zahavi and Robert Hanna for helpful comments on previous versions.

References Brentano, F. (1874/1995) Psychology from an Empirical Standpoint. Eng. edition by L. McAlister. London, New York: Routledge. Buytendijk, F. J. J. and Plessner, H. (1925) “Die Deutung des mimischen Ausdrucks. Ein Beitrag zur Lehre vom anderen Ich,” in H. Plessner (ed.) (2003), Gesammelte Schriften in zehn Bänden. Band VII: Ausdruck und menschliche Natur, Frankfurt a. M.: Suhrkamp, 67–130. Danto, A. (1965) “Basic Actions,” American Philosophical Quarterly 2, 141–148. Davidson, D. (1971) “Agency,” in R. Binkley, R. Bronaugh and A. Marras (eds.), Agent, Action, and Reason, Toronto: University of Toronto Press, 3–25. Erhard, C. (2019) “Unifying Agency. Reconsidering Hans Reiner’s Phenomenology of Activity,” Husserl Studies 35, 1–25. Frankfurt, H. (1971/1998a) “Freedom of the Will and the Concept of a Person,” in H. Frankfurt (1998), The Importance of What We Care About, Cambridge: Cambridge University Press, 11–25. Frankfurt, H. (1977/1998b) “Identification and Externality,” in H. Frankfurt (1998), The Importance of What We Care About, Cambridge: Cambridge University Press, pp. 58–68. Frankfurt, H. (1987/1998c) “Identification and Wholeheartedness,” in H. Frankfurt (1998), The Importance of What We Care About, Cambridge: Cambridge University Press, 159–176. Geiger, M. (1933) “Alexander Pfänders methodische Stellung,” in E. Heller and F. Löw (eds.), Neue Münchener philosophische Abhandlungen, Leipzig: J. A. Barth, 1–16. Hanna, R. and Maiese, M. (2009) Embodied Minds in Action, Oxford: Oxford University Press. Hildebrand, D. von (1916) “Die Idee der sittlichen Handlung,” Jahrbuch für Phänomenologie und phänomenologische Forschung III, 126–252.

94

Hans Reiner Horgan, T., J. Tienson and G. Graham (2003) “The Phenomenology of First-Person Agency,” in S. Walter and H.-D. Heckmann, Physicalism and Mental Causation: The Metaphysics of Mind and Action, Exeter: Imprint Academic, 323–340. Husserl, E. (1994) Briefwechsel, ed. by Karl Schuhmann, Vol. III/4: Die Freiburger Schüler, The Hague: Kluwer Academic Publishers. Kriegel, U. (2015) The Varieties of Consciousness, Oxford: Oxford University Press. Pfänder, A. (1911) “Motive und Motivation,” in A. Pfänder (ed.), Münchener Philosophische Abhandlungen, Leipzig: Barth, 163–195. Reiner, H. (1927) Freiheit, Wollen und Aktivität. Phänomenologische Untersuchungen in Richtung auf das Problem der Willensfreiheit, Halle: Niemeyer. Reiner, H. (1931) “Die Freiheit des menschlichen Wollens,” in C. J. Bock (ed.) (1969), Die Rolle der Werte im Leben. Köln: Wienand (First published 1931 in: Philosophie und Leben 7), 53–68. Reiner, H. (1974) Die Grundlagen der Sittlichkeit, Meisenheim am Glan: Anton Hain. Reiner, H. (1975) “Die Ausbildung und Fortbildung der phänomenologischen Methode als Methode der Ethik,” Zeitschrift für philosophische Forschung 29, 108–117. Ricœur, P. (2007) Freedom and Nature. The Voluntary and the Involuntary, trans. from the French and with an introduction by E. Kohák. Foreword to the new edition by D. Ihde, Evanston, IL: Northwestern University Press (French edition 1950: Le Voluntaire et l’Involontaire. Paris: Aubier). Sartre, J.-P. (1939) Esquisse d’une théorie de la emotion, Paris: Hermann. Scheler, M. (1921) Der Formalismus in der Ethik und die materiale Wertethik. Neuer Versuch der Grundlegung eines ethischen Personalismus (Sonderdruck aus Jahrbuch für Philosophie und phänomenologische Forschung, Bd. I (1913) u. II (1916)), Halle: Niemeyer. Schmid, H.B. (2011) “Feeling Up to It – The Sense of Ability in the Phenomenology of Action,” in A. Konzelmann Ziv, K. Lehrer and H.B. Schmid (eds.), Self-evaluation. Affective and Social Grounds of Intentionality, Berlin, New York: Springer, 215–236. Searle, J. (1980) “The Intentionality of Intention and Action,” Cognitive Science 4, 47–70. Spiegelberg, H. (1982) The Phenomenological Movement. A Historical Introduction. 3rd edition, Dordrecht, Boston, MA, London: Kluwer. Stein, E. (1922) “Psychische Kausalität,” Jahrbuch für Philosophie und phänomenologische Forschung V, 1–116. Strawson, G. (2003) “Mental Ballistics, or the Involuntariness of Spontaneity,” Proceedings of the Aristotelian Society (New Series) 103, 227–256.

95

8 MARTIN HEIDEGGER From fluid action to Gelassenheit Sacha Golob

Introduction Heidegger’s views on agency are central to his philosophy: in both his earlier and later work they form part of an intricate web of notions including responsibility, normativity and activity. At the same time, however, Heidegger’s position can be elusive. This is partly due to deep methodological and conceptual differences which make it hard to locate him in relation to the standard analytic or Kantian debates. It is also because his views are subject to a series of complex shifts – for example, during the early 1930s and then again in the aftermath of the war. There is no scholarly consensus on the exact nature of these shifts or on the degree of continuity or change that they imply: as a result, it is impossible to adequately address Heidegger’s views on “agency” or indeed any other topic in a single article without radically restricting the chronological range of the discussion. For these reasons, I am going to focus primarily on the period around Being and Time (Sein und Zeit, published 1927): Heidegger’s most influential text; it is also the best point of contact with contemporary philosophy of action or with other phenomenologists. I close by sketching more briefly how these themes develop within Heidegger’s later work. Throughout I try to avoid an excessive reliance on Heidegger’s distinctive terminology: my aim in that sense is to provide an analysis of his views, rather than a recitation of them. By “agency” I mean the capacity for action, where acts are differentiated from mere movements. Suppose you shove Paul from behind and he falls onto Sarah, hurting her. Intuitively, this is not an exercise of Paul’s agency: it was something that was done to him; you could have equally pushed an inanimate object in just the same way. One could make the same point by saying that it was not something that Paul chose to do. Agency is thus closely connected to notions like activity, freedom and blame: Paul is not blamed for hurting Sarah because he did not do so freely or willingly. This basic picture rapidly becomes more complex, however: suppose Paul shoves Sarah whilst in the grip of addiction, or because his early upbringing has made him careless and insensitive. Or consider a case where the relevant “movements” are not physical but mental. Paul is responsible, both epistemically and perhaps morally, for judgments that he makes. But is he responsible for forgetting something – a friend’s birthday perhaps? Or for thoughts that occur to him unbidden? Neither is a matter of action in any simple sense. Is he responsible for the sexual images that enter his mind or 96

Martin Heidegger

only for decision to act on them? Cases like these are designed to pressure the links between responsibility, agency and choice. One option is to try to reinforce such links: one might argue that Paul is at fault for forgetting a birthday just if he chose not to do what was necessary to remember it, such as making notes or other reminders for himself, or because his forgetfulness testifies to the problematic lack of importance which he judges the celebration to have.1 All these issues will be in background of any adequate discussion of agency, but the place to begin is with the nature of action itself.

Heidegger and phenomenological approaches to action Early Heidegger develops a highly original and systematic picture of human existence, his so-called “existential analytic of Dasein” (SZ: 13). At its core is a hierarchical structure in which certain forms of experience or understanding have explanatory priority over others. For example, he identifies “cognition” [Erkennen], in something like Kant’s sense of the term, as a “founded” or derivative mode of experience – one made possible by “being-inthe-world” (SZ: 61–62). Heidegger’s full picture is immensely complex, and resists easy summary. This is partly because it is unclear how “being-in-the-world” relates to familiar categories such as perception or judgment: from a Heideggerian point of view, this unclarity is a necessary consequence of the move beyond such frameworks and the pseudo-problems, such as skepticism, attendant on them (SZ: 56–57, 228–229). It is also because his story interweaves descriptive and normative elements: whilst the first half of SZ focuses on life in its “average everydayness” (SZ: 16), the second half addresses both the temporal structures which supposedly underlie that life and “authenticity”, a privileged form of existence that escapes the banality and “idle talk” supposedly pervasive in social contexts (SZ: §§35 and 53). Even to sketch Heidegger’s overarching framework would take a chapter in itself, and so I propose to concentrate here on what has undoubtedly been the most influential aspect of his early philosophy of action. The best way to approach that is via one of the most famous examples in the history of philosophy: Heidegger’s hammer. In developing an analysis of “average everydayness”, Heidegger introduces a characteristically artisanal scenario: a craftworker, completely absorbed in their task, silently discards a hammer that is too light and takes a more appropriate one (SZ: 69, 157). The case foregrounds several points: the paradigm case of action for Heidegger occurs fluidly and unreflectively within a meaningful context, tacitly structured by my goals and by the equipment around me (SZ: 70, 129). At its core is what he labels the “for-the-sake-of-which”: the agent’s self-understanding, where that is cashed in terms of the norms to which the agent is committed, the skills she exercises in pursuing them and the shifting configuration of the world which is thus manifest to her (SZ: 87). To take a simple example, Joan understands herself as a professor insofar as she encounters and evaluates the world in terms of that identity. Here is Heidegger’s own description of the attendant phenomenology: Coming into the lecture-room, I see the lectern . . . What do “I” see? Brown surfaces, at right angles to one another? No, I see something else. Is it a largish box with another smaller one set on top of it? Not at all. I see the lectern at which I am to speak. You see the lectern from which you are to be addressed and from which I have previously spoken to you . . . I see the lectern in one fell swoop, so to speak, and not in isolation, but as adjusted a bit too high for me. I see—and immediately so—a book lying upon it as annoying to me (a book, not a collection of layered pages with black marks strewn across them). (GA 56–57: 71–72) 97

Sacha Golob

To switch to Heidegger’s own terminology: “the referential totality of significance (which as such is constitutive for worldhood) is ‘tied up’ with a ‘for-the-sake-of-which’” (SZ: 192). The result is a picture of the world in which self and world go hand in hand: objects are encountered as means to specific ends, tasks appear as desirable or obligatory, all ordered, or in Heideggerian terms “disclosed”, on the basis of the agent’s self-understanding (SZ: 192). Many contemporary Heideggerians, drawing on Gibson, express this by talk of “affordances and solicitations” (Gibson 1979): for Joan, the pile of scripts teetering on the desk “affords” reading, whilst for Harry, the room cleaner, it is manifest simply as an obstacle to be moved aside. For the skilled craftsperson a certain weight of hammer is manifest immediately as right without, Heidegger cautions, her needing to predicate any particular mass of it (SZ: 154). This sketch introduces some central Heideggerian themes and concepts, and it raises a number of exegetical questions – for example, why exactly is the craftsperson’s experience different from one of judging that the hammer is the right weight? But my present concern is primarily philosophical rather than textual, and so I want to focus on two of the key ways in which this early Heideggerian picture has been developed. Heidegger’s own hammer example, however, hides a complex twist, and so at least initially I will use a simpler case. Suppose Tom is working out how to complete a difficult project: he sits down, weighs the options, thinks through the various pros and cons, and then acts, taking a series of steps to realize his plan. This is a prime case of “reflective” or “deliberative” action, i.e. action chosen by the agent after some extended process of rational thought. But such cases are relatively rare: the very same day, Tom performed countless other acts, few of which required any kind of reflection. For example, he drove to the shops, fluidly tracking a whole range of social and legal norms covering where, when and how he can turn or park. SZ’s focus on “average everydayness”, of the silent setting-aside of a tool within fluid and absorbed action, exemplifies this. Let’s call such acts “non-reflective”. We can now introduce the two key Heideggerian claims. First, Heideggerians argue that philosophy has over-focused on the reflective or deliberative case, and generalized from it. Wrathall, for example, accuses analytic philosophy of precisely this error: [A]nalytic agency theory takes deliberative action as its paradigm – that is, action in which the agent aims at an end or goal that he or she envisions, and pursues that end in a rational way. (Wrathall 2014: 208) Crowell and Okrent raise a similar objection to Kantian theories such as Korsgaard’s. As they see it, Korsgaard’s mistake was to analyze agency in terms of one specific mental state, self-conscious reflection (Okrent 1999: 70; Crowell 2007: 321). Korsgaard writes: [O]ur capacity to turn our attention onto our own mental activities is also a capacity to distance ourselves from them and to call them into question . . . I desire and I find myself with a powerful impulse to act. But I back up and bring that impulse into view and then I have a certain distance. Now the impulse doesn’t dominate me and now I have a problem. Shall I act? Is this desire really a reason to act?... And this sets us a problem no other animal has. It is the problem of the normative. (Korsgaard 1996: 79) Here it is reflection which separates human behavior from that of animals and so which introduces notions like responsibility and normativity. On this Kantian picture, I am not 98

Martin Heidegger

directly responsible for my impulses: they are understood as natural events resulting largely from factors beyond my control such as genetics and upbringing. But I am responsible for whether I reflectively endorse them.2 The boundary between mere “happenings”, including those that occur within my body such as “impulses”, and agency is thus aligned with reflection. But what now about Tom’s driving – his fluid switching down of gears? It obviously doesn’t seem like a series of purely animalistic impulses but nor is it a matter of a reflective decision. Korsgaard’s response is to claim that such behavior is not preceded by a decision but nevertheless still “embodies” one (Korsgaard 2009: 127). The Heideggerian suspicion is that such talk “remains a metaphor” (Crowell 2007: 328): in the crudest terms, how effective can a model based on self-conscious reflection be once reflection is removed from the equation?3 Second, Heideggerian accounts argue that it is in fact absorbed or engaged or fluid action which has explanatory priority, and that an account of agency should begin there. In Dreyfus’s influential terminology, we should start with “coping” (Dreyfus 1991: 67–75): with Tom’s driving, not his planning the project. As Crowell puts it: What is needed is an approach to the categorial structure of action that arises from a description of non-deliberated action, and this is just what Heidegger provides. (Crowell 2013:267) This move brings with it some immediate challenges. On the one hand, Heideggerians must explain why such action differs from mere instinct or from the kind of responsive behavior of which non-human animals are capable. Unless they can do this, they risk losing the links between agency and responsibility. If Tom drives into someone, he’ll be called to answer for it; this is hard to grasp unless his behavior can be separated from the kind of responsive training capacities found in dogs. Wrathall again provides a neat formulation of the issue: The problem for the paradigm of fluid action, however, is preserving the idea that bodily movements are “full-blooded actions,” as opposed to a merely instinctive, animalistic response to stimulations. (Wrathall 2014: 197) On the other hand, however, some story is still needed about the relationship to deliberative thought: it is one thing to highlight a mode of action missed by the tradition and another to claim that it has explanatory priority. To put it another way, why isn’t Heidegger simply inverting the biases of the canon, switching from an obsession with deliberative agency to an obsession with non-deliberative agency? To this end, Dreyfus developed an influential “two floors” model of agency on which deliberative action emerges only when fluid action hits some form of obstacle – thus explaining its secondary status (Dreyfus 2005 provides a particularly detailed presentation of this view). Deliberative agency would be, like Kantian cognition, a “founded” or secondary mode of “being-in-the-world” (SZ: 59). The two main Heideggerian claims I have identified are in a very literal sense phenomenologically driven: they are motivated by first-person descriptions of action which seem to undermine the deliberative or reflective model. For example, here is Wrathall introducing the key idea of solicitation, a notion that I will discuss in more detail below: Consider Superbowl MVP Phil Simms’ recollection of what it was like to play American football at a professional level: “You take the ball. You get it. And man, you react and you throw it. And you go, ‘well I don’t know why I did that, but I did it, and let’s just 99

Sacha Golob

move on.’ . . . You know, my mind couldn’t focus on anything too much. It really, it just reacts. It’s amazing. You react to: well, I saw a helmet move. And you whoa! And you think, ‘gosh, why did I do that?’ Then you see the film, and everybody parted . . . and you go, ‘ohhh, that’s why I did it.’” As this description suggests, highly skilled, fluid actions are experienced, not as the deliberative outcome of my aims and desires and beliefs, but as being drawn out of me directly and spontaneously by the particular features of the situation, without the mediation of occurrent mental or psychological states or acts. Of course, the actions that are solicited or drawn out of the agent depend on his or her current way of being involved in the situation. (Wrathall 2014: 195) Simms’ remark serves as a kind of phenomenological test case for Wrathall: whatever model of agency we come up with must do it justice. Dreyfus, in critiquing both Searle and Husserl, similarly appeals to Larry Bird’s memories of the court: Heidegger would like basketball player Larry Bird’s description of the experience of the complex purposive act of passing the ball in the midst of a game: “[A lot of the] things I do on the court are just reactions to situations ... A lot of times, I’ve passed the basketball and not realized I’ve passed it until a moment or so later.” (Dreyfus 1993: 28) One can see the similarity to Heidegger’s original scenario: just as the craftsperson reacts fluidly to the obstacles the material poses, Bird navigates the shifting challenges posed by his opponents almost as if on auto-pilot. As with Wrathall, the tactic is to use phenomenology to motivate an alternative to reflective or deliberative theories of agency. Cases like Bird’s thus form the basis of Dreyfus’s account of “skilled coping”: essentially an analysis of absorbed or non-deliberative action that emphasizes the skilled management of multiple variables – from the position of onrushing opponents to the precise angle of the bounce. As Dreyfus puts it: According to Heidegger . . . skillfull coping does not require a mental representation of its goal at all. It can be purposive without the agent entertaining a purpose. (Dreyfus 1993: 28) Again, the intended contrast is with a deliberative model on which the agent sets out some particular goal before taking steps to realize it. We can now see, incidentally, why Heidegger’s own hammer example is not the ideal introduction, insofar as it can call to mind both the fluid “coping” he wants to highlight and a prior “planning” stage in which the craftsperson deliberates over materials, goals, resources – a planning stage that fits much more naturally with a Kantian or analytic approach (Heidegger is aware of the tension and occasionally warns against the artisanal model on this basis – see, for example, GA 24: 150–151).

Phenomenology versus logical analysis? The Heideggerian points just discussed are phenomenological in a straightforward sense: they are motivated by the distinctive first-person experience of absorbed action. But how helpful is such a phenomenological approach when thinking about agency? Perhaps the best way to approach this is via the extended debates between Dreyfus and John Searle.

100

Martin Heidegger

For Searle, phenomenology is at best a starting point: as he puts it, “phenomenological approach is inadequate to solve the problems that bother me . . .logical analysis has to go far beyond phenomenology” (Searle 1999: 277). Indeed, for Searle, phenomenological theories of agency are positively dangerous because they foster what he labels the “phenomenological illusion”. We do not, when playing tennis have a conscious experience of having propositional representations of conditions of satisfaction. . . . and of course, there is no propositional content running consciously through your head. But all the same, the entire logical apparatus of intentionality applies. If you describe the phenomenology and stop there, you miss the underlying logical structures. (Searle 2004: 327) Searle’s point is this. Obviously, much action is absorbed or fluid rather than deliberative. But that is only the first step: we need then to undertake logical analysis to identify the conditions of satisfaction for such acts – conditions which hold despite the evident fact that the agent has no phenomenological awareness of them. For example, my doing x requires not just that x occurs but that it occurs as a result of my intention to do x: a case where I intend to raise my arm and it happens to go up because you shock it electrically is not a case where I have managed to act (Searle 1983: 85). In short for Searle, phenomenology barely scratches the surface: indeed, by the end of the series of exchanges he talks of the “bankruptcy” of phenomenology (Searle 1999: 277). Searle’s frustration is evident throughout his dialogue with Dreyfus, and it speaks to deep terminological differences between the phenomenological and analytic approaches to agency. For example, Dreyfus reads talk of “representations” as phenomenological and so meets it with examples such as the Bird case: he takes Bird to show that “skillfull coping does not require a mental representation” because Bird describes an experience completely absorbed in the game (Dreyfus 1993: 28). But for Searle, as for almost all analytic philosophy of mind, to say that a given act is “representational” is simply to say that it has: [T]ruth conditions, obedience conditions, conditions of improvement, and conditions of satisfaction generally. As I remarked earlier, representation for me is a logical, not a phenomenological notion. The expression “mental representation” does not imply “sentences or pictures in the head.” (Searle 2004: 327) Bird’s act of shooting is representational in this sense because it can succeed or fail: the challenge as Searle sees it is to work out what, beyond the ball going through the hop, needs to be included in those conditions – how exactly, for example, do we rule out the electric shock case just mentioned? The lesson here is not confined to the analytic case. The attacks on Korsgaard discussed earlier have validity precisely because reflection is at least partly a phenomenologically defined state: it is very hard to see in what sense Bird or Simms is reflecting. But most commentators understand Kantian talk of transcendental apperception or the “I think” in a logical fashion: as they see it, the point is not, of course, that we go around deliberating or talking to ourselves, but that our acts have an inferential and syntactic structure which depends on the faculty of understanding and which animals thus lack.4

101

Sacha Golob

The points so far might be simply summarized like this: the fact that reflection or deliberation are unusual states is entirely compatible with the view that action possesses a representational or propositional structure, and that this is well illuminated by considering those more explicit instances. More broadly, pushing the Dreyfusian approach to its extremes will create internal problems. For example, motivated by cases such as Bird’s, Dreyfus occasionally denies that absorbed action contains any awareness of the self: there is “awareness but no self-awareness” (Dreyfus 1991: 67). Exegetically, it is difficult to square this with Heidegger’s own emphasis on the selbstweltlich or “self-worldly” nature of experience: the “for-the-sakes-of which” that structure the manifold (GA 61: 95). Furthermore, as McKinney notes, it makes it hard to understand the connection to deliberative thought when the latter does occur. In Dreyfus’ story, deliberation emerges when absorbed action hits some kind of obstacle; yet to see “a difficulty as a summons to deliberation demands that I be able to regard the difficulty as a difficulty in relation to my purpose” (McKinney 2017: 78). As we will now see, removing self-awareness from the picture also obscures a key part of Heidegger’s positive story.

Normativity and Heideggerian approaches to agency The results of the previous section should make us wary of straightforward appeal to phenomenology when analyzing agency. In a sense this is unsurprising: it is hard to motivate the claim that certain facts are both phenomenologically evident and yet systematically denied by one’s opponents. In this section, I want to examine some of the ways in which the most sophisticated Heideggerian theories move far beyond such straightforward appeals, interweaving phenomenological analysis with detailed work on normativity: after all, no true phenomenologist, particularly not one with Heidegger’s hermeneutic background, would automatically take a first impression as giving the truth of the matter. To do this we need to flesh out the Heideggerian framework: there are complex textual issues in play here, and what follows is only a snapshot, but it will be enough for our purposes.5 As discussed, for Heidegger agents primarily encounter objects as instrumentally, phenomenologically and normatively structured by our “for-the-sake-of-which” or selfunderstanding: when Joan enters the classroom, she experiences it not simply as a space of a given size and magnitude, but as ordered by that identity: the pile of scripts on the far table is highly salient; it is “nearer” in that sense than the floorboards beneath her feet (SZ: 107).6 Furthermore, the pile does not just blankly loom large: it is experienced with a particular teleological and motivational valence, it is seen as “requiring marking”, a perception that may grow gradually oppressive as its size mounts. The desk at the front, meanwhile, is seen as an appropriate place for doing that marking. In contrast, for the contractor come to price new flooring, it is the floorboards that are the salient feature and the pile of scripts and the table are merely obstacles that hinder measuring up. This gives some grip on the role of normativity in the Heideggerian picture: our experience is saturated with awareness of what is appropriate or required. But what is the connection between normativity and action? Crowell, the most influential exponent of this line of thought, provides a succinct summary. He starts by glossing actions, as opposed to mere movements, in terms of intentions and intentions in terms of norms: An action, in this sense, is an attempt to realize an intention, and for this reason what belongs to the action must not only conform to the norm but must also be done in light

102

Martin Heidegger

of that norm. To be an act of revising my paper, it is not enough that my bodily movements realize that goal (as though I were a machine or a zombie). I must also be making such movements in order to revise my paper. (Crowell 2013: 264) Actions here are defined in normative terms: they are attempts to do something that can succeed or fail. Furthermore, those success conditions must be “immanent” to the act: my car exploding is evidently a failure in some sense but not an action: If there is no such standard . . . for determining what it is supposed to be, then action cannot be distinguished from mere movement. And the standard must be immanent to the act, since if it is completely external to the experience, then it is irrelevant for an account of practical intentionality. (Crowell 2013: 264) This approach has notable connections to Searle’s own: Searle analyzes action in terms of conditions of satisfaction which are precisely norms in Crowell’s sense, standards that an act may meet or fail to (Searle 1983: 90). The problem, as Crowell sees it, is how to specify these conditions or norms when one moves beyond cases of deliberation: there I decide to do x and my subsequent movements either do or don’t meet that goal, making talk of success or failure easy. But what about the act “that does not follow upon a process of deliberation? Whence does it derive its normative orientation?” (Crowell 2013: 264). It is here that Crowell appeals to Heidegger’s “for-the-sake-of-which”: it is this identity which determines the conditions by which my actions must be judged as success or failures. Were I not committed to being a writer, what I am doing could not be determined. Am I revising my paper or creating an artwork? . . . [W]hat I am trying to do depends on what I am trying to be. (Crowell 2013: 273) We have here an account which cashes out action in terms of “trying to do” and “trying to do” in terms of “trying to be”, that is, in terms of my ongoing attempt to realize some forthe-sake-of which. This meshes well with the characteristically absorbed phenomenology of everyday life: John might see himself as a tough guy and try to live up to the characteristic norms of that identity without having engaged in any reflection about it. When deliberation does occur, it will be against a backdrop of such fluid, normatively structured acts. Before assessing this account, I want to introduce a closely related position, that defended by Wrathall. As mentioned in Simms’ example, Wrathall understands fluid action in terms of solicitations: we can think of this as a complex pattern of saliencies and attractions ordered around my self-understanding, as in Heidegger’s lecture hall case. It is because Simms understands himself as a quarterback that particular movements are salient and invite, without his even thinking about it, a quick release, whilst other movements fade into the background, below the level of awareness. Actions are then distinguished from mere happenings in terms of the pattern of such saliencies: What distinguishes a doing from a happening, on the model of fluid action, is that the solicited action is the result of the individual that I am – that it would not have been

103

Sacha Golob

solicited were I not myself, and I would not be myself were I not individuated from other possible actors. (Wrathall 2014: 212) As with Crowell, actions are responses to the normatively structured situation generated by my self-understanding: the solicited pass manifests itself simply as “what ought to be done”. I now want to highlight several points regarding such accounts. First, we need to hear more about the relative roles of “trying”, “solicitation” and “skills” in these stories. Suppose that the very moment the world solicits Simms to throw an out route, I trigger an electronic charge to his hand, propelling the ball loose. Simms’ “throwing” here is not an action and Wrathall can plausibly concur on the grounds that the act is not “the result of the individual that [Simms is]” (emphasis added) – note though that it meets his other conditions, “it would not have been solicited were [Simms] not [himself ]”. Now suppose Joan is incredibly nervous because she cares so much about grading, and when trying to take a sip of water, her hand spasms, spilling the whole glass. Here the movement is the direct result of Joan’s solicitations. The individuating role of the “for-the-sake-of-which” highlighted by Crowell also still applies: the event is individuated by what Joan is trying to be, it is a ruining of papers, not a political protest. Finally, for both Crowell and Wrathall the explanatory grounds for an action must lie at least partly in the agent’s exercise of skills (Crowell 2013: 267; Wrathall 2014: 196): again, it seems like this condition is met (Joan was exercising the skill of handling the glass). Do we have an action? If so, why is there no action in Simms’ case simply because the source of the jolt lay within Joan’s body? If not, when would we have an action – what if Joan spills only a little, something a clumsy person might habitually do? Second, Heideggerian approaches raise interesting questions around the links between responsibility and action. Take the example of a friend’s forgotten birthday – clearly this was not done in order to do anything, but it makes sense to apologize and to feel guilt for. Wrathall’s model could allow us to accommodate such cases by recruiting them as acts: my forgetfulness flows directly from the way the world solicits me – a solicitation that manifests a lack of concern for someone to whom I should be more attentive.7 Alternately, one could simply severe the link between responsibility and activity: exegetically, this might be developed via Heidegger’s account of guilt as an unavoidable and ontological condition (SZ: 269). Third, the “nervous spill” case highlights an interesting methodological point: Heideggerian accounts of agency will typically be suspicious of the causal chain counter-examples that dominate the analytic literature. Consider an influential case, adapted from Chisholm. George intends to kill his uncle. Thinking about the prospect whilst driving, he becomes so excited he loses control, fatally striking a pedestrian who happens to be his uncle (Chisholm 1966: 37). It seems true that George didn’t intentionally kill his uncle, a point hard to capture on the Heideggerian accounts for the same reason as with our nervous grader: the crash was the result of an exercise of skill in accordance with the distinctive patterns of solicitations that individuate George. The interesting issue here is not whether Heideggerians might finesse their position in some way so as to avoid such cases. It is more the divergent attitude to such elaborate thought experiments: in analytic debates on agency these are regarded as central means for isolating necessary and sufficient conditions, whilst the natural Heideggerian reaction is to treat them as derivative on an account which must first be established via more everyday examples. Fourth, what about the possibility of actions unconnected to my self-understanding, to “what I am trying to be” in Crowell’s terms or to the “individual I am” in Wrathall’s? Presumably the response would be that either some such connection exists – perhaps a prima

104

Martin Heidegger

facie random act is motivated by a self-conception as unscripted and spontaneous – or that the behavior is indeed not an action but something akin to Kant’s “mere happening” where my body is. But clearly more details are needed on the precise form of connection required. Fifth, one can see how theories of agency will closely link to theories of freedom. For example, from a Kantian perspective, the natural question for Wrathall’s account is whether I have control of what solicits me, and if not whether I am able to stand back and distance myself from such draws and pulls – this emphasis on distancing, seen above in Korsgaard’s theory, is designed to guarantee my freedom to do otherwise when the pattern of solicitations remains fixed. Heidegger’s attitude toward freedom is extremely complex; what I want to flag here is simply that what is a core constraint for Kantians, preserving my ability to always do otherwise, is not in play in most Heideggerian accounts.

Authenticity and Das Man So far, I have focused on the case of individual action. I want to conclude the discussion of Heidegger’s early work by looking briefly at the social context of agency. From the perspective of Being and Time, there are two key questions. First, to what degree does the social context make individual agency possible? Many commentators, particularly those influenced by Dreyfus, take Heidegger to argue that a public or shared understanding of the world is a transcendental condition on individual agency. It is this “everyday understanding of the world as publicly available” that “allows us to identify ourselves as individuals in the same world” (Carman 1994: 218). One way to understand this is in terms of a content externalism closely linked to the for-the-sake-of-which: insofar as understanding myself as x involves the exercise of certain skills and the use of certain equipment within a public context, the available options are not simply up to me. It is impossible to now perform many of the actions that defined the samurai way of life: I could physically make the movements, but in the absence of the right social context this would only be going through the motions.8 Jonathan Lear gave a famous example of this when he argued that, after a certain point in the changes in Crow society, nothing could count as planting a coup stick anymore (Lear 2006: 32). Second, to what degree does the social context limit or restrict individual agency? The key here is Heidegger’s account of das Man: I follow Dreyfus in translating this as “the one” (Dreyfus 1991: 141–145). Drawing on familiar Enlightenment themes, Heidegger presents this as a form of conformism, motivated in part by a refusal to question pre-established social norms and conventions. By doing what “one does” agents dodge responsibility for their behavior, in favor of a “tranquilizing” conformity (SZ: 127, 178). In its most extreme form, this evasion of responsibility is so severe that Heidegger talks as if it is no longer an exercise of my agency: as he puts it, “the one” is the “realest subject of everyday life. . . the self of everyday Dasein is a one’s self ” (SZ: 129). In contrast, authentic Dasein who has freed itself from domination by “the one” lets its “ownmost self ” take action (SZ: 342). The complication is whether one should read such remarks as denying that the inauthentic have agency, or whether they express the view that such agency has been radically limited, perhaps as when in the grip of an addiction. The latter alternative seems much more plausible: after all, Heidegger seems to hold such individuals partly responsible insofar as they preferred tranquilizing certainty over radical questioning (GA 20: 384–386). A full answer, however, would have to wait on detailed discussion not just of authenticity, but also of the

105

Sacha Golob

relationship between das Man, the everyday and a third notion that Heidegger dubs “undifferentiatedness” and which seems designed to provide a neutral point between authenticity and inauthenticity (SZ: 331).9

Later Heidegger – Gelassenheit, Activity and Passivity I now want to turn to Heidegger’s later thought. As indicated, the question of how many “phases” there are to Heidegger’s development and when exactly any changes occur is controversial, and what follows is largely neutral on the details. My aim is to introduce one fundamental strand of that later thought, a strand which revolves around a challenge to agency. To start we need to distinguish two questions. One, what defines agency – what distinguishes an action from a mere happening? This has been our earlier focus. The other is whether there are privileged or problematic forms of agency? I touched on this when discussing authenticity, and in Heidegger’s later work it comes center stage. Very roughly, his view is that a range of what traditional philosophy would identify as both epistemic and social problems are ultimately rooted in the dominance of a certain model of agency. To combat this, he attempts to valorize an alternative, one which downplays activity in favor of a meditative receptiveness. The place to begin is with the problematic form of agency. During the 1930–1940s, Heidegger develops a reading of the history of philosophy as the triumph of the willing subject. Within this, both individual agents and being in general are conceived in terms of the will: In the subjectivity of the subject, will comes to appearance as the essence of subjectivity. Modern metaphysics, as the metaphysics of subjectivity, thinks the being of beings in the sense of will. (GA 5: 243) The problem as Heidegger presents it is an appropriative and domineering vision of agency: as he puts it in the Nietzsche lectures, “To will . . . is to will-to-be-master [Herrsein-wollen]” (NII:  265). Perhaps the clearest example comes in his account of modern technology as a state in which all entities, including ultimately ourselves, are understood as possible energy for exploitation within a calculative framework – “modern metaphysics” is thus able to slide seamlessly from a Nietzschean glorification of the will, a glorification of such exploitation, to a conception of individuals merely as “human resources” (GA 7: 18). Heidegger’s solution is to emphasize an alternate way of thinking about agency, one which brings “the measure of meditative thinking decisively into play against mere calculative thinking” (Gel: 21). The shift is encapsulated by the term Gelassenheit: this marks an attitude of contemplative attentiveness to entities and to being. It is not a purely passive comportment, nor is it simply a “denial of willing” (GA 77: 77). Instead, it is best glossed as a receptivity in which we wait attentively, without letting our expectations determine and limit that which we are open to (Gel: 42/68). These moves have been widely criticized, particularly by authors who read early Heidegger as a decisionist whose account of authenticity would valorize radical choice irrespective of its content. Thus Habermas: The language of Being and Time had suggested the decisionism of empty resoluteness; the later philosophy suggests the submissiveness of an equally empty readiness for subjugation. (Habermas 1987:141) 106

Martin Heidegger

This worry will be particularly acute if one reads Heidegger’s later talk of being [Sein] or beyng [Seyn] as ascribing it some form of agency, perhaps of a supernatural type. Wolin, for example, accuses Heidegger of “a secularized replay of medieval ontology”, in which our agency must be sacrificed in the face of Being-God’s (Wolin 1990: 152). Some of these objections are uncharitable: Heidegger is insistent that Gelassenheit remains a form of action and absolutely not a “letting self-will go in favor of divine will” ­­(Gel: 33–34). Indeed, this is one of several points on which he explicitly separates himself from theologians such as Eckhart (for helpful discussion, see Davis 2007: 127–140). But how are we to assess the story even rightly construed? Clearly, much hangs on Heidegger’s larger history of philosophy; much also depends on his later notion of being and on his attempt to demarcate that questioning from metaphysics (GA 65: 173). I cannot do justice to these issues here, but perhaps the best that can be said is this. Heidegger’s later work valorizes a mode of agency that is intended to allow a genuine receptivity to that which it encounters: listening, for example, is identified as prior to speech (GA 12: 142, 165) and moods such as reservedness [Verhaltenheit] supposedly prepare us to notice what is otherwise crowded out (GA 65: 33–34). As Kisiel observes, in this sense there is a close continuity between it and his earliest work, immediately after World War I, when he grapples with the issue of receptivity: Do we really apprehend, grasp, take . . . the immediacy of experience in its sense? Instead of Hinsehen, a Hingabe, a receptive submission: heeding, and not looking, more of a suffering than an action? Or somewhere in the middle, that Greek voice which will continue to recur as Heidegger moves from Paul’s verbs of God to Aristotle’s search for a middle between passion and action? (Kisiel 1993: 226) Much of Heidegger’s later thought turns around a sustained reflection on this stance between passion and action. My suggestion is that he is attempting to solve the very same problem that occupied him in Being and Time, namely how to “secure the right access” to that which he is investigating (SZ: 15). How, in other words, can one be genuinely open to what one encounters given the insight that all inquiry comes with presuppositions? In that sense, Heidegger’s later philosophy of agency remains first and foremost an exercise in phenomenology: it is an attempt to model a stance in which the phenomena might show themselves as they are, undistorted.10

Related topics Chapters 2 (on Pfänder & Husserl), 17 (on Dreyfus), 24 (De Monticelli).

Abbreviations I use the following abbreviations for Heidegger’s work. The pagination refers to the standard German edition which is given in the margins of all English translations. “GA” references are to the Gesamtausgabe edition (Frankfurt: Klostermann, 1975–;). Where translations exist, I have consulted these but often modified them; the relevant translations are listed below. Gel Gelassenheit, Pfullingen: Neske 1992; Discourse on Thinking, trans. J. Anderson and E. Freund, New York: Harper and Row 1966. NII Nietzsche. Zweiter Band, Pfullingen: Neske 1982. 107

Sacha Golob

SZ Sein und Zeit, Tübingen: Niemeyer 1957; Being and Time, trans. J. Macquarrie and E. Robinson, New York: Harper & Row 1962. GA 5 Holzwege (1977); Off the Beaten Track, trans. Julian Young and Kenneth Haynes, Cambridge: Cambridge University Press 2002. GA 7 Vorträge und Aufsätze (2000). GA 20 Prolegomena zur Geschichte des Zeitbegriffs (1979); History of the Concept of Time, trans. T. Kisiel, Bloomington: Indiana University Press 1992. GA 25 Phänomenologische Interpretation von Kants Kritik der reinen Vernunft (1995); Phenomenological Interpretation of Kant’s Critique of Pure Reason, Parvis Emad and Kenneth Maly, Bloomington: Indiana University Press 1997. GA 56/57 Zur Bestimmung der Philosophie (1999). GA 61 Phänomenologische Interpretationen zu Aristoteles (1994). GA 65 Beiträge zur Philosophie (Vom Ereignis) (1989); Contributions to Philosophy (From Enowning), trans. Parvis Emad and Kenneth Maly, Bloomington: Indiana University Press 1999. GA 77 Feldweg-Gespräche (2007).

Notes 1 For recent discussion see Clarke (2016), Smith (2005) and, from a more empirical standpoint, Murray, S., Murray, E., Stewart, G., Sinnott-Armstrong, W. and De Brigard, F. (2019). 2 These claims, often framed in terms of the “Incorporation Principle”, are central to the Kantian picture of action and ethics. For a highly influential discussion see Allison (1996: 129–133). I say that I am not “directly” responsible for impulses to avoid cases where I deliberately improve or worsen my natural inclinations, for example through some kind of training program – given Kant’s other commitments, such cases introduce complex questions that I cannot address here. 3 Crowell is discussing the original 2002 lecture presentation of Korsgaard’s position; I cite the later 2009 published version since it is more widely available. 4 This structure is set out in the Metaphysical and Transcendental Deductions of the first Critique which are thus pieces of logical analysis in Searle’s sense (Kant 1998). As Ameriks observes, it is frankly impossible to see how Kant’s talk of self-consciousness could perform any of the roles it does in those sections if it is understood in something like Dreyfus’s sense (Ameriks 2003: 12). 5 Despite the issues raised in the preceding section, Dreyfus (1991) remains by far the best introduction to these aspects of Heidegger and is the place to start for a fuller analysis. 6 Heideggerian identities are thus separable from institutional ones: Joan may continue to be employed as a teacher whilst having long lost any commitment to it; conversely, she may see life in terms of opportunities for teaching irrespective of whether she draws a salary for it. 7 On this reading, Wrathall’s account has similarities to views such as Smith’s where responsibility accrues to behaviors that reflect our evaluative judgments (Smith 2005:253). 8 For a rigorous debate centered around the role of public intelligibility in Being and Time see Olafson (1994) and Carman’s 1994 reply. 9 I discuss these issues in Golob (2014: 214–238). 10 I would like to thank the editors and an anonymous referee for their extremely helpful comments on an earlier draft.

References Allison, H. (1996) Idealism and Freedom, Cambridge: Cambridge University Press. Ameriks, K. (2003) Interpreting Kant’s Critiques, Oxford: Clarendon Press. Carman, T. (1994) “On Being Social: A Reply to Olafson,” Inquiry 37, 203–23. Chisholm, R. (1966) “Freedom and Action,” in K. Lehrer (ed.), Freedom and Determinism, London: Random House. Clarke, R. (2016) “Ignorance, Revision, and Commonsense,” in P. Robichaud and J. W. Wieland (eds.), Responsibility: The Epistemic Condition, Oxford: Oxford University Press.

108

Martin Heidegger Crowell, S. (2007) “Sorge or Selbstbewußtsein? Heidegger and Korsgaard on the Sources of Normativity,” European Journal of Philosophy 15, 315–33. Crowell, S. (2013) Normativity and Phenomenology in Husserl and Heidegger, Cambridge: Cambridge University Press. Davis, B. (2007) Heidegger and the Will, Evanston, IL: Northwestern University Press. Dreyfus, H. (1991) Being-in-the-World, Cambridge, MA: MIT Press. Dreyfus, H. (1993) “Heidegger’s Critique of the Husserl/Searle Account of Intentionality,” Social Research 60, 17–38. Dreyfus, H. (2005) “Overcoming the Myth of the Mental,” Proceedings and Addresses of the American Philosophical Association 79/2, 47–65. Gibson, J.J. (1979) The Ecological Approach to Visual Perception, Boston, MA: Houghton-Mifflin. Golob, S. (2014) Heidegger on Concepts, Freedom and Normativity, Cambridge: Cambridge University Press. Habermas, J. (1987) The Philosophical Discourse of Modernity, Cambridge, MA: MIT Press. Kant, I. (1998) Critique of Pure Reason, Cambridge: Cambridge University Press. Kisiel, T. (1993) The Genesis of Heidegger’s Being and Time, Berkeley, CA: University of California Press. Korsgaard, C. (1996) The Sources of Normativity, Cambridge: Cambridge University Press. Korsgaard, C. (2009) Self-Constitution: Agency, Identity, and Integrity, Oxford: Oxford University Press. Lear, J. (2006) Radical Hope, Cambridge, MA: Harvard University Press. McKinney, T. (2017) “Two Forms of Practical Knowledge in Being and Time,” in J. Capek and O. Svec (eds.), Pragmatic Perspectives in Phenomenology, Oxford: Routledge. Murray, S., Murray, E., Stewart, G., Sinnott-Armstrong, W., and De Brigard, F. (2019) “Responsibility for Forgetting,” Philosophical Studies 176, 1177–1201. Okrent, M. (1999) “Heidegger and Korsgaard on Human Reflection,” Philosophical Topics 27, 47–76. Searle, J. (1983) Intentionality, Cambridge: Cambridge University Press. Searle, J. (1999) “Neither Phenomenological Description nor Rational Reconstruction,” Revue Internationale de Philosophie 55, 277–84. Searle, J. (2004) “The Phenomenological Illusion,” in M. Reicher and J. Marek (eds.), Experience and Analysis, Kirchberg am Wechsel: Austrian Ludwig Wittgenstein Society. Smith, A. (2005) “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics 115, 236–271. Wolin, R. (1990) The Politics of Being: The Political Thought of Martin Heidegger, New York: Columbia University Press. Wrathall, M. (2014) “Autonomy, Authenticity, and the Self,” in D. McManus (ed.), Heidegger, Authenticity and the Self: Themes from Division Two of Being and Time, London: Routledge.

109

9 EDITH STEIN Psyche and action Antonio Calcagno

A quick scan of the scholarly literature on the philosophy of Edith Stein (1891–1942) shows a larger body of work devoted to her later philosophy, which mainly focuses on a rapprochement between phenomenology and medieval and Christian philosophy (see, for example, Manganaro 2002; Sharkey 2006, 2009; Maskulak 2007). One notes, however, a growing interest in her early, more strictly phenomenological work, especially her account of communal mind and social ontology (Moran 2004; Zahavi 2010, 2015; Calcagno 2014; Moran and Parker 2015; Szanto 2015; Ferran Vendrell 2015). There exist, however, only few sustained studies of Stein’s understanding of psychology and the lived experience of psyche (Ales Bello 2007, 2010; Betschart 2009, 2010). Stein’s essay “Psychic Causality” (Psychische Kausalität) forms the first part of her monumental Philosophy of Psychology and the Humanities (Stein 1922/2000). The essay may be read as an attempt to curb Husserl’s transcendental idealism, as found in the first version of Ideas I, and it may also be interpreted as a concrete analysis of how psyche conditions sense-making (Calcagno 2018). Here, I wish to highlight another important aspect of Stein’s discussion of psyche, namely its relation to action. I consider action in two senses. First, psyche acts upon consciousness, influencing how consciousness experiences reality and makes sense of it. Hence, psyche is an important building block of the act of sense-making (Sinngebung): for Stein, Sinngebung is not a purely logical process, as the first version of Husserl’s Logical Investigations suggests (cf. Stein 2014: 38–43). Second, psyche conditions how we ethically act: psyche is part and parcel of valuing and, therefore, conditions our ethical decisions and actions. Though I cannot show how psyche, for Stein, conditions all aspects of the life of the human person, I shall argue that psyche is vital for her understanding of action, understood in the two aforementioned senses. In particular, a hierarchy of values and acts of valuing cannot exist without the constitutive structure of psyche.

Psyche: what it is and how it manifests itself Generally, psyche must be understood as a constitutive aspect of the human person. The person, phenomenologically speaking, is defined as an I living through (erleben) a unity of body, psyche, and spirit (Geist):

110

Edith Stein

We have at least outlined an account of what is meant by an individual “I,” or by individuals. It is a unified object inseparably joining together the conscious unity of an “I” and a physical body in such a way that each of them takes on a new character. The physical body occurs as a living body; consciousness occurs as the soul of the unified individual. The unity is documented by the fact that specific events are given as belonging to the living body and to the soul at the same time: sensations, general feelings. . . . The living body. . . is characterized by having fields of sensation, being located at the zero point of orientation of the spatial world, moving voluntarily and being constructed of moving organisms, being the field of expression of the experiences of its “I” and the instrument of the “I’s” will. (Stein 1989: 56–57) Psyche works together with and is not separate (though it is distinct) from body and spirit. The realm of psyche has as its proper elements: a capacity to experience and make sense of sensory impressions, for example, the pleasure of the experience of a pleasant smell or color, feelings/emotions or affectivity, causality, an I-ness and we-ness, and the life-force. Though each of these elements may be said to belong properly to the realm of psyche, they do work together with the body and the spirit, resulting in the lived experience of a personal unity. Psyche comes to manifest itself in two primary modes: first, through an analysis of the structure of consciousness; second, through empathy (Stein 1989). In “Psychic Causality,” psyche first shows itself as an “influence” on consciousness: it acts upon and affects consciousness (Stein 2000: 14). Dialoguing with both Bergson and Husserl, Stein asserts that phenomenal consciousness must be understood as a continuous flow of pure becoming. One does not experience consciousness, she says, as a mere series of syntheses or as a serial connection of moments in the form of A + B + C+ D (Stein 2000: 9–10). Though there are times in consciousness when, for example, consciousness can become aware of specific objects, they appear as a “living and dying” of content (Stein 2000: 9–10). Consciousness is experienced as a living flow (Erlebnisstrom) and it signals its existence by how we experience certain contents, for example, the sensation of color or of a tone – what Husserl calls a sensuous impression (e.g. the softness of a color). Sensory data can arrange themselves and be enclosed by a field of consciousness. For Stein, it is the association of content that is vital, because psyche first appears with the connection between objects of consciousness. She describes the living and dying of content in consciousness as association by contact: The togetherness of different kinds of experiences in one momentary phase is the most original and premier kind of connection of experiences. Conversely, with the becoming of experiences out of phases continually overflowing one another, it makes no sense to talk about a connection. This togetherness is what phenomenally undergirds the term ‘association by contact’. (Stein 2000: 13) Consciousness is a complex (Komplexbildung) of togethernesses and successivenesses, and Stein (2000: 14) stresses that the pure becoming of consciousness is not produced as an effect. Finally, phenomenal consciousness presents its content in a unified fashion: it is a unifying consciousness. As consciousness gathers and organizes its content, it concomitantly tries to make sense of the whole and parts of the very content that it grasps.

111

Antonio Calcagno

Stein remarks that consciousness, at certain points, experiences an additional kind of association, which she calls an “operative influence” (Stein 2000: 14). Consciousness can experience itself as being influenced not only by content but by another kind of force; it may experience itself as weary, energized, and more or less alert. She notes that there are shifts in consciousness that are tied to life feelings, that is, feelings that stem from our physical, material living. For example, feeling tired makes our experience of colors less vivid, makes tones ring hollow. Stein identifies this shifting influence, which arises not from consciousness itself but from the vital force of life, as a causality that comes from the psyche. The causality of the psyche is described as being akin to natural, physical causality: a certain force or impetus, much like “mechanical production,” causes a reaction or produces an effect (Stein 2000: 15). Psychic causality, like any form of causality, is marked by an origin. Here we find the ancient and medieval notion of causality as an act that has a beginning and an end; it is understood as an originating event that carries with it a certain necessity. For example, fatigue causes consciousness to become dull. In addition to describing the psyche as marked by a certain form of causality, understood as cause-effect, stimulus-response, Stein characterizes the psyche as being experienced as a particular form of tying-together of experiences. Unlike the vital becoming of consciousness, the psyche ties its experiences together in a cause-effect structure, but whereas consciousness can become aware of itself as a pure experiencing, the psyche itself is never experienced or accessed as directly as consciousness. We feel the effects of the psyche operating upon consciousness, but we never see, feel, or experience the psyche itself. Hence, Stein (2000: 24) describes the psyche as transcendent. For Stein (2000: 24), the psyche also manifests life and a “life force” (Lebenskraft), especially when we experience the effects of “life feelings” (Lebensgefühle), such as tiredness, on consciousness. Stein’s phenomenology of psyche is marked by a profound vitalism, a vitalism that can be found in Bergson’s thought and that is taken up by later phenomenologists such as Michel Henry. Like Bergson, Stein claims that psyche causes us to consciously experience reality in more or less intense ways. She notes that, Differences of brightness for consciousness correspond to grades of tension for experiencing. The more intense the experiencing, the more luminous and alert is the consciousness of it. With that, it becomes quite clear that this being conscious, which we are claiming as a component of experiences, is not an experience itself, an act of conceiving reflection. For the more intense the experiencing is, the more ‘undivided’ [our absorption in it is likely to be, ] and the less it tolerates the splitting off of reflection. (Stein 2000: 18–19) Finally, Stein (2000: 23) claims that, unlike the pure ego of consciousness that accompanies and organizes the contents of intentional consciousness, the psyche has a real ego – that is, when we experience the effects of the psyche, we become aware that there is a subjective bearer of properties and psychic states, who feels and lives them. The real ego manifests itself in immanent data, but, like the psyche itself, it never becomes manifestly immanent – that is, it never appears directly to consciousness (Stein 2000: 23). Two questions arise from Stein’s account of the psyche. First, how does psychic causality differ from physical or natural causality, and second, how are psyche and consciousness related to one another? In response to the first question, Stein argues that psychic causality colors and affects the life of an individual, whereas natural causality is woven within a vast network of causal relations that extend throughout the physical world. She observes that 112

Edith Stein

[s]ensate [read psychic] causality differs from physical causality in the following way. With the latter, the unity of the causal occurring permeates the entire network of material nature, and single things emerge from that network as centers of occurrence. With the former, we’re confined to the sensate [read psychic] states of an individual, who as substrate of the causal occurring, corresponds to the totality of the matter while his or her properties emerge as single centers analogous to things. (Stein 2000: 25) For example, the introduction of bacteria may cause a certain infection to arise. In this case of physical causality, the bacteria and the resulting infection in a host are distinct realities that have broader connections to physical nature. In psychic causality, a certain event may cause a certain psychic disturbance, say the sudden loss of a loved one. Rather than a broader connection to nature, the individual person becomes the larger framework and his/her specific aspects, i.e. psyche and lived body, become the central focus of the lived experience of trauma. The trauma objectifies body and psyche as discrete things. In response to the second question, Stein contends that the psyche, understood as a substrate of consciousness, affects the capacity of consciousness to experience reality in varying degrees. For Stein, as for Husserl, consciousness structurally consists of (a) the possession of content, the data of consciousness; (b) an experience of the becoming or unfolding of the sense of that content; and (c) the consciousness of that experiencing, which always accomplishes it (i.e. experiencing) – in a higher or lower degree – and for whose sake the experience itself is even designated as consciousness (Stein 2000: 16–17). To the noema (a) and noesis (b) structure, Stein introduces a third aspect (c) of consciousness, namely, the force of the psyche, which ultimately “influences” the very capacity of consciousness to grasp the sense of things. Given that the psyche only manifests its effects, one can become aware of the psyche’s influence as a causal relation. Consciousness grasps the operation of the psyche as causal. Stein remarks that the real causality of the psyche manifests itself in the phenomenal causality of the experiential sphere. The enduring properties of the real ego, or sentient individual, appear as a substrate of the sensate (psychischen) causal occurrences which persists in a regulated changing of modes of those properties; so that a determinate power—lifepower—is singled out as both setting the mode of the others by its own momentary modes, and set in its own states by them in turn. The “effect” consists in the alteration of other sensate properties. There isn’t any direct causal dependence of other properties on one another without the mediation of lifepower. For example, receptivity for colors can be neither enhanced nor diminished by receptivity for sounds. Yet, the two can be enhanced together by an increase of lifepower that’s independent of both of them. Or, lifepower can be diminished by the activity of one, and in that way the other is diminished in turn. (Stein 2000: 25) To Husserl’s understanding of the foundational structures of phenomenology – namely, noesis and noema – Stein adds the psyche. It should be remarked that Husserl too, especially in his Ideas II, which Stein edited while she was his assistant in Freiburg, takes up the importance of psyche. But Husserl was very reluctant to publish this view while he was alive. Empathy provides a second gateway to understanding psyche and its acts. Stein conceives of empathy as a sui generis act of mind that enables one to enter and bring into relief the mind of another person through analogically comparing ego and alter ego. In this way, one 113

Antonio Calcagno

takes up the perspective of the other. I cannot develop a full account of Steinian empathy, as this goes beyond the scope of this chapter, but I do wish to emphasize that empathy allows one to recognize unique aspects of the structure of psyche. It does so by drawing upon the senses or meanings achieved though inner perception and the act of empathy itself. In each of these acts, psyche influences human consciousness and our person. Inner perception is distinguished from external perception insofar as the latter has an object in flesh and blood before it that lies outside of the perceiver, whereas the former experiences and grasps realities or events that occur within the human person such as a mood, feeling, or the lived experience of one’s own body. For example, I can perceive my own feeling of my sluggish body at the end of a long day. Careful study of inner perceptions helps me to grasp various aspects of my own personhood, including the various aspects of my lived body, psyche, spirit, and how they all work together as one. Again, it also allows me to grasp the various aspects described earlier through the study of conscious lived experience itself, including sensory impressions, feelings/emotions or affectivity, causality, and the life-force. Through its connection with the lived body, one can inwardly experience the effect of sense (taste, sight, sound) and sensory impressions (pleasure, intensity) on the psyche – for example, the feeling of intense vividness that comes from seeing the color red. But emotions also affect psyche. Joy or sadness affects our moods, and they come to express themselves in the lived body, but they are not reducible to the mere physical function of the lived body. What allows Stein not to reduce emotional experience to bodily functions is causality. The death of a beloved friend affects me and can produce a profound sadness. My face can certainly express my sadness, but the cause of the sadness is not certain bodily events, but rather the news of the event from the outside affecting my psyche. Understanding how causality works allows me not only to distinguish the lived body from psyche but also to understand the relation between distinct phenomena or aspects of myself and others. Finally, as mentioned earlier, one can perceive the operation of the unique effects of the life-force on our person. What is unique about empathy and its relation to psyche is that one can carry out acts of empathy in order to understand the psyche of another as well as understand how psyche works in general. I can see the drawn countenance of the other, tears welling in her eyes. Her body expresses something about her psychic state. Drawing upon knowledge from my own inner perceptions, I analogously compare what presents itself with my own knowledge, and I come to the conclusion that the other before me is terribly sad and devastated. In empathy, I bring the other’s state of mind to relief by comparing it with my own knowledge of myself. The other tells me that her brother has died. This is the final clue that allows me to confirm that she is sad. The act of empathy not only enables me to understand what the other is experiencing, namely, sadness and grief, but it also gives me insights into the nature of sadness itself, its psychic and bodily structure as it comes to express itself. Moreover, I also see how psyche is an important constituent of the personhood of the other. Empathy allows one to grasp the affective causal structure of ego and alter ego. Empathy is the building block of another form of consciousness, which Stein calls weconsciousness and which comes to manifest itself in the lived experience of community. The knowledge seized in empathy serves as material for a broader form of social cognition. In the lived experience of community, one grasps and lives in the experience of the community in solidarity. Stein gives the example of grief caused by the death of a beloved troop leader: The army unit in which I’m serving is grieving over the loss of its leader. If we compare with that the grief that I feel over the loss of a personal friend, then we see that the two cases differ in several respects: (1) the subject of the experiencing is different; (2) 114

Edith Stein

there’s another composition to the experience; (3) there’s a different kind of experiential current that the experience fits into. As to the first point, in place of the individual ego we’ve got a subject in our case that encompasses a plurality of individual egos. Certainly, I, the individual ego am filled up with grief. But I feel myself to be not alone with it. Rather, I feel it as our grief. The experience is essentially colored by the fact that others are taking part in it, or even more, by the fact that I take part in it only as a member of a community. We are affected by the loss, and we grieve over it. And this “we” embraces not only all those who feel the grief as I do, but all those who are included in the unity of the group: even the ones who perhaps do not know of the event, and even the members of the group who lived earlier or will live later. We, the we who feel the grief, do it in the name of the total group and of all who belong to it. We feel this subject affected within ourselves when we have an experience of community. I grieve as a member of the unit, and the unit grieves within me. (Stein 2000: 134) In the aforementioned example, psyche plays an important role. The emotion of sadness is what is experienced and grasped at the communal level, as a collective experience. Inner perception and empathy help us understand the nature of psyche and the role it plays for the affective and causal connection between me and other alter egos. But it is the lived experience of community that explains how affectivity and causality can help to establish a genuine we-experience by making us aware of the common cause of our collective grief, namely, the death of our beloved leader. In the end, the lived experience of community, the most powerful form of sociality for Stein, exemplifies the possibility that psyche can operate on a super-individual level. Stein discusses the example of a nation and argues that the force of the lifepower of certain individuals can act on other individuals by informing the life of the community as a whole. Stein remarks, No matter how much of their power the members of a community devote to the whole, the accumulated reserve obviously also depends upon what lifepower they can draw upon as single members, taken absolutely. One person with a very intense aliveness can accomplish more for her community when she places at its disposal only a portion of her power than someone else who places herself in its service with all her power. Thus, the level of lifepower of a community depends upon these two factors: the lifepower that its components can draw upon, and the amount of the power at their disposal that they devote to the community. Therefore, the power of a community can be increased in two ways: by receiving new powerful individuals, and by demanding more from those who already belong to it. Accordingly, it can be weakened in two ways: if its components drift off, and if the individuals belonging to it slacken in their accomplishments for the community. (Stein 2000: 206) A personality with a strong, vivid life-force can affect the psyche of the collective. But the opposite is also true that a moribund or sickly life-force can negatively affect a community or group.

Psyche and ethical action Thus far, we have examined how Stein conceives of psyche, how it acts upon consciousness, and how its actions affect the understanding of ourselves and psyche. In the foregoing 115

Antonio Calcagno

treatment, psyche largely comes to expression through the lived body, though it is not reducible to the body. The unity between body and psyche is intimate and both aspects mutually condition one another, but psyche also acts upon spirit, ultimately forming a constituent layer of all motivation. Ethical acts, for Stein, can be motivated, which means that they draw not only on psyche and motivation but also on reason and judgment. Spirit is defined as the lawful realm of human freedom (expressed in the I can of the will) and reason. The fullest or most meaningful kinds of human acts have their roots in spirit. Motivation is an act, and Stein attributes two senses to it. The first and most common sense is connected to acts of free will. All freely chosen acts are constituted by some deliberation and reflection. The justification of willed action in reflection, through reason, can push an action into being; motivation is the “force” of any freely willed act. Second, motivation may be also understood as a form of connection between the different contents of lived experiences: motivation sets one lived experience in motion in relation to another, connecting them. Stein views acts as being experienced, always from the standpoint of consciousness, according to three constitutive moments: apperception, synthesis, and being-set-in motion: In the realm of acts, we confront new means of connection that we haven’t yet encountered up to now. If the gaze points itself successively at a series of continually subsiding data, or rather points through the data at “external” objectivities, then we have not only a succession of detached apprehensions of a single shape, but one continuous apprehension, an appending of the later to the earlier (“apperception”), a combination of single apprehensions (“synthesis”) and a being-set-in-motion of the later by the earlier (“motivation”). All this makes no sense outside the realm of egoic acts. You can’t talk about taking, grasping, and moving in the sphere of pure passivity, which we were dealing with before. If we designate the connection of acts that we have in view here quite broadly as motivation, then we’re aware of departing from the customary linguistic usage which restricts this expression to the area of “free acts,” especially of willing. However we believe that this broadening is warranted, that what we now have in view is a structure valid in general for the entire range of intentional experiences, a structure that simply undergoes various configurations according to the particularity of the acts that adapt themselves to it. (Stein 2000: 40) Motivation can be part of an act, and it is viewed as that which may launch an act or sets it in motion: What is conventionally called motivation must be recognized as one such particular configuration. Motivation, in our general sense, is the connection that acts get into with one another: not a mere blending like that of simultaneously or sequentially ebbing phases of experiences, or the associative tying together of experiences, but an emerging of the one out of the other, a self fulfilling or being fulfilled of the one on the basis of the other for the sake of the other. The structure of experiences, which can enter into relationships of motivation all by themselves, is decisive throughout for the essence of those relations: that acts have their origin in the pure ego, emanate from it phenomenally and aim toward something objective. The “pivot” at which the motivation starts, so to speak, is always the ego. It executes the one act because it has executed the other. But the “execution” need not be taken in the sense of a genuine spontaneity. It’s characteristic 116

Edith Stein

of the relation of motivation that it can proceed in various forms. It can come to pass explicitly, but it can also be present only implicitly. (Stein 2000: 40–41) Stein maintains that the content of one moment of lived experience can directly and purposively bring about the content of another moment of lived experience. The sense derived from the transition of the moment before to the moment after, explicitly or implicitly, is one of a “being set in motion.” Here, we do not have a more blending or spontaneous flow of one act into another. The egoic motivational acts stand in a specific temporal relation to one another in terms of before and after; there is a strong connectivity between them, and one produces in the other a certain related and meaningful content. The association between various contents in motivation is different from that of causality. In the latter, psyche, for example, may bring forward or express an emotion, but the emotion is an automatic, driven response over which the I has no choice or will. In the case of the former, an emotion may affect us, but then through the power of reason, judgment, and ultimately will, we can choose or deliberate upon whether we wish to act based on that emotion. In this case, the emotion may signal an underlying value we hold about a person or state of affairs, for example, love and/or hate. Motivation is central to value and acts of valuing: The grasping of a value can motivate a disposition (for example, joy in beauty) and, accordingly, a wanting and doing (perhaps the realizing of a state of affairs recognized as morally right). Different kinds of acts. And yet something common to all of them: the ego executes the one experience – or the experience accrues to the ego – because the ego has the other, for the sake of the other. The commonality of the structure, as stated, isn’t altered even by the fact that motivating and motivated experience are, in a few cases, clearly set off as self-supporting acts following one after the other (like premises and conclusion, or grasping of value and will), and in other cases, bound in the unity of one concrete act (like self-grasping, and belief in perception). As motivating, here we always take the one experience for whose sake the other takes place. (Stein 2000: 42–43) Again, it is in valuing that psyche comes to expression. When we value something, we hold something to be dear or of worth to us. Values come to expression in loving, hating, in what we find beautiful and ugly. The values we have may come to manifest themselves within the realm of psyche. When we love someone of something, we become aware of the love through first being hit by certain emotions or feelings, which may also express themselves in the lived body. The love for one’s partner may, in various instances, induce certain psychic feelings and sensations that bring the objectivity of the love-value to presence. Love, understood as a value, may also motivate various possible free acts that I may carry out in the name of that love, for example, the decision to purchase a gift for the beloved or to write a poem. But values may, in turn, also inform certain feelings: The specific coloring of the feeling is dependent upon the particular kind of value. The feeling is insightfully and rationally motivated only insofar as it corresponds in all its dimensions to the value. Accordingly, whatever there is about the feeling that is not ‘owing’ to the value (its greater or lesser strength, perhaps) is unmotivated, uninsightful, and to be explained as merely the effect of the present life feeling. (Stein 2000: 76) 117

Antonio Calcagno

Psyche, in its relation to motivation and valuing, manifests the constitutive, affective layer of all valuing acts. Though a value may motivate a free act, which possesses a constitutive psychic layer, motivation alone does not fully determine the act. Motivation informs free will and the acts it undertakes. So, though my motivation may come to show itself, and though reason and judgment may come to deliberate on the motivation in question behind the act, the “I can” of free will can also reject what the motivation brings forward, hence confirming the freedom that is constitutive of the will. Stein observes: Certainly, that is often the case: but if I’m stuck in the struggle of conflicting motives, if I’m placed before a decision, still I am the one to whom the decision falls. The decision does not impose itself automatically, as the tipping of the scales toward the side of the “weightier” motive indicates. Rather, I make up my mind in its favor because it is weightier. Even if more can be said for the doing than for the abstaining, the doing still requires my “fiat!.” (Stein 2000: 55) It is in Stein’s treatment of free will that one also finds her most extensive discussion of action. The body manifests both voluntary and involuntary movements: something may cause the body to move or express itself in a certain way, for example, pain, but one may also deploy the will to make the body act in a certain way, for example, I can will to move my feet at a certain time (Stein 1989: 54). In the case of the lived body, Stein thinks of bodily action (körperliche Handlung) in terms of physical movement or locomotion, and also in terms of expression (Stein 1989: 48). The unity of body and psyche, especially in emotions, can come into relation with motivation and even strivings (Stein 1989: 56). We saw earlier how psyche, motivation, and rational judgments relate to one another within the framework of values and acts of valuing. The will, however, is identified by Stein as the source of action (Handlung). Stein defines action as being “always the creation of what is not” (Stein 1989: 56). She notes: To act is always to produce what is not present. The “fieri” of what is willed conforms to the “fiat!” of the volitional decision and to the “ facere” of the subject of the will in action. This action can be physical. I can decide to climb a mountain and carry out my decision. It seems that the action is called forth entirely by the will and is fulfilling the will. But the action as a whole is willed, not each step. I will to climb the mountain. What is “necessary,” for this takes care “of itself.” The will employs a psvcho-physical mechanism to fulfill itself, to realize what is willed, just as feeling uses such a mechanism to realize its expression. (Stein 1989: 55) The action of the will is described as a kind of creatio ex nihilo. The will may make use of psycho-physical realities and relations, but it is distinct from them. The will is described as the master of its domain (Stein 1989: 55), and what it enacts is an extension of that selfmastery or self-possession of the spiritual aspect of human persons: The will is thus master of the soul as of the living body, even though not experienced absolutely nor without the soul refusing obedience. The world of objects disclosed in experience sets a limit to the will. The will can turn toward an object that is perceived, felt, or otherwise 118

Edith Stein

given as being present, but it cannot comprehend an object not present. This does not mean that the world of objects itself is beyond the range of my will. I can bring about a change in the world of objects but I cannot deliberately bring about its perception if it itself is not present. The will is further limited by counter-reflective tendencies which are themselves in part body-bound (when they are caused by sensory feelings) and in part not. (Stein 1989: 55) The will is not to be understood in causal terms, though it can make use of causal relations, especially those between the psyche and the body, as was noted earlier in the case of motivation. The will may even be limited by certain psycho-physical processes, for example, fatigue. The fullest sense of an action, however, is that of a creative act of willing that brings something out of nothing: Action is always the creation of what is not. This process can be carried out in causal succession, but the initiation of the process, the true intervention of the will is not experienced as causal but as a special effect. This does not mean that the will has nothing to do with causality. We find it causally conditioned when we feel how a tiredness of body prevents a volition from prevailing. The will is causally effective when we feel a victorious will overcome the tiredness, even making it disappear. The will’s fulfillment is also linked to causal conditions, since it carries out all its effects through a causally regulated instrument. But what is truly creative about volition is not a causal effect. All these causal relationships are external to the essence of the will. The will disregards them as soon as it is no longer the will of a psychophysical individual and yet will. . . . Every creative act in the true sense is a volitional action. (Stein 1989: 56) For Stein, psyche is part of what it is to be a human person, to be living in the personal mode, so to speak. It is always acting, always within the unified working together of body and spirit, and it conditions the senses of ourselves as both body and spirit. Psyche must be viewed as acting within the flow of lived experience itself insofar as it is connected to motivation and valuing. It is in valuing, Stein says, that psyche can make us aware of values and we can reflect on the values that psyche manifests. But value, in turn, can affect how we live and experience certain values and even the flow of lived experiences. One of Stein’s most important contributions to phenomenology, in addition to her rich social ontology, is to show how psyche acts upon the lived body and spirit to help form and build what she calls the person. If ethics ponders the values of our actions, be they good or evil, one cannot comprehend the structure of our motivated and/or free choices and decisions without the assistance of psyche, which transmits the layers of affective and causal sense constitutive of value and valuing.

References Ales Bello, A. (2007) “The Study of the Soul between Psychology and Phenomenology in Edith Stein,” Cultura: International Journal of Philosophy of Culture and Axiology 4/2, 90–108. Ales Bello, A. (2010) “Causality and Motivation in Edith Stein,” in R. Poli (ed.), Causality and Motivation, Heusenstamm: Ontos, 135–149. Betschart, C. (2009) “Was ist Lebenskraft? Edith Steins erkenntnistheoretische Prämissen in Psychische Kausalität (Teil 1),” Edith Stein Jahrbuch 15, 154–183. Betschart, C. (2010), “Was ist Lebenskraft? Edith Steins anthropologischer Beitrag in Psychische Kausalität (Teil 2),” Edith Stein Jahrbuch 16, 33–64.

119

Antonio Calcagno Calcagno, A. (2014) Lived Experience from the Inside Out: The Social and Political Philosophy of Edith Stein, Pittsburgh: Duquesne University Press. Calcagno, A. (2018) “Edith Stein’s Challenge to Sense-Making: The Role of the Lived Body, Psyche, and Spirit,” in D. Zahavi (ed.), Oxford Handbook of the History of Phenomenology, Oxford: Oxford University Press. Ferran Vendrell, I. (2015) “Empathy, Emotional Sharing, and Feelings in Stein’s Early Work,” Human Studies 3/4, 481–502. Manganaro, P. (2002) Verso l’altro. L’esperienza mistica tra interiorità e trascendenza, Rome: Città Nuova. Maskulak, M. (2007) Edith Stein and the Body-Soul-Spirit at the Center of Holistic Formation, New York: Peter Lang Publishers. Moran, D. (2004) “The Problem of Empathy: Lipps, Scheler, Husserl and Stein,” in T.A. Kelley and P.W. Rosemann (eds.), Amor Amicitiae: On the Love that is Friendship. Essays in Medieval Thought and Beyond in Honor of the Rev. Professor James McEvoy, Leuven, Paris, Dudley, MA: Peeters, pp. 269–312. Moran, D., and Parker, R.K.B. (eds.) (2015) Early Phenomenology (Special edition of Studia Phaenomenologica: Romanian Journal for Phenomenology 15), Bucharest: Zeta Books, pp. 11–26. Sharkey, S.B. (2006) “What Makes You You? Edith Stein on Individual Form,” in J.A. Berkman (ed.), Contemplating Edith Stein, Notre Dame, IN: University of Notre Dame Press, pp. 283–300. Sharkey, S.B. (2009) Thine Own Self: Individuality in Edith Stein’s Later Writings, Washington, DC: Catholic University of America Press. Stein, E. (1989) On the Problem of Empathy, translated by W. Stein, Washington, DC: ICS Publications. Stein, E. (2000) Philosophy of Psychology and the Humanities, trans. M.C. Baseheart and M. Sawicki, Washington, DC: ICS Publications. Stein, E. (2014) Letters to Roman Ingarden, trans. Hugh Chandler Hunt, Washington, DC: ICS Publications. Szanto, T. (2015) “Collective Emotions, Normativity, and Empathy: A Steinian Account,” Human Studies 3/4, 503–527. Zahavi, D. (2010) “Empathy, Embodiment and Interpersonal Understanding: From Lipps to Schütz,” Inquiry 53/3, 285–306. Zahavi, D. (2015), Self and Other: Exploring Subjectivity, Empathy and Shame, Oxford: Oxford University Press.

Further reading T. Burns (2015) “On Being a ‘We’: Edith Stein’s Contribution to the Intentionalism Debate,” in Human Studies 38/4, 529–547. (This article looks at the discussion of group action with Stein’s social ontology.) F. Svenaeus (2016) “The Phenomenology of Empathy: A Steinian Emotional Account,” Phenomenology and the Cognitive Sciences 15/2, 227–245. (Provides an updated account of and situates Steinian empathy within current discussions in cognitive science.) F. Svenaeus (2018) “Edith Stein’s Phenomenology of Sensual and Emotional Empathy,” Phenomenology and the Cognitive Sciences 17/4, 741–760. (This article provides a good descriptive analysis of the psycho-physical unity of a person.)

120

10 ACTION IN THE PHENOMENOLOGY OF ALFRED SCHÜTZ Michael Barber

Levels of human activity, action, and motives Alfred Schütz (1899–1959) studied law and the social sciences in Vienna, and, using the work of Edmund Husserl, sought to provide a philosophical account of the social life-world to support the verstehende sociology of Max Weber. Schütz wrote his The Phenomenology of the Social World in Vienna, before fleeing Hitler’s Anschluss and coming to the United States in 1939, where he spent the next twenty years extending his philosophical work into such areas as semiotics, multiple realities, and assorted philosophical questions. Schütz’s theory of action stood in contrast to the behaviorism of his time by its typical phenomenological focus on how action and other forms of spontaneity are “experienced by the mind in which they originate” (Schütz 1962d: 209), that is, by a subject whose action is not to be studied merely as that of a physical object as one might do in the natural sciences (Schütz 2013: 217). Faced with the Cartesian alternative of mysterious inner processes pertaining to a non-observable “soul,” behaviorism opted instead for a methodology based on sensory observation that would avoid unverifiable “purposes” or motives and that instead reduced action to mere bodily responses to stimuli – more in accord with the procedures of the natural sciences. Schütz, however, argued that behaviorism operated with the unexamined prejudice of Cartesian dualism (choosing the body instead of the soul, though, as its investigative focus) rather than consider the object that the social sciences study, namely human beings in their everyday, social life-world, who understand each other quite well, encountering each other as fellow human beings rather than as “organisms” (Schütz 1962c: 55), without limiting themselves to either physically observable bodily actions or introspection into a mysterious inner sanctum of the other. One ought first to clarify what the object of the social sciences is, namely, the social world of everyday life, and then consider which methodologies are appropriate for studying that object. Having clarified the everyday life basis of the social science through The Phenomenology of the Social World, Schütz argued that that a more appropriate methodology than sensory observation of responses to stimuli would be ideal-type construction ala Max Weber that takes account of the meaning actions have to their actors – interpretations that are verifiable or falsifiable by the community of social scientists assessing such types’ accuracy. Schütz’s theory of action depends on breaking free from behaviorism and the dualism that spawned it to concentrate on a phenomenological 121

Michael Barber

description of the action that everyday actors regularly engage in and that this handbook entry will present more fully (Schütz 1962c: 52–63). In his The Phenomenology of the Social World, he initially employs the phenomenological reduction to become clear about internal time-consciousness of any actor, even though he dispenses with the reduction as he proceeds to examine social interaction and its structures (Schütz 1967: 43–44, 2013: 217, 252). Schütz (1967: 45, 51) portrays the inner stream of consciousness in terms of Bergsonian durée, understood as “a continuous coming-to-be and passing-away of heterogeneous qualities,” as undifferentiated experiences melting into one another. Although one lives in these experiences, one can cease immersing oneself in the flow and reflect on them, with the result that these experiences undergo modification and are apprehended and distinguished, brought into relief, and marked out from each other as discrete. Insofar as reflection succeeds upon and supplants the experience just preceding it, it presupposes that the experience on which it reflects is already past and elapsed. Furthermore, following Husserl, Schütz (1967: 48–50) explains how reflection has access to such just past experiences by pointing to a succession of “primary remembrances” or “retentions” that involve a being-still-conscious of the just-having-been and that build a bridge to the later reflection from the original now apprehension or impression, to which such retentions attach like a comet’s tail streaming behind its nucleus. Despite this connection, the reflective act, an act of “secondary remembrance” or “reproduction,” introduces a sharp discontinuity between itself and the experiences it grasps. Only when the reflective act falls upon streaming inner-time and grasps its experiences as well-circumscribed, past experiences, do those experiences become “meaningful” (Schütz 1962d: 210). Not all experiences are rendered meaningful so easily, since some experiences lie close to the private core of the person, and they cannot be grasped at all or can be apprehended only vaguely, and they are to a degree ineffable. These experiences, of which one can remember “that” they occurred and reconstruct only vaguely the “how” of their being experienced, are categorized as “essentially actual experiences” (Schütz 1967: 52). These include corporeal experiences, muscular tensions, physical pain, sexual sensations, and moods like joy, sorrow, or disgust. In addition, one is assailed by a “surf of indiscernible and confused small perceptions” (Schütz 1962d: 210) that provoke passive reactions, and one selects from this confused whole for pragmatic purposes some specific perceptions (such as some of the distinctive noises that emerge from the sea when one listens to it) and makes them clear, converting perceptions into apperceptions. Such small perceptions, which correspond to the “unconscious” of psychoanalysis, often determine without one’s knowing it many of one’s activities. For instance, if one is walking and talking with a friend and turns to the right rather than the left, such a movement might be prompted by a state of uneasiness about what is on one’s left. This unease, of which one may be barely aware, if at all, and which originates in small solicitations deriving from confused small perceptions, results in a movement that does not depend on entertaining any alternatives or making any choices (Schütz 1962d: 210–211, 1964a: 78, 1967: 52–53, 2011: 82; , 2013: 237, 269–270, 285–286, 292). However, beyond these experiences that are undergone or suffered, subjectively meaningful experiences arise insofar as one spontaneously takes up an attitude toward them, as when one fights against, suppresses, or yields to the pain one finds oneself beset with (Schütz 1967: 54). This type of “meaning-endowing experience of consciousness,” as Husserl (1983: 129) describes it, which becomes meaningful by being singled out from durée, also introduces a new level of meaning insofar as, for example, the pain now has the sense “something to be fought against.” One can find such meaning-conferring experiences in the automatic 122

Alfred Schütz

activities – habitual, traditional, and affectual – which Leibniz denotes as “empirical behavior” (Schütz 1962d: 211). This kind of spontaneous Ego-act, one species of spontaneous activity (or “conduct” (Verhalten) in Schütz’s terms), does not take place with the kind of intent typical of the other species, namely, “action.” When such unplanned spontaneous activity occurs overtly, through a bodily intervention, it is known as “mere doing,” and when covertly, “mere thinking” – either of which can be evoked by small perceptions (Schütz 1962d: 211, 1967: 55–56, 2011: 82). Another kind of conduct is “action” (Handeln), which is based on a project devised in advance, whether overt or covert (Schütz 1962d: 211). Should one covertly entertain a project with no intention to realize it, one engages in mere phantasying or daydreaming. But when an intention to realize supervenes, one converts the project into a purpose, as when one sets about solving a scientific problem mentally, and such a covert (i.e., without physical expression) purposive action becomes a “performance” (Schütz 1962d: 211). Purposive overt actions are known as “working,” that is, a bodily engagement with the world to bring about a projected state of affairs (Schütz 1962d: 211–212). These levels of personal activity underpin distinctions made in Schütz’s 1937 manuscript “The Problem of Personality in the Social World,” in which he distinguishes four types of “pragma,” or activities: pragma without a project or its conversion into a purpose to be realized, such as unconscious reactions evoked by stimuli (such as the knee-jerks, blinking, facial expressions, or gait that would fall under the essentially actual experiences described earlier); pragma with purpose but without project such as spontaneous habitual, traditional, or affective activity; pragma with both purpose and project, which is action in the full sense; and pragma with project but no purpose to realize it, that is, mere phantasying (Schütz 2013: 278). Moreover, Schütz shifts the Husserlian distinctions between types of knowing, such as memory or anticipating the future, in the direction of his own theory of action. Thus, in discussing the past-directed intentional orientations of memory, Husserl separates immediate retention (primären Erinnerung) and reflective reproduction or recollection (Wiedererinnerung), which, as mentioned, transforms undifferentiated experiences into discrete, elapsed experiences (Schütz 1967: 45–52, 2004: 141–142). These past-oriented intentional structures have their future oriented correlates: protentions (Protentionen), which intercept what is immediately coming, and the more reflective, reproductive-like, long-range, foreseeing expectations (Vorerinnerungen) (Schütz 1967:57–58, 2004: 153). It is the reflectively shaped project, known in advance by such future, long-range expectations, which, converted into a purpose by the intention to realize it, constitutes the goal that guides the actions and sub-actions taken to realize it. Schütz elaborates on the shaping of this project in imagination and its role in motivating action. Distinguishing the “action (Handeln),” that is, the ongoing activity, from the completed “act (Handlung),” Schütz states that one begins by phantasying the completed act as the goal toward which one’s action will lead. One imagines that act as completed, but in the future, focusing on what will have resulted because of one’s action – thinking, in other words, in the future perfect tense. This pre-phantasied state of affairs to be brought about by an action yet to come constitutes the “in-order-to” motive of the action. This future-directed motive motivates the voluntative fiat!, the decision to transform one’s phantasied act into covert performing or into an action gearing into the outer world. As a result, the project guiding the action-steps taken to realize it serves as the “primary and fundamental meaning of the action” (Schütz 1964b: 11, 1967: 88–89, 2011: 78–79). Of course, one usually already has had past experience of the act that one phantasies and the means leading to it insofar as one has already executed such a project oneself or seen 123

Michael Barber

others doing so, and consequently one has the same sense about one’s phantasied project that one has about most pragmatic behavior patterns that one has mastered in everyday life, namely that one “could do it again” (Schütz 1962d: 224, 1967: 90). In addition, there can be differences between the perspectives of an actor, who knows the in-order-to motive of her working act, and of an observer who may have access only to a limited temporal segment of such an action. The observer might see an actor complete the writing of a letter without knowing that finishing that letter was only a sub-act in pursuit of the more extended inorder-to motive of applying for a fellowship. In such a case, the observer, from an objective standpoint, would see only the product of the actor’s bodily engagement (the letter) without understanding the full subjective meaning of the actor in producing it. Further, sometimes the product-outcome may turn out differently than the actor intended or fall short of what was intended, but an observer unacquainted with the subjective meaning of the actor may not be able to appreciate such discrepancies (Schütz 2011: 85–86). There is another type of motive, however, to be found when one looks backward to past events that may have influenced one’s adoption of the in-order-to motive that guides one’s action toward the future. Hence while the bank-robber may be focused on realizing the act of robbing the bank that he phantasied in future perfect tense and decided to pursue, after the bank robbery (or during a pause in its execution), he might pause to reflect on what events in his past might have influenced his decision to rob the bank. He might, for instance, consider the violence of his upbringing as having motivated that decision, as, therefore, the “because motive” of his deciding to rob the bank. Insofar as his upbringing lies in a past more past than the past of his decision to rob the bank, looking for because motives requires one to think in “pluperfect tense” (Schütz 1967: 91–96, 2011: 79–80). Relying on this distinction between motives, Schütz opposes the utilitarian assumption that economic preferences derive from feelings of unease that require one to act in such a way as to satisfy them. On the contrary, Leibniz’s small perceptions, un-apperceived, unattended to, and unreflected upon, produce the unease prompting one to turn to the right in the garden rather than left and thereby provide a disposition to act, out of which all actions without deliberation originate. To the extent that these influences on action are recoverable by a retrospectively directed reflection, they would be grasped as because motives, but they cannot provide the final explanations of what determines activities, in particular, of a particular future-directed actor who rationally evaluates several projects, decides upon one, and commences to act on the basis of a supervening volition – all of which can be located in the domain of “in-order-to” motives. Schütz’s distinctions make possible rational action, which is taken as archetypal for all economic acts (Schütz 2011: 81–84). In addition, Schütz’s two types of motives interlock with each other in the idealization of the “reciprocity of motives.” When one requests information from interlocutor, the inorder-to motive of one’s request is to obtain information, and one assumes that one’s interlocutor will understand one’s action and be led to act in such a way (to find the information) that one will understand the interlocutor’s consequent behavior as a response to one’s query. In effect, one’s in-order-to motive will, in turn, serve as the because motive (discoverable in retrospective reflection) launching the interlocutor’s action whose in-order-to motive will be to secure for one the information requested. Many additional assumptions accompany this “idealization of the reciprocity of motives,” namely that one’s in-order-to motives will become the because motives of another, such as that the other’s motives will be like one’s own and those of others, according to one’s typical experience of how such motivations work in typical circumstances (Schütz 1962b: 23). In fact, Schütz (1964b: 14) affirms, “The prototype of all social relationship is an intersubjective connection of motives.” 124

Alfred Schütz

Essential features of action: typifications, relevances, and temporality The essential features of action, as this section will explain, consist in typifications (shaped by the past, oriented toward the future, automatized, and socially interrelated), relevances (of various types such as topical, motivational, interpretive, intrinsic, and imposed), motivations (“in-order-to” and “because”), and temporality. Typifications are a central constituent of Schütz’s conception of action. Whenever one experiences any object, such as the child experiencing a scissors for the first time, as Husserl has shown (Husserl 1960: 111), one transfers the meaning “scissors” and their power to cut to the next object encountered that resembles the first scissors. One acquires such a typification not only by repeating one’s experience, but cultures bequeath to their members whole sets of typifications, especially through language vocabularies, by which one knows “mountains,” “trees,” “salespersons,” and knows what to typically expect when one encounters such typified objects. Such typifications are taken for granted by “everyone” in one’s culture and have an objective, anonymous character, as if they are applicable regardless of one’s personal biography. Whole trains of action and routine behaviors, too, are typified, such as “going to the post office to mail a letter” or “brushing one’s teeth,” and are contained within the stock of knowledge that one possesses as part of the “biographical situation” that one brings to new experiences. One assimilates these new experiences under the typifications one already has acquired, interpreting something that looks like a dog to one’s previously formed typifications of a dog, including one’s typifications that if one mistreats this animal, it might respond by biting one, as previous dogs have behaved. Because of the typifications in one’s stock of knowledge, one operates with the idealization, that “I can do it again” as one classifies and acts in regard to new experiences (similar to past ones) or undertakes new projects of action (similar to a degree to those executed before). This idealization is basic to all pragmatic mastery of the conditions of everyday life. In sum, one phantasies one’s future project of the same type as previously performed acts that were effectively executed in response to circumstances similar in type to those one encounters in the present (Schütz 1962a: 69, 74–75, 2011: 126). This typical structure of projected acts enables Schütz to distinguish mere fancying from the projection of performances or overt actions. In pure phantasy, one is not impeded by the limits imposed by reality, and one can imagine oneself freely flying through the air or outrunning a speeding automobile – one thinks “in the optative mode” (Schütz 1962a: 73). However when one phantasies a typical project to be implemented in real life, the ability to achieve such a project, or its “practicability,” is “a condition of all projecting which could be translated into a purpose” (Schütz 1962a: 73). One thinks in a “potential mode” (Schütz 1962a: 73), that is, one seeks to be as sure as possible that the means and ends necessary are within one’s reach, and one recognizes that one cannot count on support from situational elements beyond one’s control and that the project itself would have been feasible and its means and ends available if the action had occurred in the past. One’s knowledge of all these factors depends upon one’s past experiences of typically similar projects and the means and ends involved in realizing them. As a result, it is practicability that distinguishes projects phantasied without regard for their implementation from projects phantasied with the intention of enacting them in the real world, and one determines what is practicable on the grounds of one’s past experience of accomplishing typically similar projects. The typicality of projects and the means and ends connected to them is, then, the key to whether similar projects and means and ends are considered practicable (Schütz 1962a: 73). While typifications are formed by past experience, they also anticipate the future. Operating with a typified project as the in-order-to motive of one’s action, one anticipates that 125

Michael Barber

the project will unfold in the future as it has in the past, unless counter-evidence surfaces. At the same time, however, the protentions and anticipations of what will be happening in the course of an action “are essentially empty references to the open horizons that may be fulfilled by the future occurrences or may . . . ‘explode’” (Schütz 1964c: 286). Consequently, while one’s present project is similar to previous projects, any project, including its circumstances and the course of its unfolding, will never be exactly like another (Schütz 1962a: 73, 1964c: 285–287, 293). There is, then, a horizon of indeterminacy that accompanies any action projected according to its type. On the one hand, it is likely or presumable that a project of this or that type will go through, and its typicality makes it somewhat predictable. On the other hand, whatever the final outcome of a project may be, it itself will be unique, different in the details of its progressing and results from every project similar to it. As Schütz asserts paradoxically, “in common sense thinking of everyday life whatever occurs could not have been expected precisely as it occurs, and . . . whatever has been expected to occur will never occur as it has been expected” (Schütz, 1964c: 287). This openness to novelty that lies at the core of any typification counterbalances the stabilizing tendency of typifications insofar as they function as valid until counter-evidence appears and tend to support predictability at the expense of flexibility and spontaneity – a concern that Jan Strassheim (2016: 494, 496, 500–503, 506–507) has voiced. Moreover, typified patterns of behavior can become automatized in such a way that they can act as a supportive substructure for higher level acts. Hence one can deploy language almost automatically, without giving it a second thought, as one pursues a higher level project such as requesting another to write a letter of recommendation for one’s fellowship proposal. Likewise, one can walk, eat, or smoke while one’s focus is on thinking through higher level practical or theoretical problems. In addition one can interpret automatically standardized signs, such as the gestures of a traffic officer, though Schütz cautions that it is possible to fail to understand others whether their actions conform or not to the assumed, standardized course of action. Finally, the typified routines underpinning higher level actions can experience interruption, as when, for instance, the pen with which one automatically writes runs out of ink, distracting one from the higher level thinking one might have been doing while writing and compelling one to attend to a new focus, namely filling one’s pen with ink (Schütz 1996: 126, 2011: 97–99, 173–175). While the idealization of the reciprocity of motives serves as a prototype of all social relationships, typifications, central to Schütz’s theory of action, also make it clear how social relationships pervade action. Insofar as the recognition of the other as an alter ego requires that one imputes to another, whose organism resembles one’s own, a life of consciousness similar to one’s own, this idealization of the general positing of the Thou amounts to a kind of typification of the other that one makes use of every time one encounters another like oneself and that is continually confirmed when interacting partners regularly conform with each other’s expectations, as happens almost without flaw, for instance, when the reciprocity of motives is implemented. Furthermore, many of the typical projects motivating one’s actions are acquired by watching others, by being trained by them, or by drawing on the project types deposited in the social stock of knowledge that social groups throughout history have built up and conferred approval upon. The elements within a social stock of knowledge are “taken for granted not only by me but by us, by ‘everyone’” (Schütz 1962a: 75), independently of one’s personal biography. Moreover, the typified projects and motives guiding one’s actions in everyday life belong to wider networks of socially constructed, maintained, and approved typifications that are shared by those who occupy socially maintained, typified roles such as those of doctors, priests, soldiers, and farmers everywhere, to 126

Alfred Schütz

name a few. Finally, there is a reasonableness to social relationships insofar as both parties to an interaction orient their actions in accord with certain socially approved, typified standards to which members of in-groups conform, including traditionally and habitually accepted and taken for granted norms, mores of good behavior, manners, organizational frameworks, or game rules. In other words, the very presence of typifications in projects, motives, and actions reflects the social groups that originate and transmit such typifications and that constrain and define the human actions taking place within their boundaries (Schütz 1962b: 32, 1964b: 13, 2013: 260). Action depends essentially not only on typifications but also on the interests at hand that shape thinking, projecting, and acting, and in particular the goals one seeks to attain. Certainly the in-order-to motive for the sake of which one acts constitutes an interest, or, as Schütz puts it, it is of relevance, or is a “relevance,” to the actor. The in-order-to motive, the final state of affairs to be brought about, understood as the paramount project, becomes then the ultimate motivational relevance for all the single steps and sub-acts that aim at realizing it. Likewise the in-order-to motive to realize each sub-act becomes the guiding purpose or relevance of each of the sub-sub-acts directed at realizing these sub-acts. Once the actor decides upon the governing in-order-to motive as an ultimate relevance, it diffuses value and relevance throughout a whole system of means, ends, and acts; and suddenly possibilities that had not existed before spring into existence or that previously had no relevance now become valuable (Schütz 1964d: 124). Since the in-order-to motivational relevance of any action is of such significance for it, it also becomes of pressing relevance to know whether it is practicable, even before one commits oneself to realizing it, at the least to avoid expending energy futilely. As Schütz remarks, “The practicability of carrying out the projected action within the imposed frame of reality of the Lebenswelt is an essential characteristic of the project” (Schütz 1964c: 289), and “The performability of the project . . . is the condition of all projecting” (Schütz 2011: 89). The judgment of whether an action is practicable, as a condition of projecting it, also relies on one’s typifications of similar past projects, their successfulness, and the resources required for their realization. If it is of relevance to seek to realize the project, then it is also of relevance to know if it is realizable at all. This concern for practicability, which as mentioned earlier, distinguishes phantasying in the optative mode from acting in the real world, is linked to the principle of scarcity in economic theory, which “establishes the limits, the frame within which the individual economic subject can draft his performable project. (Otherwise my fancy of a million dollars to spend daily would be economic projecting.)” (Schütz 2011: 89) The interconnection of projects and sub-projects, each as relevant to each other, suggests that there is no such thing as an isolated relevance, and, further any interest at hand belongs to a system of relevances that are part of a more or less structured hierarchy of relevances or even a plurality of systems that form an actor’s life-plans. To fully understand an action, one might inquire regressively into the ascending motivational relevances, the series of inorder-to motives, which would ultimately explain why one undertakes an action. Hence, for example, one might want to have ink to fill out an application to apply for a fellowship to allow one to complete a degree to find employment to support one’s family. Furthermore, one’s system of relevances continually undergoes modifications depending on the changing features of the action’s context, such that if a donor stepped forward and promised complete funding, the system of the actor’s relevances would be reconfigured, and the weight the author placed on certain motivations would be redistributed. Finding funding would then become irrelevant and preparing for the upcoming semester’s classes might emerge as one’s most pressing concern (Schütz 1964d: 125). 127

Michael Barber

Not only would preparing for classes become the preeminent in-order-to motive, but suddenly that project, which may have lain on the horizon of one’s attention when one was focused on finding funding, would become prominent as a new topic of attention, thereby assuming what Schütz would call “topic relevance.” Furthermore, when preparation for classes becomes the foremost topical and motivational relevance of the actor, she must also resort to a set of typifications that she may or may not have (and so have to acquire) in her stock of knowledge in order to carry out the project of preparing for classes such as typifications about what “credit hours,” “registration,” and “degree requirements” are. These typifications, now relevant for interpreting all that is involved in the new project, constitute “interpretative relevances,” which would differ from the interpretative relevances of securing funding, which might include finding and filling out sets of distinctive forms and meeting applications deadlines. One can see how Schütz’s three types of relevances, i.e. motivational, topical, and interpretative, might all be involved in elucidating an action (Schütz 2011: 107–123, 156–157). Moreover, the various social roles one occupies carry with them systems of relevances that are disparate and possibly even conflicting. For example, one’s relevance to support one’s family as the ultimate motivational relevance behind one’s applying for a fellowship coincides with relevances springing from one’s role as a parent and a spouse. In addition, though, one could further imagine how those relevances might conflict with the relevances attached to one’s role as a student, which might require spending large amounts of time in study that might subtract from the time available for one’s family. The relevances shaping an action are not necessarily homogeneous. As Schütz observes, the realms of relevances motivating actions are “intermingled, showing the most manifold interpenetrations and enclaves, sending their fringes into neighbor provinces and thus creating twilight zones of sliding transitions (1964d: 126).” Also, the socially defined roles and the relevances that are correlative to them and that motivate action reveal another way in which social relationships permeate action. Similarly, the typifications that become interpretatively relevant for carrying out in-order-motives (themselves often socially transmitted), such as the types of forms to be filled out and deadlines to be met for a fellowship, pertain to the social stock of knowledge of communities or institutions such as universities or funding agencies (Schütz 1964d: 125, 2011: 135). Furthermore, a dialectic between intrinsic and imposed relevances plays a key role in Schütz’s account of action. Intrinsic relevances are those that one has chosen or that one’s biographical situation outfits one with as one faces the potentiality of acting, and imposed relevances are those that are not connected with one’s chosen or biographically acquired interests, that do not originate in one’s decisions, and that one is unable to avoid, although one might be able to give a new meaning to imposed relevances that would accommodate them within one’s system of intrinsic relevances (Schütz 1964d: 126–127). For instance, running one’s business successfully depends upon the whole set of topical, motivational, and interpretative relevances that one has in hand, but when a major earthquake strikes one’s city, demolishing major parts of the building in which one conducts business, one finds oneself faced with new imposed relevances with which one must come to terms. Though at first one might interpret the earthquake as “utterly devastating leaving one hopeless,” one might be able to integrate the effects of the earthquake by cordoning off destroyed sections of the building, operating the business in intact sections, and applying for government assistance to rebuild – thereby revising the initial meaning one had given to the earthquake’s impact. Imposed relevances can be understood in two ways. They can be relevances of a second order, supervening upon intrinsic relevances already in place, as the example of the earthquake impacting one’s business shows. One can imagine other examples of second-order imposed 128

Alfred Schütz

relevances: terrorist attacks from a foreign power, the other person who is needed to bring one’s action to completion but whose relevances may not coincide with one’s own, the sudden onset of an illness or disability, the changing of one’s plans and the correlative discovery that one must now acquire additional knowledge for a new end, or the unanticipated outbreak of an airline employees’ strike just after one has agreed to deliver a paper in a distant country. Such imposed relevances interrupt the intrinsic relevances already guiding one’s action. Another type of imposed relevances precedes and sets the stage for whatever intrinsic relevances – whether they be topical, interpretational, or motivational – one will subsequently build up. These latter imposed relevances often appear as biographical or ontological conditions such as the situation and circumstances of one’s birth, including one’s parents, or the fact that once born one is destined to age and die. With regard to such relevances, intrinsic relevances appear secondary and derivative, supervening upon the base of imposed relevances already in place. Further examples of this second type of imposed relevances might include such things as the kind of body one was born with, its strengths or impairments (e.g. if one is color-blind), or one’s being taught a single language early in life without being exposed to other languages until later in life when one’s capacity to learn a new language is diminished. Such limits might impede or facilitate the kinds of projects and actions one might be able to adopt later. While imposed relevances can either interrupt intrinsic relevances or pre-exist them, one can also classify relevances as produced either by natural and ontological factors or by the free choices of oneself or others. Whatever action one undertakes, one will often encounter imposed relevances with which one must come to terms either by removing them, working around them, compensating for them, integrating them with one’s life-plans and/or revising, limiting, or even abandoning one’s intrinsic relevances in their face of such impositions (Schütz 1964c: 288, 1964d: 127, 1967: 128, 2011: 130, 193–194, 195, 196–199; Schütz and Luckmann 1973: 102–103, 114, 165). To understand action, it is also essential to take account of temporality. When Husserl (1960: 39–53; 1983: 9) describes intentionality, often with reference to the experience of perception, he describes continually unfolding internal time-processes in which one focuses on an object, whose horizons indicate indeterminate, but determinable horizons that one can proceed to explore, in a continual process that confirms or undermines one’s expectations. Similarly, as one sets about carrying out a phantasied project, the in-order-to motive and intentional object of one’s action, the sub-acts, also intentionally oriented, taken to see that project through to fulfillment, resemble the temporal continuity of perceptual intendings with reference to an object, insofar as these sub-acts are successful or thwarted over time. The interplay between intrinsic and imposed relevances often drives this ongoing intentional progression. One continually adjusts one’s action or recalibrates one’s projects in the face of relevances imposed by such factors as the “nature of things”; the sudden realization that one must wait to realize an act; the surprising discovery that an action might have unexpected, unwanted consequences; the vagueness of the guiding typifications that might not accommodate smoothly the unique circumstances encountered; or the emergence of a whole new project that might seem more appropriate. In this process, new topical relevances may emerge, motivational relevances can be altered or be modified, and new typifications may become interpretatively relevant for dealing with imposed relevances. Consequently, in the course of an action, it is usually the case, even if only in small details, that the typifications in one’s stock of knowledge and one’s system of relevances are constantly undergoing modification (Schütz 1964c: 286–287, 293, 2011: 156–160). In addition, in this temporal unrolling of an action, its series of sub-acts are experienced as “polythetically articulated syntheses of a higher order” (Schütz 1996: 85), that is, as 129

Michael Barber

stages toward one’s overarching action-project. However, after one has executed the action through its polythetic series of sub-actions, one can view the completed act retrospectively as a monothetic unitary act, as, for example, the completion of “one’s weekly shopping.” Furthermore, past, present, and future time segments are brought into interesting relationships with each other through action, with, for instance, the realized typified projects of the past affording a potential to which the agent in the present can always turn in order to project similar projects in the future. But given how one’s projected action faces imposed relevances and unexpected consequences, one’s knowledge at the time of projecting will be different after having performed the projected act insofar as an actor rarely escapes such impositions and surprising consequences and insofar as, at the least, one has grown older and acquired a raft of new experiences that one did not have in embarking upon the project. When one couples such unanticipated occurrences, along with the inherent vagueness of any typification of the project to be achieved, it is not surprising that, as mentioned previously, no outcome can be precisely foreseen and the outcome expected will never happen exactly as expected. Moreover, in the present when one faces the future, one operates with open and undetermined protentions and expectations; but, when looking back on the past, completed project, one surveys a field of fulfilled and completed anticipations, and the potestativeness experienced prior to acting has disappeared. Hindsight is considerably clearer than foresight. In fact, one can mistakenly overlook the difference in such temporal perspectives when, for example, one considers the outcome of action as economically irrational but forgets that, given the information available at the time of projecting and commencing the action, the decision to act may have been the best option available. These temporal modifications converge with Schütz’s distinction between the subjective point of view of an actor, such as that of the person planning a course of action, and the objective point of view of an observer who may have access to the outcome of the action without having understood the information available to the actor at the time of deciding to act (Schütz 1962a: 69, 1962b: 20, 30, 1964c: 286, 293, 2013: 226, 271). Having presented the essential features of Schütz’s theory of action, it might be instructive to consider his theory in relation to Donald Davidson’s more contemporary account of action within the tradition of analytic philosophy. Indeed, Davidson posits that a primary reason can be the cause of an action, with the primary reason being defined as a pro-attitude toward an action with a property that the action possesses under a description in which the action is typified as being able to lead typically to a certain, typified result – all within a broader context of typified goals, ends, practices, and conventions. Clearly, the pro-attitude and Davidson’s recognition of the importance of typifications suggest extensive convergence with Schütz’s notions of relevances and typifications so basic to his action theory. Likewise, Davidson’s recognition that it is not the agent’s desire for the result but rather the state of affairs that (in Schütz’s terms) will have been realized that provides guidance for the unfolding action and that nonteleological causal explanations of action are possible indicates a linkage with Schütz’s concepts of “in-order-to” and “because motivations.” Although Davidson admirably struggles against tendencies from natural scientific quarters to deny the causal force of reasons that can “make persons voluntary agents” (Davidson 1963: 700), his conflation of reasons with causes leaves him vulnerable to natural scientific incursions on his theoretical turf. It is just such incursions that Schütz resists by insisting on a theory not of causation but of “motivation,” developed by careful phenomenological analyses of how everyday actors perform in the life-world that precedes the adoption of natural or social scientific investigative attitudes. In addition, Schütz’s rich discussion of the intricate temporal frameworks within which in-order-to and because motives operate and are apprehended demonstrates the further benefits that a phenomenological approach can provide (Davidson 1963: 685–700). 130

Alfred Schütz

Ancillary considerations: choosing projects of action and rational action Striving to take account of the comprehensive context of action, Schütz investigates how one goes about choosing a project of action in the first place. Doubting or questioning about which of two or more projects to pursue arises because of conflicts within one’s system of relevances and, consequently, one is led to deliberate. Following Husserl’s account of how one might grow doubtful about what one once took to be a typical object given in perception (e.g. a human being) because of a succession of disappointed anticipations about how that object should have been given (which lead one to realize that it was really a scarecrow), Schütz discusses how projects of action approached as open possibilities can be converted into problematic ones. In the former, no one possibility has any weight over another and all possible courses of action seem equally open, with nothing speaking for one possibility over another. These open possibilities become problematic, however, when something speaks for one project over another and the projects appear competitive with each other, each with a different weight (Schütz 1962a: 77–78, 79–82; Husserl 1983: 250). While concurring with Husserl’s account of the ego’s oscillation between problematic understandings of an object, Schütz, applying the idea of problematic possibilities to choices between projects of action, insists that projects of action differ from objects since objects are ready-made and beyond one’s control, existing in simultaneity in outer time. Projections of alternative actions, however, are of one’s own making; one can control the constitution of one’s project; and projects do not yet exist at the time of one’s projecting them. Instead, the mind runs through a series of phantasying acts in inner time, passing from one problematic possibility to another, and returning to the first possibility with one’s consciousness altered by having run through the other possibilities that might suggest ways to modify that first possibility from which one began. The conflicting relevances of one’s biographical situation are at play in this transition from one version of a project to others insofar as any one version may result in different benefits and disadvantages than any other, with each version appealing to or posing problems to the assorted values in one’s relevance configuration. The ego, its motives, and the problematic possibilities compared are in a “continuous stage of becoming” (Schütz 1962a: 86) until the voluntary fiat! intervenes. This free decision, which Schütz (1962 a: 79–92) describes in terms of Leibniz’s “let us start,” occurs at the end of a complex process that depends upon previous activities, such as constituting what Husserl calls “problematic possibilities” in which something speaks for one’s side or other of a decision; Bergson’s view on the oscillation over time over the choices to be made, moving from one choice to a series of overlapping combinations gradually transitioning from one to another (which do not appear at first as two clear options); and finally Leibniz’s movement from soliciting inclinations to the counterbalancing of such inclinations to the final “in-order-to” motive that prevails when one chooses the project that motive governs. In this progression, the project becomes a purpose, and the free action detaches itself from the deliberative process like an overripe fruit from a tree. Only in retrospect does one often imagine the final choice as taking place between two distinct, clearly defined project choices, but if one reflectively transposes oneself back into the process occurring before the decision, one would appreciate the ongoing mixture, modification, and overlapping of a whole variety of projects, passing from one to another. The retrospective simplification of the preceding, messy flux as a choice between two well-delineated choices represents another version of how hindsight fails to do justice to the nebulous, indefinite character of what goes on prior to one’s decision. Schütz recapitulates this discussion in terms of Leibniz’s theory of volition in which one begins with inclinations 131

Michael Barber

to bring about a positively weighted project in the “antecedent will,” faces counterarguments emerging in the “intermediate will,” and arrives at the “final will,” in which the in-order-to motive of the prevailing project leads to the voluntative fiat of a will that can be characterized as “consequent, decretory, and definitive” (Schütz 1962a: 84–87, 88–91). To prepare for his discussion of rational action in economics, Schütz examines “rational action,” which differs from a “sensible action,” that is, one understandable to one’s partners, and “reasonable action,” which springs from a judicious choice among alternative courses of action. Rational action, by contrast, requires an actor to have clear and distinct insight into: the alternative means to an end, the relation of ends to other ends and means, the desirable and undesirable consequences of one’s action, and the relative importance of different possible ends. Rational action becomes even more complicated if one’s project of action involves another of whom one ought to know: her definition of her situation, her likelihood of understanding and cooperating with oneself, her knowledge of one’s project and its importance, her familiarity with one’s relevances, and her grasp of the interrelationship between one’s means and ends and one’s ends among themselves. Schütz recognized that such rational action represents an ideal that contrasts with what Max Weber called traditional or habitual acts and that is seldom to be found in a pure form in everyday life. However, social scientists and economists make use of this ideal of “purely rational action” as “archetypal of all economic acts” (Schütz 2011: 84), as mentioned earlier. Economists, as disinterested observers, construct models, economic “homunculi,” and fit them out with a stock of knowledge so that they could be construed to act as a human being would if the achievement of economic goals by economic means based on economic motives formed the exclusive content of her stream of consciousness. The scientist can then experimentally design how such ideal types would interact with others under different conditions such as those of unregulated competition or cartel restrictions (Schütz 1962b: 27–28, 30–32, 41, 1962d: 235, 1964a: 79, 2011: 84). Although Schütz’s general account of action is most applicable for everyday life, there are other types of action that can be found within what he calls various “provinces of meaning.” Hence, within the province of theoretical sciences, in disciplines such as economics or sociology, rational action becomes of particular importance. Likewise, there can be unique types of action, for example, in the religious or dramatic provinces of meaning, and within the province of phantasy, there are projects and actions that lack any purposive fiat.

Related topics Chapter 2 (on Pfänder and Husserl).

References Davidson, D. (1963) “Actions, Reasons, and Causes,” The Journal of Philosophy 60, 685–700. Husserl, E. (1960) Cartesian Meditations, trans. by Dorion Cairns, The Hague: Martinus Nijhoff. Husserl, E. (1983) Ideas Pertaining to a Pure Phenomenology and to a Pure Phenomenological Philosophy. First Book: General Introduction to a Pure Phenomenology, trans. by F. Kersten, The Hague: Martinus Nijhoff. Schütz, A. (1962a) “Choosing Among Projects of Action,” in A. Schütz and M. Natanson (eds.), Collected Papers 1: The Problem of Social Reality, The Hague: Martinus Nijhoff, 67–98. Schütz, A. (1962b) “Common-Sense and Scientific Interpretation of Human Action,” in A. Schütz and M. Natanson (eds.), Collected Papers 1: The Problem of Social Reality, The Hague: Martinus Nijhoff, 3–47. Schütz, A. (1962c) “Concept and Theory Formation in the Social Sciences,” in A. Schütz and M. Natanson (eds.), Collected Papers 1: The Problem of Social Reality, The Hague: Martinus Nijhoff, 48–66.

132

Alfred Schütz Schütz, A. (1962d) “On Multiple Realities,” in A. Schütz and M. Natanson (eds.), Collected Papers 1: The Problem of Social Reality, The Hague: Martinus Nijhoff, 207–259. Schütz, A. (1964a) “The Problem of Rationality in the Social World,” in A. Schütz and A. Brodersen (eds.), Collected Papers 2: Studies in Social Theory, The Hague: Martinus Nijhoff, 64–90. Schütz, A. (1964b) “The Social World and the Theory of Social Action,” in A. Schütz and A. Brodersen (eds.), Collected Papers 2: Studies in Social Theory, The Hague: Martinus Nijhoff, 3–19. Schütz, A. (1964c) “Tiresias, or Our Knowledge of Future Events,” in A. Schütz and A. Brodersen (eds.), Collected Papers 2: Studies in Social Theory, The Hague: Martinus Nijhoff, 277–294. Schütz, A. (1964d) “The Well-Informed Citizen: An Essay on the Social Distribution of Knowledge,” in A. Brodersen (ed.), Collected Papers 2: Studies in Social Theory, The Hague: Martinus Nijhoff, 120–134. Schütz, A. (1967) The Phenomenology of the Social World, trans. by G. Walsh and F. Lehnert, Evanston, IL: Northwestern University Press. Schütz, A. (1996) “Basic Concepts and Methods of the Social Sciences,” in A. Schütz with H. Wagner, G. Psathas, and F. Kersten (eds.), Collected Papers 4, Dordrecht: Kluwer Academic Publishers, 121–130. Schütz, A. (2004) Der sinnhafte Aufbau der sozialen Welt: Eine Einleitung in die verstehende Soziologie, in A. Schütz, M. Endress and J. Renn (eds.), Werkausgabe. Vol 2, Konstanz: UVK Verlagsgesellschaft, mbH, 75–450. Schütz, A. (2011) “Choice and the Social Sciences,” in A. Schütz and L. Embree (eds.), Collected Papers 5: Phenomenology and the Social Sciences, Dordrecht: Springer, 75–92. Schütz, A. (2013) “The Problem of Personality in the Social World,” in A. Schütz and M. Barber (ed.), Collected Papers 6: Literary Reality and Relationships, Dordrecht: Springer, 242–312. Schütz, A. and Luckmann, T. (1973) The Structures of the Life-World. Vol 1, trans. by R.M. Zaner and H.T. Engelhardt, Jr., Evanston, IL: Northwestern University Press. Strassheim, J. (2016) “Type and Spontaneity: Beyond Alfred Schütz’s Theory of the Social World,” Human Studies: A Journal for Philosophy and the Social Sciences 39: 493–512.

Further reading H. Nasu and F.C. Waksler (eds.), Interaction and Everyday Life: Phenomenological and Ethnomethodological Essays in Honor of George Psathas (Lanham, MD: Lexington Books, 2012). (Action in relation to others and social science.) J. Renn, G. Sebald and J. Weyand (eds.), Lebenswelt und Lebensform: Zum Verhältnis von Phänomenologie und Pragmatismus (Weilerswist: Velbrück Wissenschaft, 2012). (Contrasts phenomenological and Wittgensteinian approaches to action.) J. Strassheim and H. Nasu, H. (eds.), Relevance and Irrelevance: Theories, Factors and Challenges (Berlin: Walter de Gruyter, 2018). (Careful consideration of relevances and their importance for action.)

133

11 DETERMINED TO ACT On the structural place of acting in Sartre’s ontology of subjectivity Simone Neuber

Preliminary remarks John Atwell observes that the eclectic and disparate nature of Sartre’s thinking is nowhere exposed more extremely than in Sartre’s considerations on acting. As Atwell remarks: Very frequently, I find, Sartre begins a discussion with relatively familiar, i.e., “analytic,” considerations only to turn abruptly to thoughts of a quite different sort, sometimes idealistic and sometimes phenomenological. Often, I think, he makes a rather sudden turn in order to carry out an analysis which his starting point will not accommodate. Nowhere is this more evident than in his treatment of the nature and explanation of action, for there he begins with (1) considerations much like those urged by many current analytic philosophers, goes on to (2) views normally called idealistic, and finally depends on (3) theories propounded by phenomenologists. (Atwell 1972: 143) I am not convinced that this exposition is helpful. Talking about a digression from a systematically oriented starting point to idealistic and phenomenological theories obfuscates Sartre’s theoretical setup and thereby the place Sartre ascribes to his considerations on agency. Sartre does not start from a systematic reflection on acting in order to “digress” into “idealistic” and “phenomenological” theories. Sartre works on a unified theory of subjectivity – i.e. an ontology of subjectivity; and it is within that framework that actions find their place and that acting (in French: faire) is specified as the definiens of consciousness: “A first glance at human reality informs us that for it being is reduced to doing” (Sartre 1992: 612/521)1; “human reality is act” (Sartre 1992: 615/523). Atwell recognizes these concerns and is eager to correct this merely first impression to stress that the three strands are interrelated. Despite this, he nonetheless abstracts from Sartre’s cumbersome ontology as if Sartre’s notion of action could obviously be exposed without it. But this is far from obvious. What is obvious, however, is that this abstraction from Sartre’s ontology is quite common when Sartre is brought into a dialogue with “analytically” oriented philosophers. This move is understandable since it tames Sartre; but it does not reflect that Sartre – in our case 134

Jean-Paul Sartre

of action – reflects on the issue from within the standpoint of his ontology and that he even stresses that his reflection on action is the ultimate task of his ontology: “Ontology must be able to inform us about [the problem of Acting]; this is by the way one of ontology’s essential tasks if the for-itself is the being which defines itself by action” (Sartre 1992: 558 modified/475). The following exposition refrains from this abstraction in order to expose the systematic place of acting in Sartre’s overall theory and to reflect that Sartre’s take on acting is shaped by the question of why subjects are agents essentially (i.e. why they define themselves by action). Thus, I will not proceed from a glimpse at current issues in the theory of agency to ask how Sartre would relate to them. Rather, I will start with an extensive outline of Sartre’s ontology to then sketch the resulting take on acting.

The ontological framework: Sartre’s dualism and his notion of consciousness as néant According to Sartre’s dualism there are two irreducible regions of being: non-conscious being-in-itself (être en-soi) and the being of consciousness, i.e. being-for-itself (être pour-soi). Both appear in the title of his Essai d’ontologie phénoménologique as l’être and le néant. If one defines substance in terms of independence and as that which does not depend on anything else for its existence,2 then Sartre is no substance dualist. Only one of the two regions of being is independent: l’être en-soi. L’être pour-soi is sketched as a relational being presupposing an ontologically independent relatum in its being. Sartre hopes to ground this essential implication by outlining consciousness as a néant. Since this move is controversial and essential for Sartre’s theoretical reflections on acting, and since it serves quite a number of systematic expectations (SE), a brief reflection on this ontological determination may be useful. Translating néant as “nothingness” fails to recognize the fact that néant sounds like the present participle of néer (cf. Gadamer 1988: 47). Now, there is no such verb in French; but that does not matter here, since there is no proper German verb nichten either, and still, it is this Nichten Heidegger has exposed as the ontological mark of “das Nichts” (the nothing) in his What is Metaphysics? (cf. Heidegger 1976). According to Heidegger, it is wrong to say of das Nichts that it is; but admitting this does not imply that it has no mode of being; and according to Heidegger its mode of being is its Nichten (its nothinging). Heidegger forces us to note that “[d]as Nichts nichtet” (“the Nothing nothings”), and this is something we should hear sounding it Sartre’s néant. For Sartre, a néant exists as mere power to negate, i.e. pouvoir néantisant, which has its being in acting out this power. This ontological specification may be difficult to accept, but Sartre expects much of it. One can identify five SE that néant-structure is invested in, some of which will be important later on. (SE1) Sartre accepts Heidegger’s claim that any “negativistic” comportment, be it “the pain of failure” (Heidegger 1976: 117), “the bitterness of dispense” (Heidegger 1976: 117), or negative judgments, presupposes some basic understanding of nothingness, i.e. some preconceptual acquaintance. Sartre is happy with this idea – and this will be important for his considerations against determinism (see below) –, but unhappy with its theoretical articulation with Heidegger’s claim who grounds this acquaintance in a transcendence into “nothingness.” For Sartre, transcendence already is a negative comportment. Thus, if Heidegger’s initial claim is correct, there must be an even more basic source of this acquaintance which Heidegger misses in Sartre’s eyes. For Sartre, this more basic acquaintance is rooted in nothing but the – essential! – self-acquaintance of consciousness in so far as it exists as néant. 135

Simone Neuber

(SE2) Sartre relates this structure to what he gathered from his eclectic readings of Hegel. There he found a specification of what it means for something to exist for consciousness (in Sartre’s language: to be an intentional object) linked to a self-differentiating of consciousness from that which it intends and thereby determines as an object. As Hegel writes: “Consciousness simultaneously distinguishes itself from something, and at the same time relates itself to it” (Hegel 1977: 76). Given that Sartre accepts the Brentanian claim that intentionality is the mark of consciousness, and given that Sartre adopts this Hegelian (and also Reinholdian) notion of being-forconsciousness, intentionality is explicated as a complex structure containing (i) a determining of the thing as “not being consciousness” (Sartre 1992: 242/210), i.e. as non-moi (“non-ego”), and (ii) a determining of consciousness “as not being the thing.” But in Sartre’s eyes, this self-differentiating imposes restrictions on what can possibly manifest such a structure; and for Sartre, such a structure can only be manifested by a pouvoir néantisant, i.e. a néant. (SE3) Sartre aims at a theory of consciousness leaning toward realism and intrinsically immune to the constitutive idealism he finds in the transcendental Husserl. Here is how this negativism is meant to support this aim: consciousness is seen as a néant, i.e. a power to negate. But according to Sartre, any operation of negation is conceptually and ontically secondary in so far as it presupposes some positive given on which it can operate. Thus, consciousness as néant seems by necessity “born supported by a being which is not itself ” (Sartre 1992: 23/28). Given that what a néant can contribute to being is only negative and “ideal” (Sartre 1992: 264/228), it is full of a determining power but devoid of any ontically relevant productivity. Whatever contribution it makes to being, for Sartre, this contribution must be causally ineffective and ontically conservative. Thus, in so far as it is a néant, Sartre’s consciousness is sketched as something like a parasite living on and fed by robust being and beings. But still, and this is the fourth and related point. (SE4) Sartre takes this parasitical structure to be structurally and ontologically robust in so far as he regards negativity as underivable from being (cf. Sartre 1992: 57/57). With this notion of ontological underivability Sartre joins Husserl in assuming that consciousness is an absolute being. But whereas Husserl (1980: § 49) ties this absoluteness to the ontological priority of consciousness, Sartre reserves3 it to its ontological irreducibility which – concerning the ground of the néant – hints at consciousness’s spontaneous and self-determining structure: “in so far as it makes itself [se fait lui-même] it is an absolute” (Sartre 1992: 787/667). (SE5) Linked to this last point is a fifth: Sartre hopes to expose consciousness as beyond any causal determination. Assuming that negativity as well as negative comportments cannot be logically derived from or physically caused by being (or any positive given) and assuming that the theoretically intentional as well as the practical life hinges on negative comportments (such as positing something as non-ego, positing something as not yet the case or discovering that milk is lacking in the fridge), Sartre concludes that the intentional and practical life has to be logically and causally undetermined. It cannot but emerge spontaneously from that néant. In its radical spontaneity and in its transcending any psycho-physical causation Sartre regards a néant to manifest the structure of freedom: “there is no difference between [this mode of being] and being-free” (1992: 60/60). As controversial as these points may be, one (Megarian4) point is clear for Sartre: doing is vital for such a consciousness; as a mere power to nihilate/negate, its being (“existence”) lies in some kind of doing, i.e. in acting out its power. But so far, this doing is highly abstract and “philosophical,” and it is unclear how all this is meant to relate to (observable) bodily movements leading to alterations within the world’s arrangement, i.e. comportments like raising one’s arm or drinking a cup of coffee. But if 136

Jean-Paul Sartre

we are interested in a philosophical account of actions, we are interested in an account of the category into which those examples belong rather that in operations of consciousness even if they may be a vital dimension in Sartre’s claim that the being of consciousness is some kind of faire. Sartre has more to say about this category, but what he has to say is not independent from his ontological sketch of consciousness. First, especially (SE1) and (SE5) will be especially important. Further, Sartre expects a philosophical elucidation of acting to illuminate why subjects are agents essentially. For this, however, reflecting on the nature of action in isolation will not suffice. Rather, reflection has to start from a consideration of consciousness to expose why there cannot be a purely contemplative consciousness. However, since nothing can cause consciousness to do anything (SE5), the task is to expose why consciousness spontaneously determines itself as an engaged consciousness. This is how Sartre hopes to expose this selfdetermination as an agent thereby exploiting (SE2) and (SE4).

Toward essential agency: l’en-soi, le pour-soi, and the emergence of the value As Sartre grounds negativity in consciousness and takes the en-soi to be conceptually and ontically prior to consciousness, the ontological sketch of l’en-soi may not presuppose any negativity, hence a rather Parmenidean picture results (see also Gadamer 1988: 45). For Sartre, l’en soi is, amongst other things, (i) uncreated, (ii) non-relational, (iii) full positivity (being), and (iv.) a limit case5 of self-identity. As néant d’être, consciousness is meant to determine itself as not all this. It results a selfdetermination as (1) existence pour-soi (existence through itself ) or causa sui (Sartre 1992: 27/31), (2) essentially relational, (3) pure negativity, and (4) an existence which is not identical to itself. But for Sartre, consciousness does not only determine itself as not all this; it also determines itself as lacking these features; more concretely (and above all with respect to features 2–4), it determines itself as ontologically imperfect. As Sartre holds: “In its coming into existence human reality grasps itself [se saisit] as an incomplete being” (Sartre 1992: 139/125). For Sartre, this self-apprehension of consciousness as lacking ontological perfection is tantamount to its (pre-reflective) self-determination as engaged. This self-determination consists of three essentially interrelated moments: (M1) consciousness in so far as it apprehends itself as ontologically imperfect; (M2) the value as the background against which consciousness forms this apprehension; and (M3) properties of objects calling for and demanding operations in the world, i.e. motivational qualities (see Morris 2010: 147). Here is how they are interrelated from Sartre’s vantage point. The interrelation of M1 and M2 follows from the fact that, for Sartre, apprehending oneself as lacking an ontological perfection is never an immediate apprehension. Lacking something does not mean positively having a property (i.e. a lack or an imperfection) which could be intuited in isolation and self-ascribed. It means not to have a property in relation to something having it. Thus, the self-determination/self-apprehension is mediated; more precisely, it is mediated by something thereby taken as an ontological norm. This norm and the respective self-apprehension “form a dyad” (Sartre 1992: 141/127). Sartre calls the mediator of the respective self-apprehension in so far as it presents some norm the value (le valeur); given that it is the background for the self-determination of consciousness as being ontologically imperfect, the value is nothing less than what Sartre calls the perfect being (l’être parfait). The claim that apprehending oneself as an imperfect being presupposes the apprehension of a perfect being is, of course, familiar from Descartes’ third meditation, and Sartre is eager 137

Simone Neuber

to stress this.6 But Sartre’s indebtedness to Descartes is limited. Sartre only endorses one aspect of Descartes’ attempt to counter an objection to his argument from ideas, i.e. the objection that the idea of an infinite being (which Descartes’ first proof of the existence of God presupposes) might not be a true idea (vera idea, AT VII: 45; cf. Descartes 2007: 31) perceived directly, but only indirectly arrived at by negating a finite (per negationem finiti, AT VII: 45; cf. Descartes 2007: 31) of which one has a true idea only. To this, Descartes replies that it is evident (manifeste intelligo, AT VII: 45) that an infinite substance surpasses a finite substance in its reality and that therefore the perception of the infinite is prior to that of the finite; and at this point, Descartes alludes to the apparently obvious fact that I could not possibly account for any self-apprehension in terms of imperfections if not granting that such self-apprehensions take place before a background of a presupposed perfect being: For how could I understand that I doubt or desire – that is, lack something – and that I am not wholly perfect, unless there were in me some idea of a more perfect being which enabled me to recognize my own defects by comparison? (AT VII: 46; Descartes 2007: 31, modified) Descartes regards this line of reasoning not only as a piece of Scholastic metaphysics but as evident; and he has Sartre on his side – at least with respect to the claim in the last quote. But Sartre refuses to go further: he only grants epistemological priority to the perfect being given this factive self-apprehension; he does not grant it any ontological priority. The strategy of ridding himself of the ontological priority is vital for Sartre’s action theoretical turn. Sartre takes it by criticizing Descartes for regarding the thinker’s selfapprehension as imperfect as revelatory of a real relation to a perfect being which I (qua thinking creature) essentially have. For Sartre, this goes too far. All that is revealed in my self-apprehension as imperfect is that I now happen to understand myself in terms of something which I accept as a norm. Thus, it is my prior and implicit acceptance of the norm which must be acknowledged and theoretically accounted for. But this, for Sartre, does not imply that there is an object which I thereby accept or relate to and whose existence I implicitly endorse. But what functions as the required norm which I do accept in apprehending myself as imperfect? Although Sartre stresses that consciousness determines itself as imperfect with regard to l’être en-soi, l’être en-soi is not Sartre’s être parfait. The reason is that, notwithstanding that consciousness lacks its properties in not being an omnitudo realitatis, l’en-soi has deficits with regard to consciousness as well: it is not conscious and not spontaneous. But the perfect being is beyond both deficits. It does not lack robust being as consciousness does but enjoys the ontological privileges of l’en-soi (i.e. independence and full being). Further, it does not lack consciousness as l’en-soi does, but it is self-conscious and spontaneous. Thus, the implied perfect being is an omnitudo realitatis which is also a self-conscious causa sui; and this is why Sartre chooses to call this perfect being “God.” As stated previously, Sartre takes l’être parfait to be epistemologically prior to this selfapprehension. However, he regards any such being as conceptually and metaphysically impossible (cf. Sartre 1992: 140/126). Consciousness cannot but come as a néant; and a néant cannot be an omnitudo realitatis. Thus, Sartre goes beyond saying that the relevant norm is not necessarily identical to some pristine object. For Sartre, it cannot be identical to any such object since l’être parfait cannot have what Descartes has called formal reality. Thus l’être parfait cannot be ontologically prior. Consequently, we may not say that l’être parfait is presupposed by consciousness if this is taken to entail that it is implicitly posited by consciousness. This is due to the fact that l’être parfait is nothing which one could ever be posited in Sartre’s framework. 138

Jean-Paul Sartre

If this is the case, however, two tasks emerge: (i) the mode of givenness of l’être parfait must be determined given that it still somehow is for consciousness, albeit not as what Sartre would regard as an intentional object in an unqualified sense, and (ii) the kind of intentional relation that pertains to l’être parfait must be elucidated, given that it is not posited by consciousness but still mediates the self-apprehension of consciousness. In response to this double task, Sartre introduces three technical terms deserving further elucidation in the next section: the notion of haunting, that of a project, and that of engagement. With respect to the first task, Sartre claims that the value is given to consciousness in so far as it haunts pre-reflective consciousness (see Sartre 1992: 141/127). With respect to the second, Sartre claims that l’être parfait is projected by consciousness and that any such projection is tantamount to an engagement into realizing it. Thus, it is this very engagement which also gives an answer to the question of what it means to accept l’être parfait as the ultimate ontological norm.

On haunting, projections, and engagement The technical term haunting requires and the earlier “grasps itself as imperfect being” (se saisit) invites clarification. As for the latter, Sartre does not claim that each of us has conceptual knowledge of being ontologically imperfect (such that the imperfection would be posited). In fact, Sartre assumes we will not reach knowledge of this ontological deficiency even in ordinary reflection; such knowledge is only reached by pure reflection. The purity of pure reflection is due to its mere explicatory status of whatever is non-positionally but pre-reflectively understood (cf. Sartre 1992: 217ff/190ff ). Thus, se saisir alludes to this implicit non-positional and pre-reflective understanding of the ontological imperfection. Sartre speaks of a “lived” (vécu) ontological imperfection, a notion echoing (i) a phenomenal coloredness of this selfdetermination, (ii) that this ontological imperfection is existentially carried out, and (iii) that it is not reflexively given. Minimally, living something indicates that what is lived matters to the intentional life of consciousness without this mattering consisting in the fact that consciousness poses its imperfection as an object. The elusive notion haunt is Sartre’s technical term for a mode of givenness of that which is no object for consciousness but only has some “inapprehensible presence” (Sartre 1992: 364/312) so that it can affect the intentional life. Since for Sartre, being an object for consciousness is the correlate of positional consciousness or intentionality as implicit in (SE2), the beingtoward that which haunts consciousness cannot be an intending in this sense. For Sartre, it is obvious that each mode of givenness of – loosely speaking – “something” (if the given is an object or not) has a structurally correlative being-toward of consciousness. As we have seen in the last section, Sartre regards l’être parfait as conceptually and metaphysically impossible; it is no possible object for consciousness; and it is nothing which could possibly be posited by consciousness. But what, then, does consciousness do when being toward or for l’être parfait? Sartre’s answer is: it does not posit l’être parfait; it projects l’être parfait. Thus, to say that l’être parfait haunts pre-reflective consciousness is just the flipside of saying that consciousness projects l’être parfait. To say that the relevant being-toward is a projecting echoes two aspects: (i) first, it echoes an ontological dependence of the value on consciousness. As seen, the value is not equivalent to l’en soi which is presupposed by consciousness and ontologically independent, but a (confused) projection of a synthetic unity of the two poles of the intentional “relation” (en-soi and pour-soi). Its presence is as robust as that of cinematic screen images; if you switch off the projector, they vanish. (ii) A second claim resounds in Sartre’s notion of projection: the claim 139

Simone Neuber

that I would not apprehend myself as imperfect if I did not accept l’être parfait as a norm. But qua norm, l’être parfait is apprehend as worthy of being attained (that’s why Sartre calls it a value); and for Sartre, this very apprehension is nothing over and above the attempt to live up to that norm. That is: apprehending l’être parfait as worthy of being attained simply is striving toward living up to it. Thus, projecting the value is tantamount to engaging into the realization of value.7 And here we see why value matters for Sartre’s self-determination of the agent: for Sartre, being toward value simply is being practically engaged. With respect to this last point, Sartre’s notion of a project translates Heidegger’s notion of an Entwurf (i.e. a project in a practical sense) and circumscribes the ultimate “Worumwillen” or telos of consciousness, i.e. that for the sake of which consciousness ultimately lives and exists. As the telos of a striving, the value is projected as a possibility for consciousness, i.e. what Sartre calls the ultimate existential possibility. Obviously, Sartre does not thereby presuppose that existential possibilities are conceptually or metaphysically possible. They are just what consciousness projects and engages into.

Engagement, intentionality, and motivational properties Engagement is non-positional with respect to value. If one defines intentionality by the positional aspect along the lines of (SE2), then one may not call engagement intentionality. But semantically, we are not forced to do so but can also say that any being-toward defines intending. Doing so, we can say that engagement is a kind of intending: not positional intending (intending POS) but a non-positional projective intending (intending PRO). IntendingPOS is meant to be a mark of consciousness. It is not the only mark, however. For Sartre, consciousness is by definition intentionPOS and a non-positional and pre-reflective selfapprehension: “every positional consciousness of an object is at the same time a non-positional consciousness of itself ” (Sartre 1992: 13/19). Given the earlier point, we can add a third vital aspect and say: each consciousness is non-positionally directed toward a value, thereby living its ontological imperfection; or: consciousness essentially exhibits intentionalityPRO. In fact, however, the third aspect comes down to the second. IntentionalityPRO is a kind of self-awareness: it is how subjects non-positionally live their ontological imperfection while being non-positionally toward the value. Thus, it is a mediated non-positional self-awareness or, as one could say, one which is stretched out toward an ultimate existential possibility. This, as a matter of fact, is not only some kind of self-awareness; for Sartre, this is selfawareness properly understood. Not only is Sartrean selfhood defined by projecting. Furthermore, Sartre also cashes out any “state-consciousness” in terms of intentionalityPRO. For Sartre, a self-ascribable “mental state” cannot but be a haunter and a project of engagement for pre-reflective consciousness. In that respect, “mental states” are also values (see Sartre 1992: 122/111, 138/124; for more on this see Neuber 2017). According to Sartre, positional consciousness and pre-reflective self-awareness exhibit a two-way entailment. Thus, intrinsically two-dimensional intentionalityPRO (twodimensional with reference to the imperfection and the value) has its own positionality, the positionality of structures that seem to allow for a realization of the ultimate existential possibility. Thanks to intentionalitPRO, situations are apprehended as calling for certain modifications; they confront us with tasks (Sartre 1992: 274/236); they raise “pure demands which rise as ‘voids to be filled’ in the middle of [the world]” (Sartre 1992: 274/236), claiming “a right of the real” (Sartre 1992: 150/136). And this is where the above-mentioned last moment (M3) enters, i.e. the idea that there are motivational qualities, i.e. properties of objects calling for and demanding actions in the world. 140

Jean-Paul Sartre

Sartre tries to shed light on the constitution of those structures by giving examples that require a cautionary remark and a clarification. As for the former: the following may sound as if some previously intended world of objects suddenly gained a new, additional structure. This is not intended. Rather, the respective constitution of demands is primordial. As for the latter: the examples introduce a situational concretion8 of the abstract notions lived ontological imperfection and value. The former comes, e.g. as a lived thirst (elsewhere Sartre introduces a lived hunger or a lived tiredness), the latter as the concrete existential possibility against which9 the concrete “imperfection” has its specific nature (e.g. being saturated, being quenched, resting). This situational concretion follows from the assumption that the former rather abstract structure cannot but appear as situationally defined. It is the concrete situation I find myself in as an embodied being which allows for quite specific ways to engage into the value and to live my imperfection. Still, these examples maintain the two-dimensional non-positionality of intendingPRO. Sartre assumes that intendingPRO is related to a specific kind of intendingPOS in the following manner: The possible which is my possible [i.e. the existential possibility/value into which I engage] is a possible for-itself [i.e. a future state of consciousness in which the value is realized and in which it apprehends itself as the value] and as such a presence to the in-itself as consciousness of the in-itself [i.e. as such it intends some object]. What I seek [is this existential possibility, i.e. this consciousness]. But this possible which is non-thetically absent-present to present consciousness is not present as an object of a positional consciousness [. . .] The satisfied thirst which haunts my actual thirst [. . .] is a thetic consciousness of the glass [. . .] and a non-positional self-consciousness. It then makes itself transcend to the glass of which it is conscious; and as a correlate of this possible non-thetic consciousness, the glass-drunk (from) haunts the full glass as its possible and constitutes it as a glass to be drunk from. (Sartre 1992: 157/141, modified) What Sartre indicated in this rather complicated passage is that what I engage into is a (sought and thus a future) state of consciousness; abstractly speaking, it is one in which consciousness can apprehend itself as omnitudo realitatis. More concretely, it is some state in which a current thirst “grasps and incorporates repletion into itself ” (Sartre 1992: 138/154) to thereby “pass[] on to the plenitude of being.” But if the engaged-into is a future state of consciousness, the engagement necessarily is engagement into a positing, in this example, into a positing of a glass of water in so far as it is empty (or drunk from). This is crucial for the rather eccentric haunting of the value. What Sartre observes is that when the value (here: the future state of consciousness) haunts actual consciousness, then a haunting on the positional level takes place; more precisely, Sartre observes some superimposition of the intentional object of that future consciousness onto the intentional object of the current consciousness. The not-yet intended object into whose positional consciousness actual consciousness engages comes to infect the actual object of consciousness. It is due to that infection that the actual object gains a property, namely the property of being-to-be-emptied.10 It thus demands being drunk from. According to Sartre, those emergent structures are apprehended as intrinsically motivational: “their character [. . .] is manifested to the unreflective consciousness by a direct and personal urgency which is lived as such without being referred to somebody or thematized” (Sartre 1992: 274/236). Thus, being responsive does not mean noticing, but being practically responsive. Sartre’s subject unreflectively or pre-reflectively feels addressed and feels torn toward a practical response and, thus, simply follows the call of the objects to give them their right to be real. Responding to value in a concrete situation, reflection is completely bypassed. 141

Simone Neuber

Since, according to Sartre, intentionalityPRO and intentionalityPOS are interrelated (with respect to demands) and since apprehending demands is just being practically responsive to them, intentionalityPRO comes as praxis: Our description of freedom [i.e. consciousness] [. . .] compels us to abandon at once the difference between the intention and the act. The intention can no more be separated from the act than thought can be separated from the language which expresses it; and as it happens that our speech informs us of our thought, so our acts will inform us of our intentions – that is, it will enable us to disengage our intentions, to schematize them, and to make objects of them instead of limiting us to live them – i.e. to assume a nonthetic consciousness of them. (Sartre 1992: 622/529) This motivates a reflection reflecting on possible dangers for a theory of action. Sartre at one point observes: As soon as we formulate the problem of action, we risk falling into a confusion with great consequences. When I take this pen and plunge it into the inkwell I am acting. But if I look at Pierre who at the same instant is drawing up a chair to the table, I establish also that he is acting. Thus there is here a very distinct risk [. . .] of interpreting my action as it is-for-me, in terms of the Other’s action. This is because the only action which I can know at the same time it is taking place is the action of Pierre [or some Other]. I see his gesture and at the same time I determine his goal: he is drawing a chair up to the table in order to be able to sit down near the table and to write a letter which he told me he wished to write. (Sartre 1992: 422/359f.) Sartre’s approach to acting as a pre-reflective practical response to demands – Sartre speaks of lived intentionality – is a direct reaction to the danger articulated in the just quoted passage. Sketching acting as being practically responsive to a world of demands is thus Sartre’s attempted first-person approach. Sartre hereby formulates an approach where acting is nothing but a practical reading11 of a motivational structure – which is itself a reflection of existential possibilities. This practical reading bypasses reflection and any self-related positing, be it that of a project (or my project) or that of an objectively given desire. All objects which exist for consciousness are the objects out in the world calling for modifications within the world. For Sartre, it is only a secondary objectification from a third-person perspective that turns the lived lack into a “desire” and the lived value into an “intention” the subject “has formed.” And although sketching the details of this process of objectification is beyond the task of this chapter, Sartre is eager to stress that accounts explaining actions in terms of causally efficient beliefs-cum-desires simply fall into such a “confusion with great consequences” by ignoring a more fundamental layer of analysis.

The “structures contained within the very idea of acting” – Sartre’s “translations” The account developed earlier is the foundation for Sartre’s attempt “to make explicit the structures contained with the very idea of acting” (Sartre 1992: 559/477). The relevant chapter in Sartre’s magnum opus has three aims: (i) translating the earlier scheme into a more 142

Jean-Paul Sartre

“action theoretical” vocabulary in order to expose action as a manifestation of an internal tripartite structure, (ii) exposing how this structure can reconcile intuitions of both “determinists” and proponents of a liberum arbitrium indifferentiae, and (iii) providing the basis for a hermeneutics of agency which gives way to Sartre’s existential psychoanalysis at the end of Being and Nothingness. With respect to (i), Sartre starts by identifying three concepts he seems to take as generally accepted core notions of action theory: (1) le motif,12 (2) le mobile, and (3) la fin (the purpose). Sartre assumes that the former two notions are regarded as either unrelated or even as belonging to “two radically distinct layers” (Sartre 1992: 577/491), and his task is to show that this is wrong. Thus, Sartre’s opponent is someone quite distinct: it is someone assuming that le motif is that which justifies the act (cf. Sartre 1992: 575/490); that le mobile is the non-rational “ensemble of desires, emotions, and passions, which urge me to accomplish a certain act” (Sartre 1992: 576/491), and that an ideally rational act is one exclusively determined by un motif. Thus, the opponent is someone assuming that there can be acts which are only determined by le motif. For Sartre this is a poor abstraction which is blind to the constitution of motifs. To expose this, Sartre hopes to show that the three notions can be neatly mapped onto the earlier three moments M1, M2, and M3. The purpose takes the place of M2, i.e. of the value or of that into whose realization we engage. Le mobile takes the place of M1 and replaces lived urgency/lived ontological imperfection. Now the crucial point for Sartre is that le motif is in fact only a variant of M3, i.e. of objectively given motivational structures. But if the structure of le motif mirrors the constitutive structure of demands or objectively given motivational properties, then there is no tradeoff between le mobile and le motif; both are but moments of an interrelated complex structure of being toward a project. Given this interpretation of le motif along the lines of M3 and along the constitution of objectively given motivational structures, it is clear for Sartre that there are only actions performed in light of “reasons” (motifs). But there are never actions which we opt for by balancing “neutral” reasons. Instead, engaging into a project is finding a reason for that project – and there is no other access to practical reasons apart from that: “far from determining the action, [the motive] appears only in and through the project of an action” (Sartre 1992: 578/492). In a way, le motif is itself an “‘irrational’ fact” (Sartre 1992: 578/493). This view has consequences for Sartre’s take on deliberation, which for him is just a way to be engaged into the project. Putting theoretical weight on deliberation or assuming that there is something pristine to being determined by one’s best reasons appears as a mistake about the ontology of motifs, i.e. the mistaken assumption that motifs can be apprehended neutrally in order to weigh them against one another. For Sartre, however, what we will find in weighing pros and cons for a project does not inform us about the value of the project (since there is no such independent value) but about whether we have already started to engage into the project. “When I deliberate, the chips are down” (Sartre 1992: 581/495). Acting is thus not only trivially acting for a reason but also acting for the subjectively best reason, i.e. the reason on which we would have acted on had we deliberated. In that respect, mindless and blind acting and reflected acting only differ in the form of engagement.13

The limits of self-knowledge, freedom, and determinism Obviously, a Sartrean framework has no place for akrasia, weakness of will or an invitus facere. But of course, it has place for being mistaken about one’s intentions. My reflective selfascriptions may be guided by a self-image I wish to be true to, so that I self-ascribe the more 143

Simone Neuber

appreciated wish. When doing something that does not reflect a wish that I aspire to have, I might come to claim that I somehow acted against my real intentions. For Sartre, such claims do not reveal hidden intentions. They only reveal that I can reflect impurely. That we can be radically mistaken about our intentions, plans, and projects would not be surprising for Sartre. Nevertheless, there cannot be psychological motives or forces that prevent us from doing what we pre-reflectively intend to do (naturally, we can fail if someone or something hinders us). All we can be wrong about is our intentionalityPRO; and we are prone for mistakes once we start reflecting on it to ascribe intentionalityPRO in terms of concrete “intentions.” As Sartre suggested before: what ultimately reveals my intentions is not an act of reflection but what I have ended up doing. Although each project is “irrationally” chosen, what I end up doing it is not arbitrary. Rather, all my actions cohere in virtue of what Sartre calls an initial project (Sartre 1992: 588/501) or profound intention (Sartre 1992: 582/496) functioning as an all-encompassing frame of coherence. It is the ultimate “project of myself ” (Sartre 1992: 594/505) and projected in a “fundamental act of freedom” (Sartre 1992: 594/506). This initial project is usually not discovered in ordinary reflection and needs a hermeneutics of concrete subjectivity which Sartre’s existential psychoanalysis wants to provide. But apart from its function to expose action as ontologically meaning ful it is meant to help reconcile the correct but one-sided assumptions of “determinists and the proponents of a liberty of indifference [liberum arbitrium indifferentiae]” (Sartre 1992: 563/480, modified). Whereas the former assume that my acting can be entirely explained in terms of psycho-physical causal laws and forces, the latter assume that we can imagine a situation in which two agents end up doing different things although their respective psycho-physical determinants are identical. In order to explain this, they allude to an unconditioned power of the will. Sartre agrees with the libertarian that (i) psycho-physical forces cannot cause my actions. This follows from the fact that intentionalityPRO is a being-toward something which still lacks and thus a comportment implying negativity (see Sartre 1992: 560/478). According to (SE3), however, this is only possible for a néant which according to (SE5) exposes the structure of radical spontaneity and causal unreachability. Sartre furthermore agrees with the libertarian that (ii) there are projects that are not chosen in virtue of one’s reasons. But this does not mean that they are not made in the light of one’s reason. As soon as I engage into a possibility, there is a network of sufficient reasons; as a consequence, the apparent situation cited by the libertarian is not possible. Given that my projects cohere in light of my initial project, others may be quite reliable in predicting my behavior. This is what Sartre grants to the determinist. Perfect prediction is nevertheless impossible for both epistemic and structural reasons. According to Sartre, the initial project is free in so far as I could have chosen otherwise. The initial project thus gives coherence to all my projects and in some way determines my projects. This allows for some predictability. But there are two factors that limit this determination: first, this determination by the initial project is just self-determination and thus nothing which could be played off against Sartrean freedom. Second, although nobody can act against her initial choice, everyone can, in principle, alter her initial choice to inaugurate a radical transformation of selfhood. Thus, at each time, I could have acted otherwise because I could have just now altered my initial project. But this, according to Sartre, comes at quite an existential expense, and thus it is rare.

Conclusion: Sartre’s primary interest Sartre has a theoretical interest in acting insofar as he takes actions to be aspirations to achieve the value. In that respect, there cannot be acting for Sartre if there is no intentionalityPRO. 144

Jean-Paul Sartre

This is why Sartre assumes that acting is “on principle intentional,” and why he regards the “adequacy of the result to the intention” (Sartre 1992: 560/477, my translation) as a criterion for action: “The careless smoker who has through negligence caused the explosion of a powder magazine has not acted” (Sartre 1992: 560/477). When Sartre assumes that the criterion for speaking of an action is nothing but the coincidence of project and result, he is talking from the perspective of a hermeneutics of subjectivity interested only in those “results” that are indeed a revelatory manifestation of an intention PRO. But even from that rather restricted stance, this approach is tricky. Since Sartre’s agent does not have to know her intention but may often learn it from what she does, Sartre assumes that my actions in the world ultimately teach me what I really intended PRO. However, I can only make a difference between actions of mine and slips of my body by checking whether the latter correspond to intention PRO of mine. This is only one of many problems14 that arise even on a sympathetic reading. On a less sympathetic reading, one may regard Sartre’s ontology as far too speculative or simply absurd. But the aim here was not to offer a defense or a critique but an exposition of what take on action results from Sartre’s ontological framework. That this perspective does not even touch upon many issues currently discussed stresses the impact of this framework. What Sartre is after is not a theory of action but an (ontological) account of the non-reflective life of a consciousness which is determined to act and which determines itself to act.

Notes 1 I usually follow Hazel Barnes’ translation unless marked by “modified”; the page numbers refer to Sartre (1992). The page numbers of the French version are included after a slash. They follow Sartre (2010). 2 As Descartes (1983) suggests in his Principia Philosophiae, First Part, § 51: “Per substantia, nihil aliud intelligere possumus quam rem quae ita existit, ut nulla alia re indigeat ad existendum.” “/By ‘substance’, we can understand nothing other than a thing which exists in such a way that it needs no other thing in order to exist.” 3 Sartre’s notion of the absolute is also linked to the transparency of consciousness: “it is because of this identity of appearance and existence within it that it can be considered as the absolute” (Sartre 1992: 17/23). 4 See Aristotle’s Metaphysics IX: 3, 1046b29–1047b3. 5 See Sartre (1992: 28/32) on the merely regional status of the principle of identity. 6 Sartre habitually but wrongly refers to “the second Cartesian proof ” of the existence of God (see e.g. Sartre 1992: 139/126). What Sartre hereby refers to, however, is only a moment within the dialectics of weighing arguments against Descartes’ argument from ideas; and the argument from ideas is Descartes’ first proof of the existence of God, not his second one. 7 Cf. Sartre (1992: 140/127): “Nothing can hold out against this self-evident truth: consciousness can exist only as engaged in this being [of the value].” 8 For Sartre, desire “bears witness to the existence of lack in human reality” (Sartre 1992: 137/124). 9 Without being haunted by its existential possibility of being quenched, “thirst” would be no more than an affective tenor (cf. Sartre 1992: 562/479). 10 For a fine delineation of the background of Gestalt-psychology, see Morris (2010: esp. 147). 11 Cf. Sartre (1992: 275/237): “[T]he order of instruments [and demands] in the world is the image of my possibilities projected in the in-itself, i.e. the image of what I am. But I can never decipher this worldly image. I adapt myself to it in and through action.” 12 Barnes translates Sartre’s motif as “cause” and mobile as “motive” which leads to confusions when thinking bilingually. The terms are thus left untranslated. 13 The mode varies with the profound intention. See Sartre (1992: 581f/495f ). 14 A further crucial problem is linked to Sartre’s ambivalent sketch of the value. On the one hand, the engagement into value is what bears the entire framework of an engaged subjectivity which Sartre is so keen on. On the other hand, Sartre regards this engagement as a source of an essential

145

Simone Neuber alienation and takes it to be nothing but a source of bad faith and a manifestation of unhappy consciousness: “The being of human reality is suffering because it rises in being as perpetually haunted by a totality which it is without being able to be it, precisely because it could not attain the in-itself without losing itself as for-itself. Human reality therefore is by nature an unhappy consciousness with no possibility of surpassing its unhappy state” (Sartre 1992: 140/126f ). See also Sartre’s concluding reflections: “Every human reality is a passion in that it projects losing itself so as to found being and by the same stroke to constitute the In-itself which escapes contingency by being its own foundation, the ens causa sui, which religions call God. Thus the passion of man is the revers of that of Christ, for man loses himself as man in order that God may be born. But the idea of God is contradictory and we lose ourselves in vain. Man is a useless passion” (Sartre 1992: 784/662). However, here – after his discovery of an existential psychoanalysis – Sartre is at least open to the possibility of “put[ting] an end to the reign of this value” (Sartre 1992: 798/675). However, it is revealing that one of the last questions that Sartre raises in his book is the question: “And can one live this new aspect of being?” (Sartre 1992: 798/767) – i.e. that which is not under the reign of the value.

References Atwell, J. E. (1972) “Sartre’s Conception of Action and His Utilization of Wesensschau,” Man and World 5, 143–177. Descartes, R. (1983) Principles of Philosophy, trans., with explanatory notes by V.R. Miller and R. P. Miller, Dordrecht: Reidel. Descartes, R. (2007) Meditations on First Philosophy, trans. and ed. by J. Cottingham, Cambridge: Cambridge University Press. Gadamer, H.G. (1988) “Das Sein und das Nichts,” in T. König (ed.), Sartre. Ein Kongress, Hamburg: Rowohlt, pp. 37–54. Hegel, G.W.F. (1977) Phenomenology of Spirit, tr. A.V. Miller, Oxford: Oxford University Press. Heidegger, M. (1976), “Was ist Metaphysik”, in Heidegger, M., Wegmarken (GA 9), ed. by F.-W. v. Hermann, Frankfurt: Klostermann, pp. 103–122. Husserl, E. (1980) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, translated by E. Klein and W.E. Pohl, The Hague, Boston, MA, London: Nijhoff. Morris, K. (2010) “Sartre on Consciousness. Remembering the Gestalt Psychology Context,” in Mirvish, A. and van den Hoven, A. (eds.), New Perspectives on Sartre, Newcastle upon Tyne: Cambridge Scholars Publishing, pp. 142–160. Neuber, S. (2017) “Sartre über präreflexives Bewusstsein. Eine kritische Relektüre,” in G. Hindrichs (ed.), Konzepte Band 3: Bewusstsein, Frankfurt am Main: Klostermann, pp. 47–100. Sartre, J.-P. (1992), Being and Nothingness. A Phenomenological Essay on Ontology, New York et al.: Washington Square Press. Sartre, J.-P. (2010), L’être et le néant. Essai d’ontologie phénoménologique, Paris: Gallimard.

Further reading J.E. Atwell, “Sartre and Action Theory,” in H. Silverman and F. Elliston (eds.), Jean Paul Sartre (Pittsburgh: Duquesne University Press, 1980, pp. 63–81), outlines the interrelation of intention and action in Sartre and draws many lines to (then) contemporary discussions. K. J. Morris, “Sartre,” in T. O’Connor and C. Sandis (eds.), A Companion to the Philosophy of Action (Malden, MA et al.: Wiley-Blackwell, 2010, pp. 570–577), discusses Sartre as a proponent for the claim that reasons for actions cannot be causes for actions which is not challenged by standard refutations.

146

12 EMMANUEL LEVINAS Freedom and agency Michael L. Morgan

A natural place to begin thinking about how agency figures into Levinas’s “ethical metaphysics” and his conception of human existence is to consider what freedom means for him as a capacity of the will. In Totality and Infinity, in a chapter of Section I entitled “Truth and Justice,” Levinas uses the expressions “freedom called into question” and “the investiture of freedom, or critique” (Levinas 1969: 82–90). Agency or the capacity to act would be truncated and defective if it were not somehow conditioned by the ethical. That is, the freedom of agency, to be what it should be, must be “called into question” or submitted to “critique,” where the engine of that critique is the ethical or what Levinas calls the subject’s encounter with the face of the other. If the word “freedom” calls to mind “true agency” or action under the control of the actor and if agency is a capacity of individuals as embodied persons, then agency is only free and only true if it is susceptible to ethical orientation. Levinas makes this point on several occasions, and it is central to his conception of how we live our lives and ought to live our lives. In this chapter, I want to explore what these formulations mean. Before I turn to the two sections from Totality and Infinity that I cited earlier, I want to begin by looking at the final section of Levinas’s paper “Substitution,” which is entitled “Before Freedom.”1 The paper, originally given as a lecture in 1967, was published first the next year and then, in 1974, it became the core of the central chapter of Otherwise than Being. This later book is a deep exploration of the subjectivity associated with the face-toface or what we might call “ethical subjectivity.” This is not the subjectivity associated with everyday, embodied agency; it is the subjectivity that is correlated with the self ’s or ego’s encounter with the other person at a primordial and transcendental-like level. The last section of the chapter asks how the outcome of this deep exploration, the conception of this Ur-subjectivity as substitution, responsibility, persecution, and being a hostage, is situated in our full-blown, embodied personhood alongside the individual’s freedom to act. It attempts to answer the question: how does the before-freedom of ethical subjectivity influence the with-freedom of everyday life? Or, alternatively, how is ethics a critique of free agency?

Subjectivity before freedom There are trends in philosophy of the past half century that are anti-humanist or antisubjectivist in a strong sense. Some reject the primacy of the individual or subject; some 147

Michael L. Morgan

reject the primacy of humanity or human freedom; some reject the very existence of subjectivity. Levinas has been figured among those who oppose the centrality of the free individual subject insofar as he claims regularly that the other and the impact of the other on the self or subject are prior to the freedom and activity of the subject. There is good reason to accept such a view of Levinas, that he is an advocate of a “humanism of the other.” If so, then any examination of agency in Levinas’s account must appreciate how derivative a notion it is and how the subjectivity of the individual person as an agent must be grounded in some important sense in an Ur-subjectivity that is not an agent at all but something else. Levinas’s account of ethical subjectivity is that it involves a kind of emptying-out of the self and what he here calls “submission.” He calls it a “Yes older than naïve spontaneity” (93); that is, it is a kind of passive receptivity prior to any agency or act of acceptance. The naïve spontaneity to which Levinas here refers is collectively the “acts of free will” that “idealists” take to be the origin of the world, the creation of the world. This idealist claim is the presumption concerning which God challenges Job, Levinas says, when God asks if Job was present at the world’s creation. God’s question asks, is idealism the whole story? God implies it is not, and Levinas reiterates the challenge. As he goes on, “the distinction between the free and the nonfree is not therefore ultimate” (93). “Prior to the Ego taking a decision, the outside of being, where the Ego arises or is accused, is necessary” (93). And this beyond-being or accusation “occurs not through freedom but through an unlimited susceptibility, anarchical and without assumption . . .. The birth of the Ego in a gnawing remorse, which is precisely a withdrawing into oneself; this is the absolute recurrence of substitution” (93). With this notion of the subject “withdrawing into oneself ” Levinas may here be alluding to a doctrine central to the Kabbalah of Isaac Luria, according to which the divine Ein-Sof or Infinite, prior to the creation of the world, withdraws in order to permit the creation of what is other than God to come into being. The Infinite must contract in order for the finite to occur. As Paul Franks has shown, this motif is found in Fichte and then Hegel, inherited from Friedrich Jacobi and Solomon Maimon, who draws on it from the kabbalistic readings of Spinoza familiar from the late 17th and early 19th centuries. In the idealists, this tzimtzum or contraction becomes the ground of the possibility of interpersonal interaction and recognition, the act whereby the I withdraws in order to make it possible to posit first the not-I and then the alter-I of the other person. It is to this stage of the inheritance that Levinas here may well be referring.2 Moreover, this Ur-self hood or Ur-subjectivity – what I have called “ethical subjectivity” – “is not originally an auto-affection presupposing the Ego but is precisely an affection by the Other, . . . a traumatism of responsibility and not causality” (93–94). That is, ethical subjectivity is wholly a passivity, but it is, as Levinas often says, a passivity beyond all passivity – i.e. prior to the distinction between activity and passivity. And these claims about what condition or dimension of our self hood “precedes” or occurs prior to the everyday experiences of life, of freedom and agency, underscore how the domain of responsibility and the domain of freedom are distinguishable for Levinas: to be a ‘self ’ is to be responsible before having done anything. It is in this sense to substitute oneself for others. In no way does this represent servitude, for the distinction between master and slave already assumes a preestablished ego(94). Thus, “to say that subjectivity begins in the person, that the person begins in freedom, that freedom is the primary causality, is to blind oneself to the secret of the self and its relation

148

Emmanuel Levinas

to the past” (94), which is another way of calling attention to “the idea of responsibility preceding freedom.” Levinas does not here refer explicitly to predecessors such as Kant, Fichte, and Hegel, or to contemporaries such as Husserl, Heidegger, and Merleau-Ponty, but he could have. The Ur-subjectivity he is here seeking to excavate and characterize precedes agency and hence freedom, where it is the capacity of agency to be spontaneous, to initiate new causal chains, actions under the self ’s control. Thus, this Ur-subjectivity or ethical subjectivity is not a subject that is responsible or that substitutes or that is hostage; rather it is responsibility through and through, substitution itself, being a hostage. And in another metaphor that Levinas uses, it is not an obsessed agent or subject; it is obsession. Hence, in order to clarify Levinas’s conception of agency further, we need to consider these two dimensions of the living, embodied person, first the naturally existing agent, and second, the ethical subjectivity that orients and directs that naturally existing agent. Ultimately, we want to understand how Levinas understands everyday agency as a critique of freedom; how does ethics orient decisions and actions? In the spirit of Kant, Levinas calls freedom “spontaneity” and refers to it as a “form of causality.” Some of our conduct, as embodied subjects, embedded in the world of objects or what Levinas calls “elements,” involves living off of such objects as water, food, and even air – we do breathe the air, and are regularly enjoying this process of being nourished by what we drink, breathe, and eat. We also, in various ways, exert ourselves; labor or work often requires significant effort – to build or dig or plant or gather. Need and especially our corporeal needs, our materiality, drive us to work and labor.3 Levinas does not talk about what makes such labor or work activity or actions, but presumably it is that it is our conduct, arrived at by decisions frequently based on deliberation and reflection, and intentional, i.e. done for reasons that are transparent to us, at least to some degree and some of the time. To one degree or other, this work or labor is constrained and restricted; we can dig and cultivate only the soil available to us and plant only what we have in hand. But to a degree, it can be free or voluntary, insofar as the decision what to plant, where, and when is the result of our consideration and thought, and when we act on such a decision, the action is intentional, done for reasons, and to some degree up to us. Various phrases and expressions Levinas uses suggest that such a general view about agency is one that he takes for granted. Eventually, what is of special interest to him is not what agency is and how we can distinguish between actions and other forms of behavior on our part. Rather it is that beyond the instrumental and prudential considerations that apply in this broad domain of our lives as natural beings, there is something more within us that brings with it moral considerations that direct or orient or determine our capacity for free agency in very particular and very significant ways. On the one hand, Levinas does spend some time revising Heidegger’s account of what our existence as being-in-the-world comes to, in terms of our being situated in a world of objects and living off of them, enjoying them, and so forth. On the other hand, Levinas identifies how many of our moods, conditions, and feelings, from fatigue to insomnia, testify to the looming, foreboding, threatening “density” of being itself, to what he calls the il ya [there is]. Free activity, in a way, provides us with some control over our lives, given this background of threat and anxiety. But this control is not sufficient by itself to enable us to overcome this foreboding, indeed to “escape” from it, and to find some meaning and significance in our lives; for that we need to be attentive to the needs, concerns, and vulnerabilities that others, in their encounters with us, expose to us and in virtue of which they are dependent upon us. As Levinas puts it, what gives our lives meaning is not confronting and coping with the

149

Michael L. Morgan

prospect of our own death; rather it is confronting and coping with the possibility of the death of others – and indeed, beyond their death, even their concerns, interests, pain, and suffering. So freedom by itself does not enable us to live authentically and to “escape” bondage to being itself; only freedom oriented to our responsibility to and for others can do that. It is this possibility that comes to embodied human existence by means of our face-to-face encounters or relations to others.

An example: judicial agency But this dyadic, second-person relation or nexus, while it is present in every actual, everyday relation with others that occurs in our lives, no matter how near or far the other person is, no matter what relational context we live in, nonetheless never occurs alone. In Levinas’s terms, we are always present to myriad others, near and far, familiar and unfamiliar, and so forth. And this fact of the plurality of third parties extends our network of relatedness to all others, so that our everyday conduct and actions always occur within a network of such relations. At the same time that each of us engages with every other as a face or at the same time that each of us is responsible to and for each and every other person, in every way, all the time, we are also related to others, individually and in groups or collectivities, in mediated ways, in terms of principles or relations described in particular ways or programs and policies or institutions and roles. Sometimes, such principles, descriptions, and so forth are determined by prudential or instrumental considerations, but sometimes they are ramifications of the “original” responsibility-for-others that arise with the face-to-face. When this is the case and when our agency is thus guided or oriented by our sense of “justice,” our relations with others take on a moral character. That is, when this is the case, then our freedom is subjected to critique, or our freedom is “called into question.”4 That is, we can choose to act morally to one degree or other, to do what is right or wrong. Levinas gives an example of this condition in a comment that he makes on several occasions regarding a Talmudic text, where the question is raised about two Biblical verses that seem to contradict one another. One, Deuteronomy 10:17, says that God does not lift up his countenance, while the other, Numbers 6:26, says that he does. How can God both show favor and not show favor? The passage occurs in the Babylonian Talmud, Tractate Rosh HaShana, pages 17b–18a. In it, the Talmud records a parable intended to illuminate the apparent contradiction. The parable describes two men who arrange a loan, with the borrower swearing on the king’s life to repay it. When the loan comes due, however, the borrower cannot pay it and seeks to excuse himself before the king. The king accepts his excuse but orders the man to seek forgiveness from the lender. The standard rabbinic interpretation is that the parable is distinguishing between offenses against God and offenses against another person, the implication being that God can show His favor only in the former cases. But Levinas ignores this section of the Talmud and looks instead to the Talmud’s report that Rabbi Akiba had read the apparent conflict in a different way. To Akiba, “one text – presumably the Deuteronomy text – refers to God’s attitude before the final sentence, the other to His attitude after the final sentence.” Levinas takes Akiba to be referring to a trial or situation of judgment, to a court: with regard to the verdict, the principles of justice must be applied impartially, but after the verdict is handed down, when it comes to sentencing, the convicted party must be heard, and special considerations can enter into the court’s decision, such as a special plea for forgiveness and the submission of a very particular excuse for the failure to pay, as in the case of the borrower. As Levinas puts it: “Do not look at the face before the verdict. Once the verdict has been given, look at the face.”5 To him, then, the Talmudic passage expresses 150

Emmanuel Levinas

the idea that the state’s responsibility is to formulate laws and the judge’s responsibility is to adjudicate according to them, and this means the courts must apply them fairly and uniformly, and without any special attention to the particular claimant or defendant. But after a decision is rendered, and it is time for the sentencing or the application of the penalties, then the judge can take into consideration special features of the case and of the persons involved. This Levinas calls a “surplus of charity or of mercy.” Judgments in court have two stages: trial and sentencing. One requires impartial treatment according to the rules of evidence and interpretation of the law; the other, however, gives the judge discretion to take many features of the case and of the convicted party into consideration. Moreover, even though Levinas describes this as not attending to the face and then attending to it, this formulation is compressed. What he means is that justice in the domain of the third-party or of institutional life requires impartiality and the general application of principles and norms, which is itself what responsibility to and for others requires in such general and everyday settings. During sentencing, however, responding to and taking into consideration the utter particularity of the convicted party, the precise other with whom one is engaged, allows for very specific treatment, indeed for sensitivity and perhaps compassion. At such a stage, acknowledging and respecting the convicted party may involve mitigating the sentence or, one can imagine, making it especially severe or selecting one form of punishment over another. Levinas does not ask specifically about the agency of the judge in such cases. He does not ask how he might deliberate or inquire into the aspects of his decisions and actions, nor does he seek to clarify the sense in which his is first-person agency and the sense in which it is third-person agency or collective agency. But we can fill in some of the details for him and identify how the case of the judge can involve freedom and a critique of freedom.6 It is helpful to begin by remembering that the judge is an individual, embodied and embedded in the natural and social world, who is playing a particular role. As a physical being, the judge has his own interests in those objects around him that support his or her natural existence – the air he breathes, the water he drinks, the food he eats, and all that provides him with nourishment, enjoyment, and resources for living. He also must have a place to live, a dwelling that is home to him, where he keeps his clothes and belongings, where he sleeps and eats much of the time, and so forth. In short, as a physical being the judge lives his life in and from the world; many of his interests arise from his needs and enjoyments, what sustains him and gives him pleasure. Much of his experience is a matter of him receiving the sights, colors, sounds, and more from the objects around him but also reacting to things with pleasure, satisfaction, and such. As Levinas also points out, the judge will act, expending effort by working on things, laboring with them, and such, basically to facilitate the availability of food, clothing, and shelter, but beyond that to interact with things around him. In the course of this labor and activity, the judge becomes skilled or at least practiced in acting in various ways; he realizes tendencies, talents, and dispositions toward doing what is required in order for him to go on living and enjoying his life. Some of the judge’s labor is expended instrumentally, to do what is needed to achieve certain ends, but some is simply done for the sheer enjoyment of it. And at times his actions not only accomplish certain ends or further certain means, but they include or bring with them feelings and emotional reactions, and at times his actions originate with such feelings or emotions. The judge also interacts with other persons, talking to them, acting together with them, interacting both physically and in other ways. Levinas notes that the dwelling or home we establish for ourselves, extending our location in physical space to a place that is ours, is also the place where we live with others, in particular one or more others who care for it and care for us in it, welcoming and serving us in various ways. The judge has such a home, and he 151

Michael L. Morgan

may have a family. There is one or more others who live with him in it, who both serve him and show him what it means to serve others, who extend themselves for him and care for him and who are his introduction to what it means to do so for others. And of course, daily and regularly, the judge lives with others with whom he regularly interacts, from members of his family to neighbors and friends to casual acquaintances and then to strangers and even distant others. The judge’s everyday life, both personal and professional, is lived with a vast plurality of others, and his days are filled with interactions of all kinds, conversations, communications, and more. All this is the province of what Levinas calls “politics” or “life in the city, in the social world,” and it is a world filled with roles, norms, practices, institutions of all kinds, and so forth. In it, the judge, as all others, interacts at times accidentally and unintentionally, but also at other times intentionally. Often, he and others interact in terms of their roles and the expectations for conduct built into those roles; in addition, he and others regularly react to one another emotionally, with indifference or attentiveness, with pleasure or jealousy or joy or with worry or distress or resentment or anger. Now let us turn to the judge’s actions in the courtroom, with a plaintiff or someone charged of a crime standing before him, when it is what he does as a judge, given the expectations of his role and the legal-juridical system, that matter most. From Levinas’s point of view, one might think of the judge as confronted with two large-scale considerations, each of which is of course broken down, in terms of his conduct, into various smaller or narrower choice-situations. The one consideration is how he ought to conduct himself as a judge in the particular context of the specific issues being brought before him in court, what the law is in this particular context, how the case is presented to him by the attorneys or legal representatives, the kind of evidence available, and so forth. In other words, to simplify, the judge must respond to the requirements of the case as a legal case, and this means he must treat the claimant or defendant fairly and impartially, based on what his role and the law require. But another consideration that the judge must confront is his relationship to the witnesses, say, to the attorneys or legal representatives, and ultimately to the defender or claimant. Thinking back to the parable in the Talmud, if the judge is the court’s representative, to whom the debtor is appealing when he is unable to pay the loan, then the second context is the judge’s particular encounter with this particular debtor, who has defaulted on the loan he owes, has offered various excuses for his failure, and has indeed asked for some kind of forgiveness or leniency, given his circumstances. Levinas’s reading of the two Biblical passages about “God’s lifting up His countenance” or showing special favor is that the context is a trial or hearing and that one passage, where there is no special favoritism, refers to that stage in a judgment where the judge must act impartially and according to rules of justice formulated to guarantee as much fairness as possible. These rules concern the norms and practices that make life among many people, in a complex and large society, possible and that spell out what justice requires for such a plurality of parties living together in a community or social group. Here the judge, in court, is not being asked to evaluate the legal procedures or the laws about defaulting on a loan. He is being asked to apply the law fairly and to judge guilt or innocence based on the evidence, the testimony, and “facts” about the loan and the debtor’s defaulting on it. But Levinas distinguishes this venue from the later one, once the judgment has been made, and the judge must hand down a sentence. Here, Levinas claims, the judge may have general guidelines, but he should, Levinas says, also consider carefully and attentively the special features of the person standing before him, the particularities of what he sees and hears and what the defendant says and how he acts. That is, Levinas’s point is that the judge not only can and must respond to all that is distinctive about the debtor; he 152

Emmanuel Levinas

must consider how best to acknowledge him, respect him, show him concern and act on it, and so forth. In a sense, Levinas suggests, both the judge’s response to the debtor who has defaulted as a general category and his response to the debtor in all his particularity should both be guided by – i.e. moved by and determined by – the judge’s sense of responsibility to and for the debtor – as well as to and for other particular persons in the court and all others in the community or group. At one level, he owes fairness and impartiality; at the other, he owes sensitivity and compassion. As a rational and free agent, the judge should be reflective in both dimensions, as it were. His relation to the debtor is both general and particular. But he should not forget that the second and not the first is deeper, more primary, and “orienting.”

Freedom’s shame Of special interest to Levinas, as I mentioned earlier, is the fact that ethics is a critique of freedom, and given what we have just said about how the judge conducts him- or herself in the courtroom, we can take this to mean that the judge’s deliberations, decisions, and judgments ought to be oriented by the judge’s responsibility to and for the debtor and many others, who play a role in the court’s proceedings. But how is the judge’s freedom itself not only subject to ethical critique but also itself not possible without ethics, i.e. without the primordial relation between the self and the face of the other, without subjectivity beingfor-the-other, as Levinas puts it? This is a deeper question about freedom, where freedom means the subject’s capacity to make decisions and act on them. How is such freedom, such taking control of one’s own will and acting on it, somehow deeply indebted to the subject’s infinite responsibility to and for each particular other?7 Levinas does in fact help us to see this dependency of freedom on ethics through his phenomenological or quasi-phenomenological examination of how the face of the other, at the most primordial level, is experienced by the subject or, to put this otherwise, how subjectivity is realized in its primordial relatedness to transcendence, to the other.8 The key to clarifying how Levinas’s account contributes to our understanding how freedom is possible is the recognition that the more Levinas considers the subject’s attention to its own desires, interests, and needs, the more he realizes that the Ur-subjectivity of the subject or agent is wholly evacuated in its relation to the other’s dependency upon it. This project is the central task of Otherwise than Being. The reason this point is central for Levinas is that this detachment from the subject’s own needs and interests is, in a way, facilitated by the claim of the other but also a prerequisite for the subject’s reflective assessment of those needs and interests. And even prior to any “reflective assessment,” this detachment from one’s own needs and interests, what Levinas calls “disinterestedness” (désintéressement), is a detachment that enables a kind of identification with what matters most in one’s deciding to act and in one’s acting. Free choice can only occur when such “stepping back” or “detachment” or reflective assessment is possible. In short, it is the other’s claim upon the subject that makes it possible, among many things, for the subject to choose and to act freely. Let me explain. In the wake of Harry Frankfurt’s famous paper on free will, many others have taken up the thought that freedom requires a detachment from our first-order desires, needs, and interests, and a reflective assessment of these first-order desires. One can find such a view in Charles Taylor, for example, and in a neo-Kantian constitutivist such as Christine Korsgaard.9 But little attention if any is paid to how such a reflective standpoint is taken up by a subject and why. Is it a brute expression of the self ’s capacity to take control of its desires and conduct, of its autonomy? Or can more be said about it? Is it simply a higher order 153

Michael L. Morgan

expression of the very freedom that such an account seeks to elucidate and understand? It is here, with questions such as these, that Levinas can be seen to offer a proposal. Rather than take our self-identification with our desires, to be primitive or basic, Levinas seeks to show phenomenologically how there is a sense in which our most basic self hood, our Ur-subjectivity, is both immersed in its desires, needs, and so forth and yet also in a sense detached from them, or, somehow “evacuated of them.” This condition is what Levinas calls the subject’s “distinterestedness” (désintéressement). The term refers to the subject’s condition of being “dis-engaged” from its own interests in favor of standing in for the other as responsible for the other’s interests, needs, and so forth. The nexus of being-for-the-other or responsibility-for-the-other evacuates the subject of its own interests, desires, and so forth; the subject becomes infinite responsibility for everything about the other and hence “substitution” for the other and “obsession” for the other. This is what Levinas means when he says that the ego or subject is a “me” before it is an “I.” In short, one might well read Levinas’s mature understanding of Ur-subjectivity as responsibility-for-the-other and radical disinterestedness as making possible the kind of detachment and reflective distance required for freedom and free will. To be sure, Levinas’s account of such detachment and disinterestedness, a kind of giving-oneself-over-to the interests, needs, and concerns of the other person, does not claim that this is already the reflective distance in which free will operates. These forms of self-detachment or self-distancing are not exactly the same. But my suggestion is that Levinas’s point accounts for how any distancing of the self from itself takes place at all, and hence it accounts for how the mode of self-distancing that many associate with free will can occur. His central insight is that it is something that must be facilitated or brought about, as it were, by something outside the subject and not by the subject acting on its own. The possibility of being free is something that must be “given” to the subject and not something that it can create for itself. But if so, there is a problem that arises for Levinas. Does he not say that freedom must be subjected to critique? And does this not suggest that freedom is already in place, prior to such critique and indeed that there can be freedom prior to and independently of that critique? In other words, does not Levinas’s own account require that freedom be a feature of the human natural condition and hence that freedom does not require the interpersonal and the ethical in order to be what it is? On the one hand, it looks like freedom for Levinas is given prior to or at least independently of the face-to-face, while on the other hand it looks like freedom is only possible insofar as the subject is related to the other person as responsibility and substitution. Levinas’s answer to this objection is that there is no fact of freedom that precedes freedom’s shame of itself, that shame comes first, as it were, which means that there is no freedom without such shame, i.e. without the subject’s being exposed to the other and responsible to and for it. Or, to put the point otherwise, the philosophical analysis that parses our everyday human existence into the subject’s embeddedness in the world of objects or its natural existence, on the one hand, and the subject’s being-for-the-other, on the other hand, may expose a certain priority of the ethical over the ontological, but in everyday experience, these dimensions constitute one whole. The dimensions engage with one another and constitute the whole of ordinary experience and ordinary life. The freedom of the agent both depends upon the subject’s unlimited responsibility to and for the other and is directed by it toward enacting that responsibility in the best ways, in ways that serve the good, that serve the other, acknowledge and accept the other, care for her and seek to ameliorate her pain and suffering.10

154

Emmanuel Levinas

Agency or how we act as individual subjects, then, for Levinas, requires both our natural capacity to initiate actions, to be in control of what we do, and our “ethical capacity” as infinitely responsible to and for others. As he puts it, we are free, but our freedom must be subject to critique. Moreover, as we have seen, this critique can be more or less effective in one’s daily life. And this fact suggests that we might look at the agent’s conduct from two points of view. That is, one might consider Levinas’s understanding of agency in two contexts or from two perspectives. One context calls attention to those moments of our everyday lives when we are most responsive to our sense of obligation toward others, when the claim of the other person is most compelling, and when we act without giving a second-thought to reaching out to the other person and extending ourselves in her behalf – or even without giving even a “first-thought” about it. That is, in Levinas’s view of the human condition, the influence of the face-to-face and our sense of infinite responsibility to each and every other person, the normative force of it, occurs with a kind of immediacy or directness, prior to any reflection or deliberation. It may be that in general such a claim or such responsibility registers in our thinking about what we ought to do, but Levinas seems to think that also such a claim can register “directly” in our acting in the other’s behalf. Elsewhere I have discussed an example of such “immediate” responding to the needs and vulnerability of the other person (see Morgan 2007). In the 1980s, on several occasions, Levinas refers to a passage from Vasily Grossman’s novel Life and Fate, where Grossman describes how a grieving mother, having lost a daughter during the battle of Stalingrad, and now watching and jeering as a captured German soldier is seen dragging the body of a young female from the ruins of a building, at first appears to lunge toward the soldier with a stone, as if to beat his head in, but suddenly drops the stone, pulls a crust of bread from her pocket, and offers it to him. Levinas takes this to be an act of “holiness” or of “saintly” generosity, of responsibility of a pure and compelling kind. Here all thought and even all typical emotions – anger, deep resentment, and more – are set aside in favor of a response to the soldier’s needs and vulnerability that is spontaneous and direct and even, dare one say it, transcendent. Here is pure love or “charity.” Unlike the actions of the judge and the discretion in sentencing, with his or her deliberations and judgments, the woman’s act here is prior to all reflection and as purely ethical as one might imagine is possible in our everyday lives. On the one hand, then, ethics can under normal circumstances orient and direct our will to judge and to act; on the other hand, there are times, more rare, when ethics can seemingly preempt our will altogether and move us to act directly and surprisingly in behalf of a vulnerable other person. My sense is that for Levinas such pre-reflective, immediately responsive conduct in behalf of another is a kind of action; it is to be directly “compelled” by an impulse to respond with “love and charity,” as he might put it. It is compulsion by the Good (see Murdoch 1971). The second context, to return to the case of the judge in court, is more typical of our daily circumstances and daily lives. Here, our will operates as a stage in our thinking about our situation, our considering what we might do and why, and finally deciding what to do. In such normal circumstances, we are, as Levinas sees it, both utterly particular in our relations and general in how we are related to the other person. The other person, to use his language, presents us with a face and also, in a sense, is a third-party; the other person is given to us in the second person and also in the third person, as it were. In these sorts of cases, which are most common, our responsibility for the other person in all her particularity enters into a deliberation that considers one’s role-bound responsibilities, the other’s situation in all its complexity, the availability of other sources of assistance or help, the demands

155

Michael L. Morgan

of considerations such as fairness and deservedness, the degree of need, other demands on time and resources, and so forth. Levinas does not explore the process of deliberation that leads to a decision and then to action. He does not ask how the will works, when principles or guidelines come to mind, and when we feel bound by rules or directed by considerations such as the availability of alternative resources. What Levinas does do is to make clear that our daily lives do operate with the guidance and direction of concepts, principles, reasoning, and such, and that all of these tools and devices take on significance for our social relations insofar as they are ways in which our responsibility to and for others is organized and generalized beyond the case of one particular other person so that we can deal with a vast plurality of others and the complex network of responsibility relations which tie people together. As individuals, this is the most ordinary setting in which we think, respond, and act. Nowhere in his writings, to my knowledge, does Levinas set out an analytical account of what makes us agents, what an action is, and how we might interpret what we are as voluntary agents. But his primary interest in exploring and exposing how interpersonal responsibility plays roles in our lives and how it provides our natural existence with ethical significance does provide resources for saying something about these matters. That is what I have tried to do in this chapter.

Notes 1 See Levinas (1996: 93–95). References in the text. 2 It is an allusion to the kabbalistic notion of divine withdrawal or contraction, in Hebrew tzimtzum, already found in Totality and Infinity, p. 104. 3 See, for example, Levinas (1969: 116–117). 4 See Levinas (1969: 84–85): “Conscience welcomes the Other. It is the revelation of a resistance to my powers that does not counter them as a greater force, but calls in question the naïve right of my powers, my glorious spontaneity as a living being. Morality begins when freedom, instead of being justified by itself, feels itself to be arbitrary and violent.” And, in a passage that surely alludes to Sartre, Levinas says: “Existence is not in reality condemned to freedom, but is invested as freedom . . .. To philosophize is to trace freedom back to what lies before it, to disclose the investiture that liberates freedom from the arbitrary.” 5 See Levinas (1990: 194). See also, Levinas (2001: 69). 6 I discuss Levinas’s reading of this Talmudic passage in Morgan (2016: 6–10). But there I do not ask about the judge’s agency in particular. 7 For an extended treatment of this subject, see Houser (2019). 8 Levinas often refers to his method as “phenomenological,” and his descriptions of the self ’s experiences of both the “face of the other” and of its “Ur-subjectivity” as evacuated, “disinterested,” and so forth are very much in a phenomenological style. But, strictly speaking, “being for the other” and subjectivity as “substitution” are beyond description and expressivity. The relation of the self ’s subjectivity at this level and the other person is a kind of transcendental structure and hence experience of it is beyond “ordinary, everyday” terms. I call Levinas’s many attempts to depict such experience “quasi-phenomenological.” For discussion, see Morgan (2007: 300–335). 9 See Frankfurt (1971), Taylor (1989), Korsgaard (1996). 10 In this way, for Levinas, self hood or agency depends upon the nexus he calls “encountering the face of the other” or, in later works, simply “being-for-the-other” and “responsibility-for-the-other.”

References Frankfurt, H. (1971) “Freedom of the Will and the Concept of a Person,” The Journal of Philosophy 68/1, 5–20. Houser, K. (2019) “Levinas and the Second Personal Structure of Free Will,” in M. Fagenblat and M. Erdur (eds.), Levinas and Analytic Philosophy: Second-Person Normativity and the Moral Life, London, New York: Routledge.

156

Emmanuel Levinas Korsgaard, C. (1996) The Sources of Normativity, Cambridge: Cambridge University Press. Levinas, E. (1969) Totality and Infinity, orig. French, 1961, Duquesne: Duquesne University Press. Levinas, E. (1990) “In the Name of the Other,” in E. Levinas (2001) Is It Righteous to Be? Interviews with Emmanuel Lévinas, Stanford: Stanford University Press. Levinas, E. (1996) “Substitution,” in A. T. Peperzak, S. Critchley, and R. Bernasconi (eds.), Emmanuel Levinas: Basic Philosophical Writings, Indiana: Indiana University Press, pp. 19–96. Levinas, E. (2001) “Interview with François Poirié,” in E. Levinas (2001), Is It Righteous to Be? Interviews with Emmanuel Lévinas, pp. 23–83. Morgan, M. L. (2007) Discovering Levinas, Cambridge: Cambridge University Press. Morgan, M. L. (2016) Levinas’s Ethical Politics, Bloomington: Indiana University Press. Murdoch, I. (1971) The Sovereignty of Good and Other Essays, orig. 1970, New York: Schocken Books. Taylor, C. (1989) Sources of the Self, Cambridge, MA: Harvard University Press.

157

13 HANNAH ARENDT Plural agency, political power, and spontaneity Marieke Borren

Engaging with the work of the philosopher who is best known for her work on the “active life” (vita activa), Hannah Arendt, this entry deals with a particular type of agency that is rarely accounted for in phenomenology: political agency. For Arendt human beings are actors only in particular instances, that is: if one follows the standard account of agency as entailing the capacity for intentional and goal-directed action, which I will call the model of agency as sovereignty. In this model, someone is considered an actor if she knows what she is doing and is more or less in control of the outcomes of her deeds, so that those outcomes can indeed be attributed to the enactment of her intentions (even if granted that she may of course accidentally fail in achieving her goals). According to Arendt, it is only when making things—as homo faber, “working” man, in her own words— that people exercise agency in the sense of sovereignty. As embodied beings—animal laborans, the “laboring” animal—people are not sovereign, as the unqualified needs of their bodies and its affects are given and fixed. As pathos, in the double meaning as passion and suffering, needs and affects befall people. Moreover, the moment people enter the public realm as citizens (zoon politikon) and engage in “action” sensu stricto, i.e. the very moment that for Arendt apparently is key to a properly human life, they are not self-sufficient either.1 The political actor “is never merely a ‘doer’ but always and at the same time a sufferer. To do and to suffer are like opposite sides of the same coin,” she writes (HC: 190). Whereas the non-sovereignty of animal laborans precludes political agency, this is entirely different when people act as citizens. Even if they are not sovereign, Arendt does not deny their agency. Nor does she consider political agency a delusion, as if humans were puppets, robots, or vehicles of supra-individual historical or natural forces. She takes great pains to argue that as soon and as long as people act in public space, they are neither determined, nor themselves determining sovereign actors. In fact, the temptation to replace non-sovereign action by work and to conceive of political agency in the image of making something—“the traditional substitution of making for acting” (HC: 220)—is as widespread in political theory and praxis as it is dangerous. This substitution is indeed the principal metaphysical prejudice about the active life that Arendt sets out to deconstruct, because it closes down the key condition of political action: plurality. Several manifestations of this prejudice will be touched upon below, because they lead to

158

Hannah Arendt

misunderstandings about political agency, particularly the collapse of freedom and sovereignty on the one hand, and of power and violence on the other hand. Drawing on her two major philosophical works, The Human Condition (1958) and The Life of the Mind (1971/1978), I will demonstrate in this contribution that Arendt conceives of political agency as plural agency and untangle why, in her view, agency precludes sovereignty in the public domain and why non-sovereignty is even the condition of citizens’ power and freedom when they act in the presence of each other. The particular type of agency that comes with political action in the Arendtian sense is expounded by juxtaposing it to that of the other human activities. The stakes are further clarified by contrasting it with the predominant aspiration in the philosophical tradition to resolve the frustrating non-sovereignty of action through a number of, what Arendt calls, metaphysical prejudices and fallacies, on the one hand, and to a, mostly (neo-)Nietzschean undercurrent in philosophy that, reversely, puts forward the complete disavowal of the agent (whether or not political), a “doer behind the deed,” on the other hand. The hypothesis that this contribution seeks to defend is that acknowledging Arendt’s distinctively phenomenological approach of politics is crucial for understanding her account of the question of human agency, particularly political agency. This aspect of her thought renders it an original contribution, not just to the phenomenology of agency, but also to political theory. Arendt’s phenomenological perspective is based on an analysis of the lived experiences of the types of human activities which she calls “labor,” “work,” and “action.” The reception of her work has mainly taken place in political theory, probably as a result of the Anglo-American predominance in Arendt scholarship. However, being immersed in the emerging phenomenological movement in German academia in the first half of the 20th century at a formative age, her philosophical habitus is deeply shaped by phenomenological concerns and approaches. Since it does not fit into the phenomenological orthodoxy (i.e. Husserl) and because Arendt keeps her method largely implicit, it took some decades after her death for the phenomenological and hermeneutic inspiration to be appreciated.2 This inspiration goes a long way in explaining her non-theoretical approach of the question of human agency. She wished to examine human affairs without theoretical (metaphysical and ideological) prejudice, i.e. preconceived ideas, or as she put it in an interview, “to look at politics . . . with eyes unclouded by philosophy” (EU: 2). Her work is committed to understanding rather than explaining political phenomena and to be faithful to reality, that is: phenomenal reality, reality as it appears in the world and so is visible and common to all people. This commitment to reality also made Arendt aversive to any type of wishful thinking, including not just romantic or radical utopianism, but also normative political theory such as ideal theory and normative value theory. In her view, political thinking is not about designing alternative, i.e. better, more just, political orders, nor about justification or prescription.3 Therefore, I will start with a reconstruction of the features of Arendt’s phenomenology which additionally allows for an introduction of the notion that is key to understanding her account of political agency: plurality. Section “Agency across the human activities” proceeds with a discussion of the distribution of agency and sovereignty across the various human activities. Whereas section “The non-sovereignty of the political actor” highlights the non-sovereignty of political agency, contra the metaphysical prejudices and fallacies, section “The doer and the deed” demonstrates its agentic dimension, contra (neo-)Nietzschean post-metaphysical thought. The agentic dimension of Arendtian plural agency is finally (section “Political agency: freedom and power”) fleshed out by a discussion of the phenomena of power and freedom as the manifestations of respectively the interaction and the initiative dimension of political action.

159

Marieke Borren

Arendt’s phenomenological approach to active life As said, Arendt’s engagement with the question of human agency is informed by a typically phenomenological approach. Her œuvre shares several motifs and assumptions with that of other phenomenologists, in particular those working within its hermeneutic branch, while politicizing phenomenology in important respects. Like other phenomenologists, she engages in descriptive analyses of phenomena, that is, of things as they appear to human beings in lived, pre-reflective experience. Like Heidegger in particular, she holds that experience is constituted by an implicit, pre-reflective understanding of phenomena, which comes about through people’s familiarity with them in their practical dealings. In contrast to Heidegger, Arendt’s work brings out the political aspects of human existence—what she calls “the human affairs” to indicate its plural nature—focusing on phenomena such as public space, power, and freedom, in order to uncover the experiences that underlie them. Like other postHusserlian phenomenologists such as Heidegger and Merleau-Ponty, Arendt often takes a via negativa, or negative approach, to the phenomena and experiences she investigates. In order to elucidate daily and allegedly “normal” experience, they, each in their own way, attend to the non-everyday: limit cases, pathologies, or more mundane instances in which people’s ordinary routines break down. In Arendt’s case, the phenomena of revolution, but above all totalitarianism (terror and ideology) are the “unprecedented” experiences that set off her thinking. Arendt’s phenomenological method precludes the empiricist and disengaged perspective which is typical for the majority of social research (including political science), on the one hand, and the theoretical approach that is dominant in political theory and political philosophy on the other hand. Like any phenomenologist, she always takes a relational point of view with respect to the things she studies. The perceiver is not opposed to or separated from the perceived. Things and events are not seen in isolation, as entities or realities external to us. So-called “objects” are always things and events that show themselves to a perceiver and somehow make sense to them. Thus, instead of “objects,” phenomenologists speak of “phenomena” or “appearances”: i.e. that which appears to a perceiver. The perspective the perceiver takes upon things is therefore central to the phenomenologist’s attention. In addition to most other phenomenologists, Arendt’s emphasizes the plurality of perspectives that those who perceive—spectators—take upon the phenomenal world. She stresses that appearance implies the “presence of others” (HC: 50, 95, 188, 199), i.e. of a plurality of spectators, as a matter of course. Appearance is always an appearance to others (LOM I: 19). The human “sense of reality” (HC: 208) is guaranteed “by the presence of others, by its appearing to all” (HC: 199). I will call this the reality effect of plural appearance and action (HC: 50–51, 95, 199, 208; LOM I: 19). Thus, Arendt infuses the basic phenomenological notion of intentionality with multi-perspectivism.4 Like most phenomenologists, Arendt’s engagement with metaphysics is highly critical. The object of Arendt’s deconstructions is the Western metaphysical legacy in political thought, more particularly the prejudices about the active life (labor, work, action) and the fallacies about the life of the mind (thinking, willing, judging) it propelled. The metaphysical prejudice and fallacy most relevant to the issue of agency that Arendt discusses are, respectively, the substitution of making for acting and the “two world theory.” The latter, epitomized in Platonism, concerns the presupposition of two worlds, on the one hand the true and eternal world of ideas which is only accessible through thought—Being—and on the other hand the misleading, fluid, and contingent world of phenomena accessible through sense perception—appearance. These two worlds traditionally do not only constitute a 160

Hannah Arendt

dichotomy but also a hierarchy, as the first of them is supposed to be of a higher rank than the latter. On this view, that which does not appear to the senses is more significant and true than that which does appear. Instead, Arendt asserts the coincidence of appearing and being. The human world, for all intents and purposes, is the phenomenal, visible world, which human beings have in common to the extent that is principally open to different perspectives. The dualism of being and appearance is obsolete, since appearances are all people will ever encounter, both within and outside of the political domain; they are simply incapable of telling appearances from being (LOM I: 3–65; Taminiaux 1997: chs. 1–2). Arendt’s deconstructive method, the application of which will be demonstrated in sections “Agency across the human activities” and “The non-sovereignty of the political actor,” entails, first, a genealogical investigation of the way the history of political experiences and phenomena is condensed or sedimented in language, that is, either revealed or concealed. Theoretical reflection (in which the mind is oriented to itself as a principle) can never equal the insight that paying close attention to discourse (which is oriented to the world) generates (HC: 94). Like Heidegger, Arendt therefore typically starts her phenomenological investigations with an analysis of concepts, because these provide privileged access to underlying fundamental human experiences, “not because [they] reveal the phenomenon in any straightforward way, but because [they] carry the record of past perceptions, true or untrue, revelatory or distorting” (Young-Bruehl 1982: 405). The aim of such a genealogical project is to disclose the experiences that underlie these concepts in order to achieve a better understanding of the phenomena they refer to. Subsequently, Arendt offers a phenomenological description of their relevance to specific, often conflicting, experiences one has when engaging in active life or in mental activities. Most of the time, Arendt argues, philosophical concepts either express generalizations of particular phenomenal experiences of the political, or amalgamations or confusions of different phenomenal experiences, for instance work and action, sovereignty and freedom, or violence and power. These confusions inform harmful metaphysical prejudices about political life, for example that politics is fundamentally about rule, that political power rests on violence and that sovereignty is its key principle. Arendt’s deconstructions of metaphysical prejudices and fallacies feature two dimensions: critique and experiment. Arendt’s work consists in “essays” or “exercises in understanding” (BPF: 14). It has the critical dimension of destruction (“dismantling” (LOM I: 212) and the experimental dimension of storytelling. Through stories, Arendt aims to retrieve forgotten experiences or “lost treasures” (BPF: 4), not just, as is often thought, the political experiences of the ancient polis, but also the modern experiences of revolutions and the institution of civic councils (see section “Political agency: freedom and power” below). Critique is directed toward the past, the given order; experiment toward the future; the new and unexpected which defies what is given. Still, the critical and experimental moments of understanding are connected for Arendt. Critique without experiment results in cynicism (“debunking,” BPF: 14), whereas experiment without critique all too easily leads to the kind of utopianism that she seeks to avoid. After making manifest generalizations and amalgamations, she criticizes them, and subsequently describes the phenomenal distinctions to save the experiences covered up by them. This practice of discrimination— between the various activities within the vita activa, power and violence, freedom and sovereignty, the private and the public spheres—is an important, but no doubt the most controversial, feature of Arendt’s phenomenology. Understanding for Arendt essentially consists in discriminating: “This is so and not otherwise” (Vollrath 1979: 101). Arendt’s distinctions aim at phenomenological clarification. A careful analysis that is to bring out the specificity of distinct phenomena requires discrimination, not generalization. 161

Marieke Borren

For Arendt, distinguishing between experiences does not mean separating them. Distinctions always contain distinct but related dimensions, for example, natural and worldly aspects of one and the same issue. Because of her phenomenological commitment to descriptive analysis of lived experience rather than theory, Arendt is particularly sensitive to paradoxes, perplexities, and ambiguities inherent in the human condition. This phenomenological attentiveness to the paradoxes inherent in human activities is one of the original features of her approach to politics that generates many novel insights; yet this aspect of her thought is often poorly understood. It has caused Arendt’s work, mostly wrongly I believe, to be labeled as conceptually inconsistent. Arendtian paradoxes express apparently contradictory experiences, yet nonetheless provide fundamental and unexpected insights in the human condition. Real life is full of experiences that appear to contradict each other, but which are not merely apparent and hence false. Rather than misconceptions, paradoxes in lived experience express counter-intuitive insight (from the ancient Greek para-, counter-, and doxa, opinion). The paradoxical experiences are both different (distinct) and related, like two sides of a coin. The irredeemable tension between them “makes sense” on further consideration. Both experiences are “real,” i.e. meaningful, and so the ensuing tension cannot eventually be reconciled, explained away or resolved by sound (i.e. logical) thinking or by gaining further (scientific) knowledge. In short, Arendtian paradoxes follow not from theoretical reflection (in which case they would be symptoms of either conceptual inconsistency or theoretical fetishization), but from lived experience itself that her phenomenological descriptions bring to light.

The paradox of plurality Plurality, arguably the “core phenomenon” in Arendt’s thought (Loidolt 2018: 2), provides the most insightful example of Arendt’s attentiveness to the paradoxes of lived experience. Since plurality is decisive for the distribution of agency and sovereignty across the different human activities, a brief exposition of the paradox of plurality seems warranted. Plurality is the key condition of political life, Arendt holds. Without a plurality of people, there is no politics (HC: 7; PoP: 93). Political action, freedom, and power are meaningless unless they involve interaction with other people who are both equal to and different from me at the same time. So plurality constitutes a paradox, as it simultaneously involves difference and equality (HC: 175–176, 178; PoP: 93). For Arendt, difference refers to individual distinctness or uniqueness; it is what prevents people from being exchangeable. What makes people distinct is that they are situated beings with a unique biography, and with each of them taking a different perspective on the world. Plurality involves alterity, being other than something else, but otherness in itself is not yet plurality since, strictly speaking, copies, specimens, and replicas already possess this quality of alterity. Distinctness can only be enacted in and through speech and action “in the presence of others,” that is, in public space, or what Arendt calls the “space of appearances” (HC: 199). Human beings are not born as unique individuals, but only individuated through interaction with others. Neither is equality a natural given, such as an inalienable property that all human beings allegedly have, by virtue of being born (natural law doctrine) or because they are created equally in God’s sight (the biblical doctrine of equality). For Arendt, equality is a political notion, in that it is entirely artificial or conventional. Like difference, public space is needed for equality to be enacted, for it only comes about the moment people interact with others who are irredeemably different from me. Equality should not be confused with sameness. Sameness indicates the naturally given similarity of human beings. As animal laborans, human beings possess a shared biological 162

Hannah Arendt

constitution that accounts for a limited number of basic needs. Indeed, as animal laborans, “we are all the same,” to a large extent. Plurality means that as soon as people enter the public sphere, they obtain and grant equality by showing in words and deeds their particular perspective of the world vis-à-vis one another. Put differently: political equality pertains to people who are by nature unequal as a principle (HC: 205). The paradox of plurality implies that difference (distinctness) is not opposed to equality but that the two mutually presuppose each other. In fact, both equality and distinctness are opposed to sameness. Equality, indeed, presupposes that a person is equal to a different person. If human beings were all the same it would make no sense to pursue equality. In the same way, relevant differences can be identified only with respect to the norm of equality.

Agency across the human activities The purpose of the previous section has been to foster appreciation of Arendt’s phenomenological approach to political phenomena. This section discusses Arendt’s descriptions of labor, work, and action—in order to explore the implications of this approach for the questions of human agency and sovereignty. The if and how of human agency depend on the particular mode of human action that people engage in (labor, work, action), and the typical mentality (animal laborans, homo faber, zoon politikon) and accompanying orientation to their environment (the earth, the thing world, and the immaterial world of the human affairs) that they take up in the exercise of these activities. In their capacity as laboring, consuming, care-needing and -giving beings, human beings approach their environment as animal laborans. This environment is earthly nature. As specimens of the human animal species, homo sapiens, human beings are embedded in earthly nature and embodied. All human activities (labor, work action, and including the activities of the mind, thinking, willing, and judging) are earth-bound, but nowhere is the correspondence between humans and the earth stronger than in labor. Like every other organism, the physical constitution of humans is adapted to the (climatological, geological) conditions of life on earth and its natural resources. Arendt calls the earth “the very quintessence of the human condition” (HC: 2). On account of their embodiment, human beings are subject to what is given: earthly conditions and the needs and vulnerabilities of their bodies and its affects and passions: the human condition of “life itself ” (HC: 7). Labor encompasses the activities that serve the self-preservation of the human organism, both on the level of the species and the individual, e.g. typically food production and consumption, reproduction, housekeeping, and care. The products of labor are consumables. They literally merely feed back into the laboring process itself. As a consequence, labor is endless in the double meaning of the word: devoid of a purpose beyond itself and as a principle never finished. The natural needs of the human body are cyclical and continuous, and can never be satisfied once and for all. As soon as I have cleaned up and tidied the house, dirt and junk start to pile up again; eating and drinking only suspend my becoming hungry and thirsty again. From the perspective of life itself, bearing children merely enables my children to bear children in their turn (HC: 30–33, ch. III) Life itself is typically experienced as coercive—we don’t control it, it controls us—fixed and unchangeable. It is, in the words of Judith Butler, “unchosen” (Butler 2012, 2015).5 As embodied and therefore vulnerable, passionate, and affective beings, people are subject to the facticity of “sheer passive givenness” (HC: 208). Moreover, people’s inner life, with its affects and passions, is quite literally invisible. It does not appear to different others. In fact, the person’s inner life is even opaque to itself. 163

Marieke Borren

It does not appear to itself in introspection and needs the detour through the phenomenal world to gain solidity and a sense of reality. Due to their unworldliness, needs and affects escape their being shared with others. This is the solipsism of animal laborans. The noncommunicability of extreme pain is a case in point. As it is deprived of the “reality effect” of appearance to others, one’s inner life may have the same “weird”—unreal or surreal— quality as dreams or nightmares. As reports of experiences of solitary confinement and exile confirm, people need the world and others to redeem the “darkness of the human heart” (HC: 237, 244; LOM I: 35; BPF: 144): its contradictions, ambiguities, equivocalities, volatility, and unreliability. Labor is a non-individuating activity. Torturers know only too well that the body’s susceptibility to suffering, pain, and pleasure is more or less uniform, and cases of extreme destitution seem to show the same. Qua animal laborans, differences between people merely pertain to alterity, not to plurality, as people lack the individuality that distinguishes them from others and that allows for equality and being-together to come about in the first place (LOM I: 35, 72). The variety of unqualified human bodily needs and affects is significantly more limited than the diversity of humans’ outward appearances (of course these affects are transformed if people give shape to, for example, their sex drive, fear (LOM I: 35–36), or even hunger,6 i.e. make them visible in the world by articulating and sharing them with others). As the most “pathic” mode of being in the world, animal laborans is neither sovereign nor agentic. This is entirely different in the case of homo faber, that is, human beings’ experiential position and mentality whenever they engage with the world in the mode of work, the activity of making (producing, manufacturing) things (HC: 34–42, ch. IV). If ever people enact sovereign agency, it is in this mode. Based on a phenomenological analysis of the lived human experience of strength (including violence), Arendt calls homo faber “the lord and master of the whole earth” (HC: 139). Working implies having a clear purpose or end (a product) in mind, determining and calculating the means necessary to achieve it, drawing up a plan and subsequently executing it, following particular rules or procedures. The results of work are both predictable and reversible: what is produced can as a principle be undone, or otherwise artifacts simply decay if they are not maintained. Homo faber, the “builder of worlds and the producer of things” (HC: 160), brings into being the thing world or the artifact, including furniture, buildings, infrastructural works, institutions, culture, and technology. Together these man-made things constitute the human home on earth. Arendt calls the human condition that corresponds to work “worldliness” (HC: 7). It represents the elements of artificiality and utility in human existence. Humans take an instrumental or utilitarian orientation to the world in the mode of homo faber because working involves tailoring means to given ends. The reification and de-naturalization inherent in the activity of work provide a stable and durable foundation for human life to protect them against natural forces. Although this durability is relative—man-made things are obviously subject to decay—it is more stable than both nature and the immaterial world of the human affairs which are equally ephemeral as a matter of course. Thanks to its mastership and authorship, homo faber possesses sovereign agency. To the extent that they belong to homo faber, people know what they are doing, they are self-sufficient and in control as a principle, because they are able to predict and reverse the things they make. However, they do not have political agency, because work is not pluralistic. The activity Arendt calls action is the only truly political one among the human activities, because it is only in action in public space that plurality can flourish. As citizens, zoon politikon, human beings display an entirely different type of agency than they do when appearing 164

Hannah Arendt

as animal laborans or homo faber. Partially a sufferer, a pathic and hence passive creature like animal laborans, the citizen is non-sovereign; partially a doer or active being like homo faber, she is agentic, at least potentially. Unlike both of them she has the potential to develop political agency. Arendt’s genealogy and deconstruction of the concept of action in antiquity reveal two aspects of its lived experience which sets it apart from both labor and work: initiative (archein/ agere) on the one hand and enactment (prattein/gerere), its reception and continuation by others on the other hand (HC: 189).7 Arendt speaks of initiative because action always involves beginning, the start of something new. Action has no given end like work and always constitutes a surprise, an interruption in the course of events. This is the “miraculous” quality of action, although there is nothing super-natural about the unexpected happening. Second, action always implies interaction (as I call the dimension of enactment) as it takes place in the presence of and together with others—co-actors and spectators—by definition. Therefore, action requires a public space of appearances: a space to show oneself in deed and word to others who are different, to be seen and heard by them, while they mutually expose their uniqueness and achieve equality: the paradox of plurality. No matter if they agree or disagree, people bring about a shared world of meaning, and “sense of reality” when they start to act. To clarify the “reality effect” of plural action, it is helpful to contrast the activity and “world” of action with those of work. Whereas work is a world-building activity, that makes the thing world, action is a world-disclosing activity. Contrary to the tangibility of the thing world that people create in their capacity as homo faber, the world of the human affairs that action discloses is entirely symbolic or discursive. It is the public and common space of stories, opinions, judgments—in short: of shared plural meanings—that weaves a “web of relationships” between people (HC: 181). To explicate the practice of world disclosure, Arendt’s uses the metaphor of illumination, more particularly the lighting of a stage. The metaphor illustrates the basic phenomenological assumption that phenomena—in Arendt’s case political phenomena—always appear against a background of concealment (HC: 71; OR: 98). In addition, spotlights serve to lighten up a person or an act in a play, by providing focus. This example also shows that in the process of world disclosure, individuals disclose themselves as political actors as well, even if the “disclosure of the agent in speech and action” (HC: 175) is not the intended purpose of action and speech.8 Only by acting in the world do people individuate, according to Arendt (HC: 97; EU: 23). In sharp contrast to phenomenologists such as Heidegger, she emphasizes that sharing, interpreting, judging, and discussing one’s opinions and evaluations with others are indispensable for world disclosure. By “talking about” (HC: 183) things and events (which involves the perspectives of others as a principle), people make them meaningful, i.e. disclose their meaning.

The non-sovereignty of the political actor To expound on the notion of plural agency and to show what is, and is not, meant by the non-sovereignty of Arendt’s political actor, it will be contrasted, in this section, with dominant tendencies in the metaphysical tradition (driving home the discussion of the deconstructions of the metaphysical prejudices and fallacies in section “Arendt’s phenomenological approach to active life”) and with the disavowal of agency in neo-Nietzschean anti-metaphysical thought in the next section. Arendt takes issue with the traditional identification of political agency and freedom with sovereignty: “[S]overeignty, the ideal of uncompromising self-sufficiency and mastership, 165

Marieke Borren

is contradictory to the very condition of plurality. No man can be sovereign because not one man, but men inhabit the earth” (HC: 234; cf. BPF: 163). The non-sovereignty of the political actor is the flipside of the potential greatness of action. Acting presents people with frustration, risk, and frailty. The failure to achieve one’s aims is not accidental, as it would be in work. Why so? Action is not endless, like labor (which keeps people moving within a closed or vicious circle, the eternal return of the same), nor is it finite like work (in which the end product brings the working process itself to a closure, too), but open-ended. This indeterminacy allows for great deeds to happen, but it is also a constant source of its frustrations (HC: 183, 190, 233; EU: 23). Adventure and venture in one, action is uncontrollable. Every deed weaves a new strand into an already existing web of relations and meanings that precedes it. Action is boundless, because it almost always sets into motion a chain reaction, in which every reaction is a new action in itself (HC: 190). The consequences of a deed do not cease with the deed itself, they lag behind. It may augment and build upon itself like a snowball effect, for better or worse. Its consequences may also boomerang back. Unlike God (at least in monotheistic religions), human beings are incapable of controlling the if, what, and how of the outcomes of their deeds. Only when acting completely solitarily, in the vacuum of a laboratory, could one at least hypothetically control the results of one’s deeds. In real life, however, people will never find themselves in such a situation, because they always act in a plural world. One should note that Arendt does not consider non-sovereignty a weakness. The political actor’s non-sovereignty does not follow from her primary dependency on others for survival, like human beings indeed simply cannot do without the help and care of others in their capacity as animal laborans (HC: 234).9 Because of the human condition of plurality, action is, first, marked by unpredictability, because it is impossible to secure its consequences and meaning in advance. Citing Luke 23:34, “Forgive them, for they know not what they do,” Arendt points to the daily and common human experience that people cannot foresee the future, whereas the consequences of what people do or say will (or fail to) be felt exactly there (HC: 239; cf. HC: 233, 237, 240; EU: 23). People’s knowledge is deficient as a principle when they engage in action. It is impossible to know all the intentions and motives of the others involved, because they do not appear in the world. The latter is equally true of people’s own intentions and motives. Introspection cannot ascertain them, since they are, just like passions and affects, part of one’s inner life. They are unreliable and volatile as long as they do not appear in the world and become visible to all (OR: 98).10 In addition, others have motives of their own as well, and since action as a principle takes place among (many) others, these may thwart one’s plans, redirect, enhance, or attenuate them. As a consequence, unexpected things may happen which transcend the original intention (BPF: 84; HC: 184, 201; EU: 320). Plurality, second, ensures that people cannot undo what they have done or said, hence the feature of irreversibility. For example, one cannot choose to live in a world in which European powers had never launched imperial conquests and settler colonialism.11 The irreversibility of action also holds for one’s own past deeds, even if most of the time the impact of them is felt only by a handful of people. A final frustration of action is “the anonymity of its author” (HC: 220). Because of the plural nature of action, no one ever accomplishes something entirely on his or her own, not even the likes of Gandhi, Rosa Parks, or Mandela nor dictators such as Hitler, Stalin, or Mao. To the extent that action is brought about by heroes and heroines, the latter are like protagonists in ancient Greek tragedy: both doers and sufferers of the consequences of their deeds rather than solitary and self-sufficient strong wo/men. 166

Hannah Arendt

Traditionally, philosophers have overwhelmingly responded to the frustrations inherent in action with either hostility or despair. If the meaning of events and deeds can never unequivocally be attributed to the intentions of individual actors, and if acting humans hardly ever achieve their goals, as history proves time and again, then history is a meaningless and sad affair, the argument goes. To save meaning, philosophers such as Hegel, Marx, Smith, and, in some respect, Kant (i.e. in his philosophy of history) resort to the two world doctrine which Arendt considers the foundational metaphysical fallacy. They explain away human agency entirely by transferring it to the anonymous collective agency of forces that transcend the human affairs such as History (be it of the Geist or of class struggles), Nature (Kant’s “ruse of nature”), or the Market (Smith’s “invisible hand”). These collective intentionalities are assumed to rule behind the scenes (i.e. the visible world of the human affairs) and through or behind the backs of the individuals who are seen as merely executing the “rule by nobody” (HC: 44–45; cf. LOM II: 154–155, 179, 180; EU: 430–431; LOM I: 95–96; BPF: 82). More importantly, in political praxis the frustrations of action very often, and understandably, rouse the temptation to replace political action by work and to conceive of political agency in the image of the sovereign agency of homo faber: the metaphysical prejudice of the “substitution of making for acting.” For work, precisely unlike action, does have a clearly recognizable actor in control and its course is predictable and reversible. Totalitarian rule is the epitome of this prejudice. Like the carpenter who processes wood to make a table, totalitarian regimes treat human beings as the material to be processed and transformed. The identification of politics and rule is not the prerogative of totalitarian regimes, though. It is reflected in the technocratic or neoliberal discourse in liberal democracies that reduces politics to governance, management, or administration. Another case in point is the surge of “strongman politics”12 and worship of authoritarian populist leaders, the fiction of the one strong wo/man (HC: 188–189) that the world currently witnesses. Whatever the specific constellation, the substitution of making for acting destroys plurality as the key condition of politics.

The doer and the deed Does the notion of political agency still have any credibility if one accepts (and even embraces) the non-sovereignty of action, as has been advanced in the previous section? To many post- or anti-metaphysical philosophers, most notably Nietzsche and his followers, it does not. Because of her emphatic deconstructions of metaphysical prejudices and fallacies in favor of the appearing world and actor, Arendt’s account of action has often been put on a par with Nietzsche’s. In his Genealogy of Morality, Nietzsche famously denounces the presupposition that action needs a foundation. In his view, language, i.e. grammar, seduces people into believing that there is an actor behind acting: “‘The doer’ is merely a fiction added to the deed—the deed is everything” (Nietzsche 1969: 45). Especially in the 1990s, neoNietzscheanism abounded within radical democratic theory. Judith Butler made Nietzsche’s anti-metaphysical or -foundational thought fruitful for her feminist and queer concept of performativity and her radical critique of sovereign agency (Butler 1990: 25). Within Arendt scholarship, the agonistic theorists, Dana Villa and Bonnie Honig, read Arendt’s account of action through the lens of Nietzsche’s aestheticism (Villa 1996; Honig 1993).13 In this section I will argue against this reading. To make this argument is not to deny the significant parallels between Arendt’s and Nietzsche’s thought. Indeed, like Nietzsche, Arendt believes that the actor does not exist independent from the act, at least in the domain of political action. Action, she writes, is “sheer 167

Marieke Borren

actuality” (HC: 207, 208), i.e. phenomenality. In The Human Condition, she argues that the doer appears in the doing. In acting in public space and talking about a worldly issue, the actor discloses “who” she is. This who refers to the actor’s distinctness, her unique and incomparable life-story, as opposed to “what” she is, the sum of her given, unchangeable, and objectifiable features, including markers of collective identity such as gender and ethnicity. Arendt compares the political actor to the protagonist of a story or an actor performing in a play: someone discloses herself in acting on stage. The who which appears is an effect of interaction and appearance in deeds and words rather than a project, the reception of which the actor can manipulate, control, and master completely (HC: 175–181; LOM I: 37).14 The argument of the disclosure of the person in action that Arendt makes in The Human Condition is consistent with the non-expressivism of appearance she defends later on in The Life of the Mind. Here she argues that the appearance of the actor does not imply the expression—literally: “pressing out”—of “‘something inside’, an inner disposition or quality of the person” (i.e. what she used to call the “what”) or “an idea, a thought, an emotion” (or a motive or intention, I would add). Appearance in the Arendtian sense is self-referential, because “it ‘expresses’ nothing but itself ” (LOM I: 30; cf. idem 29; BPF: 144). The dualism between the actor’s invisible inner life and its outward appearance, together with the assumption that the former is more real and causes the latter, is but another version of the metaphysical two world theory that needs to be deconstructed in Arendt’s view. These parallels go a long way to explain the neo-Nietzschean “performative” and aestheticist reading of Arendt’s notion of agency.15 However, such a reading sits uneasily with the equally consistent emphasis Arendt puts in The Human Condition on the “actualization of the sheer passive givenness of [men’s] being” (HC: 208). She describes who one is as “the living essence of the person as it shows itself in the flux of action and speech” (HC: 181), and a “latent self ” that manifests itself by acting (HC: 175; cf. idem 208, n.41). In The Life of the Mind, she even more explicitly argues that appearance is something which humans take upon themselves actively. Appearance in a politically meaningful sense hence transcends merely or passively appearing; it takes a deliberate effort to make one’s appearance in speech or action on stage. Reversely, interaction in public always implies not merely “self-display,” but “self-presentation” in deeds and words (LOM I: 34). Decisive for the disclosure of the who is the active, conscious, and deliberate choice of how one wishes to appear, “what to show and what to hide” (LOM I: 34), and of what one thinks is “fit to be seen and what is not” (LOM I: 36). These arguments are not symptoms of residual metaphysical foundationalism or essentialism, or of a relapse into the model of political agency as sovereignty, but they follow from Arendt’s political-phenomenological take on the actor. Even if Arendt considers Nietzsche one of her strongest allies in deconstructing the two world fallacy, as Villa and Honig rightly point out, the latter ignore the phenomenological rather than the aestheticist background from which Arendt’s foregrounding of appearance arises. For Arendt, performance pertains, not to performativity in Butler’s sense, but to the disclosure of the world and the person, and to the enactment or interaction dimension of action. Most of all, it is because of the human condition of plurality that “action almost never achieves its purpose” (HC: 184). Far removed from the solipsism of the Nietzschean master, the aristocratic “exceptional man” and his celebration of the individual will to power, she argues that although the doer chooses to make her appearance in deeds and words, she does not determine the meaning of these deeds, because that is in the eye of others, i.e. its audience of co-actors and spectators (HC: 10–11, 178–188, 193, 206, 211). This is the paradox of suffering and doing.

168

Hannah Arendt

Arendt believes that people do have motives and intentions; these are not merely delusions created by language, history, nature, or the market. The disclosure of the person in action does not imply that its appearance is a false belief, as Nietzsche thinks, or a facade, caused by false consciousness, as Marx holds (HC: 183). Motives may incite people to take initiative, one of the indispensable elements of action. However, the relation between motives and deeds is contingent because one never acts alone, in other words: because of the human condition of plurality.16 It is the world instead of a self-sufficient actor that solicits action: When I make [the] decision [to appear to others], I am not merely reacting to whatever qualities may be given me; I am making an act of deliberate choice among the various potentialities of conduct which the world has presented me. (LOM I: 37)17

Political agency: freedom and power To flesh out the agentic dimension of Arendtian plural agency, this section will expound two main features of action: its typical mode of freedom—spontaneity—and its mode of power—citizen empowerment. The former corresponds to the aspect of initiative, the latter to interaction. Action is not necessary like labor, as it does not serve self-preservation, nor is it useful like work, because it has no exterior purpose, i.e. a product. Instead, the meaning of action is freedom. Arendt rejects the identification of political freedom and sovereignty and argues that the two are in fact mutually exclusive: “If men wish to be free, it is precisely sovereignty they must renounce,” she writes, because “in the realm of human affairs sovereignty and tyranny are the same” (BPF: 165).18 Instead, spontaneity—or natality—is the type of freedom people exercise when they act within the space of appearances. Whereas the principle of freedom as sovereignty implies controlling the outcomes of one’s deeds—the agency of homo faber—the principle of freedom as spontaneity refers to the initiative dimension of action: the human capacity to begin, to initiate something that did not exist before and which cannot be deduced from precedents or a preconceived ideology. Because of its spontaneity, action is both fundamentally contingent and inherently open and creative.19 Contingency and freedom are therefore closely related. The uncontrollability and contingency of action rule out complete causal determination. On the one hand, to exercise agency does not mean to cause the outcomes of one’s deeds and words. On the other hand, action is not causally determined, either by “the system,” one’s past or by one’s genes, not even by one’s motives or goals. As fundamentally spontaneous, action resembles improvisation. It is independent, i.e. not bound by rules, protocols, procedures, or models such as the work of homo faber. Arendt learned about the fundamental spontaneity of human action through a via negativa, namely her analysis, in The Origins of Totalitarianism (1951), of life in the concentration camps of the Third Reich. She writes that it is exactly human spontaneity that totalitarian regimes try to eliminate, by pairing ideology and terror, including horrific medical experiments. The concentration camps “serve as the laboratories in which the fundamental belief of totalitarianism that everything is possible is being verified” (OT: 437), including the experiment of turning human beings into “conditioned and behaving animals” (HC: 45). Under non-totalitarian conditions, however, real life never takes place in a laboratory but is always lived among others. Human beings can never be completely reduced to interchangeable “bundles of reactions” (OT: 438), i.e. organisms which completely obey a stimulus response

169

Marieke Borren

model, like puppets or robots (OT: 438, 455–457) whence “spontaneity as such, with its incalculability, is the greatest of all obstacles to total domination over man” (OT: 456).20 The ineradicability of human spontaneity and indeterminacy eventually brought to Arendt’s attention the human condition of natality as the principle of plural freedom and the initiatory dimension of plural action (OT: 478–479). The eventual failure of the experiment in the camps to reduce human beings to “mere life,” i.e. animal laborans, informs her descriptions, not just of action but of labor and work as well. For instance, it taught her to appreciate the artificiality of the thing world that homo faber creates. The reduction of freedom to sovereignty is a version of the substitution of making for acting. Holding onto this deeply unrealistic view of freedom, the failure to keep control in the public realm may drive people to despair and cause them to withdraw from the public world altogether (as for instance in Stoicism), which comes at the expense of the human sense of reality (HC: 234–235). There is a close link between Arendt’s description of freedom as spontaneity on the one hand and power on the other hand. Both are opposed to sovereignty. Political power is what emerges whenever people mobilize and organize around a particular worldly issue, in short: when they engage in non-violent “action in concert” (HC: 199–200; CR: 95, 98, 143). Arendt’s description of this phenomenon closely resembles “empowerment,” as this concept captures the positive and active, i.e. interactive, dimensions of the power of citizens. Arendt unravels the lived experience underlying power by, first, genealogically investigating the ancient Greek and Latin concepts of dynamis, potentia, and energeia, demonstrating that each of them refers to “enactment” (HC: 200, 206; cf. VA: 252, 261–262) and, second, by studying modern European and American history. She describes how throughout modern history, usually in the slipstream of revolutionary upheaval, voluntary associations of citizens, sometimes federally connected into council republics, surface spontaneously every now and then, to vanish, at least for that moment, after some time.21 For Arendt, power consists in a potential—a “potentiality in being together” (HC: 201)—that may be actualized whenever people start to act together. Like action, power only exists in its enactment. Genuine political action is almost always short-lived and small-scale; yet the possibility of making a new beginning is never exhausted. This potential is always there to be re-enacted by citizens at any time in history and in any place. Political power is not a property of an individual or a group, like sovereign power, as it is dependent upon interaction between different and equal citizens, i.e. “upon the unreliable and only temporary agreements of many wills and intentions” (HC: 201). The power of citizens does not rest on a homogeneous collective, such as “the people,” in sharp contrast to Rousseau’s conception of “indivisible” sovereign popular power. Arendt is of course aware that in daily speech, “power” is often used to refer to the authority of an individual or a small group of individuals, for example a head of state, the executive, or any other type of leader. However, saying that someone is “in power” means that she is empowered by people to act on their behalf. In other words, the power of a single person is derivative of the pluralistic power of action in concert. The experiences underlying power are completely different from those underlying violence—strength and force (HC: 200ff ). Both in political theory and praxis, power and violence are often identified, another example of the substitution of making for acting.22 In political theory, the model of sovereign power serves as the basis of justifications of the necessary use of violence in politics. For modern “realist” political theorists from Hobbes to Weber, state power ultimately rests on the legitimate use of violence. Against this view, Arendt argues that the derivative power of the state erodes when it no longer rests on the 170

Hannah Arendt

plural power of citizens. State violence crops up whenever the plural power of acting citizens crumbles. As a consequence, the use of violence is a symptom of impotence, not the condition of power. Very much like realist theorists, many radical political theorists, such as Rousseau, Marx, Sartre, and Fanon, attribute an instrumental role to violence in politics, as the means to create a new—free, more just—society, sometimes even a new man. When this ideological view of power and violence spills over into revolutionary praxis, it usually leads to glorifications of violence (CR: 105–198). Arendt’s works offer a wealth of case studies of “spontaneous rebellions” (HC: 216 n.52), citizen councils and associations, civil disobedience, and civil rights movements that illustrate the phenomena of freedom and power in action. The Hungarian revolution (OT), the US civil rights movement of the 1950s and 1960s, the student protest movement, and the antiwar movements of the 1960s and 1970s (CR; EU: 22) are but some cases she describes in detail. Similar examples which occurred after Arendt’s death include the eastern European dissident movements and “velvet” revolutions of “1989,” and the protest movements of “2011” (Occupy and the “Arab Spring”). What these examples have in common is that teach of them were non-violent movements that were launched spontaneously and mostly to their own surprise by citizens gathering in public space. These movements mobilized people with very different backgrounds and political views, often without even a formal affiliation to a political organization and did not have formal leaders, nor a shared ideology.

Conclusion Arendt repeatedly observes that the political actor is as much a sufferer as a doer (HC: 184, 190, 233–234). She rejects the model of political agency as sovereignty for its disregard of the plural and worldly quality of action. Collapsing the political agency of citizens and the sovereignty of homo faber leads to a destruction of plurality. She is equally far from asserting that the political actor is powerless, or fictitious, on account of her awareness of humans’ spontaneity, the ineradicable capacity to make a beginning, and the inexhaustible power potential inherent in action in concert. Arguing that political power is generated by irreducibly different individuals who act in concert, Arendt brings forward an account, not of collective, but of plural agency. Also, political agency cannot be reduced to the model of agency as resistance. The power of citizens is not limited to challenging given states of affairs (as in many contemporary agonistic democratic theorists who draw on Arendt’s account of action), but also includes the capacity to set into motion new states. Arendt’s account of political agency contains important lessons for both phenomenologists and political theorists. Arguing that appearance always implies appearance to— different yet equal—others, she brings the condition of plurality to phenomenological notions of agency. On the other hand, thanks to her phenomenological approach she provides a non-theoretical perspective to debates on agency in political theory. This approach explains her sensitivity to the paradoxes inherent in the human affairs. These paradoxes point not to Arendt’s conceptual inconsistency or to a theoretical obsession, but to lived experience itself. For example, doing and suffering are different but inseparable experiential dimensions of political agency, corresponding to the aspects of initiative and interaction. This approach also leads her to reject normative theory and to embrace, what I call, a particular phenomenological kind of “realism.” It helped her to demonstrate that the model of political agency as sovereignty is as “unrealistic” and illusory as is the complete disavowal of agency. Arendt’s realism does not imply a commitment à la Hobbes and Weber 171

Marieke Borren

to Realpolitik, nor does it follow from cynicism, or from resignation to what happens to be the case. Instead, it implies “the unpremeditated, attentive facing up to, and resisting of, reality—whatever it may be” (OT: viii). This disposition is consistent with Arendt’s phenomenological habitus of taking seriously the res, the matter or issue, that one examines— the “things themselves”—and of being faithful to reality as it appears in the world and so is visible and common to everyone.

Notes 1 Throughout this chapter, the word “citizen” is used to refer to humans in their capacity as “political animals” (zoon politikon), rather than to formal membership in a state. In this usage, even an undocumented immigrant may be a “citizen,” namely as soon as she starts to act politically in public space. 2 However, recently, the first systematic treatment of Arendt’s phenomenology was published, the excellent book by Loidolt (2018). Among the earlier exceptions is the work of Vollrath (1977, 1979) and Taminiaux (1996, 1997, 2000). Also see: Birmingham (2006), Borren (2013), Hinchman and Hinchman (1984, 1991), Mensch (2009), Ricœur (1983, 1992), Topolski (2015), and Vasterling (2011a, 2011b). 3 On Arendt’s commitment to a “proto-normative” “phenomenological form of practical reason,” that is attuned to the “ethical demands” implicit in the human activities themselves, see Loidolt (2018: ch. 6). 4 See Arendt’s discussion of Husserl’s conceptualization of intentionality (LOM I: 45–46). 5 From her earliest to her most recent work, Butler has challenged the model of sovereign agency. Whereas in her early work on performativity her arguments were inspired by Nietzsche’s disavowal of a ‘doer behind the deed’ (see section “The doer and the deed”), her recent arguments against sovereign agency draw on Levinas’s work: human beings are not sovereign, because of their shared condition of vulnerability and precarity and because of their “constitutive dependency.” This perspective informs her reading of Arendt’s work on action (2012, 2015). Even if Butler shares with Arendt the assumption that the body and life itself are given and passive, she criticizes Arendt for excluding the suffering and vulnerable body from political, on account of her supposedly rigid distinction between the private and the public spheres. This criticism partly rests on a misunderstanding of the phenomenological background of Arendtian distinctions (see section “Arendt’s phenomenological approach to active life”), partly on fundamental disagreements (see section “The doer and the deed” below). 6 See Arendt’s example of the labor movement in HC: §30. 7 In HC, Arendt translates prattein/agere as ‘achievement’ (189). In her own German translation of HC, she uses the verb vollziehen (substantive: Vollzug) (VA: 235), which translates as enactment, performance, or actuality. Cf. Loidolt (2018: 87, 201–202). 8 The disclosure of the actor is closely related to the appearance of the “who,” as will be argued in section “The doer and the deed.” 9 Arendt attributes the physical dependency argument against human sovereignty to the metaphysical tradition “since Plato” (HC: 234). However, note that Judith Butler’s argument against sovereignty rests on exactly the same assumption, i.e. constitutive dependency (Butler 2015). 10 As a consequence, intentions and motives do not constitute “who” we are according to Arendt, because the who appears in the world, as will be argued in the next section. 11 The fact that, obviously, the impact of imperialism is felt with different intensities across the continents and has various meanings for different individuals in no way contradicts Arendt’s account of world disclosure, because of its multi-perspectivism. 12 I borrow this phrase from Barack Obama’s Nelson Mandela Annual Lecture, delivered on July 17, 2018, in Johannesburg. 13 Recently, Villa has reconsidered Arendt’s “Nietzscheanism” (Villa 2008). 14 On the disclosure of the person and storytelling, see Ricœur (1983). 15 Arendt most explicitly draws a parallel between action and performance in HC and BPF. 16 For a similar argument, see Zerilli (2005: 11–13, 17).

172

Hannah Arendt

173

Marieke Borren Butler, J. (2015) Notes towards a Performative Theory of Assembly, Cambridge, MA: Harvard University Press. Hinchman, L. and Hinchman, S. (1984) “In Heidegger’s Shadow. Hannah Arendt’s Phenomenological Humanism,” Review of Politics 46/2, 183–211. Hinchman, L. and Hinchman, S. (1991) “Existentialism Politicized. Arendt’s Debt to Jaspers,” Review of Politics 53/36, 435–468. Honig, B. (1993) Political Theory and the Displacement of Politics, Ithaca, NY: Cornell University Press. Loidolt, S. (2018) Phenomenology of Plurality. Hannah Arendt on Political Intersubjectivity, London: Routledge. Mensch, J. (2009) Embodiments. From the Body to the Body Politic, Evanston, IL: Northwestern University Press. Nietzsche, F. (1969 [1887]) On the Genealogy of Morals, transl. W. Kaufmann, New York: Vintage. Ricœur, P. (1983) “Action, Story and History. On Re-reading The Human Condition,” Salmagundi 60, 60–72. Ricœur, P. (1992) Oneself as Another, Chicago, IL: University of Chicago Press. Taminiaux, J. (1996) “Bios Politikos and Bios Theoretikos in the Phenomenology of Hannah Arendt,” International Journal of Philosophical Studies 4/2, 215–32. Taminiaux, J. (1997) The Thracian Maid and the Professional Thinker. Arendt and Heidegger, Albany, NY: SUNY Press. Taminiaux, J. (2000) “Athens and Rome,” in D. Villa (ed.), The Cambridge Companion to Hannah Arendt, Cambridge: Cambridge University Press, pp. 165–177. Topolski, A. (2015) Arendt, Levinas, and a Politics of Relationality, London: Rowman & Littlefield. Vasterling, V. (2011a) “Hannah Arendt,” in S. Luft and S. Overgaard (eds.), Routledge Companion to Phenomenology, London: Routledge, pp. 82–91. Vasterling, V. (2011b) “Contingency, Newness, Freedom. Arendt’s Recovery of the Temporal Condition of Politics,” in C. Schues et al. (eds.), Time in Feminist Phenomenology, Bloomington: Indiana University Press, pp. 135–48. Villa, D. (1996), Arendt and Heidegger: The Fate of the Political, Princeton, NJ: Princeton University Press. Villa, D. (2008), “How ‘Nietzschean’ was Arendt?,” in H. W. Siemens and V. Roodt (eds.), Nietzsche, Power and Politics. Rethinking Nietzsche’s Legacy for Political Thought, Berlin: De Gruyter, pp. 395–408. Vollrath, E. (1977) “Arendt and the Method of Political Thinking,” Social Research 44/1, 160–182. Vollrath, E. (1979) “Politik und Metaphysik. Zum politischen Denken Hannah Arendts,” and “Hannah Arendt über Meinung und Urteilskraft,” in A. Reif (ed.), Hannah Arendt: Materialien zu ihrem Werk, Wien: Europaverlag, pp. 19–57and 85–107. Young-Bruehl, E. (1982) Hannah Arendt. For Love of the World, New Haven, CT: Yale University Press. Zerilli, L. (2005) Feminism and the Abyss of Freedom, Chicago, IL: University of Chicago Press.

174

14 MERLEAU-PONTY AND AGENCY Thomas Baldwin

“You are your act” – Merleau-Ponty quotes from Saint-Exupéry at the very end of Phenomenology of Perception (Merleau-Ponty 2012: 4831), where the passage occurs as if it is the conclusion of the book. Saint-Exupéry is describing a situation in which there is great danger, and in other references to it, Merleau-Ponty writes that “It can even happen that when I am in danger, . . . my body completely merges with action” (Merleau-Ponty 2012: 86), and that in these extreme situations one becomes aware of “the presence of self to self ” (Merleau-Ponty 2012: 426). So they reveal something important: “You are your act”. The emphasis here on action contrasts with the emphasis on perception that is characteristic of Phenomenology of Perception. We shall see how Merleau-Ponty seeks to present a new conception of human life which avoids the picture of us as “an activity tied to a passivity, a machine surmounted by a will” (Merleau-Ponty 2012: 452). Before coming to this new conception, however, it is important to understand his reasons for rejecting the picture of us.

Beyond dualism In simple terms, this picture is that suggested by the traditional dualism of the mind and body. For Merleau-Ponty this dualism is one between a consciousness whose essence is “being for itself ” – so that “in consciousness, appearance and reality are one” (Merleau-Ponty 2012: 310) – and a body whose essence is “being in itself ” (so that its identity is determined by its intrinsic physical structure). Merleau-Ponty’s critique of this dualism is set out most clearly in a long note in Phenomenology of Perception in which he discusses the analysis of apraxia advanced by the German neurologist Hugo Liepmann. Liepmann’s diagnosis of apraxia, which is an inability to carry out bodily movements, despite having the physical ability to perform them, was that because of brain abnormalities or damage the subject lacks the power to convert their mental representations of bodily movement into appropriate movements. Merleau-Ponty comments that the fact that Liepmann’s analysis involves mental representations shows that he thinks of acts of will as a kind of “motor consciousness” which represents desired bodily movements and normally gives rise to these movements by means of connections between these representations and the body; and, he argues, this position gives rise to the problem of determining “through which magical operation the representation of a movement gives rise in the body to precisely this 175

Thomas Baldwin

very movement” (Merleau-Ponty 2012: 525). The solution Merleau-Ponty then proposes involves a radical rethinking of the nature of consciousness and of the body: We will not render apraxia comprehensible nor make sense of Liepmann’s observations unless the movement to be accomplished can be anticipated, but without being so through a representation. This is possible only if consciousness is not defined as the explicit positing of its objects, but rather more generally as a reference to an object that is practical as much as theoretical. That is, if consciousness is defined as being in the world (être au monde), and if the body is defined in turn not as one object among others, but as the vehicle of being in the world. (Merleau-Ponty 2012: 525) Liepmann might respond that the mental representations he is concerned with are fundamentally neural states, so that there need be nothing “magical” about the connection between a mental representation and physical movements guided by it. Merleau-Ponty would challenge the assumption that this is possible, and the issues at stake here will come up later. But we should note now the way in which Merleau-Ponty connects his conception of our practical “reference to an object” with two of the central themes of his treatment of agency – first, that, as agents, we are beings whose being is being in the world; and, second, that our body is not just “one object among others” but our embodiment, the vehicle of our actions. Merleau-Ponty sets out his alternative conception of action in the following passage in which he starts by alluding critically to the conception of kinesthetic sensations employed by some psychologists: What they expressed, admittedly poorly, by “kinesthetic sensation” was the originality of movements that I execute with my body: my movements anticipate directly their final position, my intention only sketches out a trajectory in order to meet up with a goal that is already given in its location, and there is something like a seed of movement that only grows later through its objective trajectory. I move external objects with the help of my own body, which takes hold of them in one place to take them to another. But I move my body directly (directement), I do not find it at one objective point in space in order to lead it to another, I have no need of looking for it because it is always with me. I have no need of directing it toward the goal of the movement, it touches (touche) it from the beginning and throws itself toward it. In movement, the relations between my decision and my body are magical. (Merleau-Ponty 2012: 96–97) Some points here are straightforward, such as the fact that we move our bodies “directly”, without the help of another body. But what is more significant is that the account of the practical “reference to an object” is elucidated. The suggestion seems to be that once an intended goal for action has been identified (presumably by perception), there is no need for a further representation of the limb movements (of the kind hypothesized by Liepmann) required to achieve this goal. Instead “from the beginning” our movement reaches out toward that goal; there is a fundamental teleology to bodily movement (and thus action) – “a seed of movement” which “grows later through its objective trajectory”. What is then a little odd is the way in which, in the final sentence, Merleau-Ponty describes this position as one involving a “magical” relationship between one’s decision and one’s body, since in the discussion

176

Maurice Merleau-Ponty

of Liepmann he uses the term “magical” to criticize Liepmann’s position. But I take it that Merleau-Ponty is not here turning against his own account; instead he is just acknowledging that the inherent teleology of bodily action can seem to be “magical”, a way of “touching” the goal one is reaching toward before one has reached it. Merleau-Ponty says a good deal more about bodily movement in the context of his extended exploration of human embodiment in Part One (“The Body”) of Phenomenology of Perception. Central to this account of the body are the two themes encountered earlier – that, as agents, we are beings whose being is being in the world, and that our body is the vehicle of our actions. One aspect of this embodiment concerns the role of our sense organs in providing us with our perceptions; another concerns the role of our limbs and other parts of the body as the “living envelope of our actions” (Merleau-Ponty 1963: 188), and these roles are combined in the functional conception of the “phenomenal body” as “a system of motor powers or perceptual powers” (Merleau-Ponty 2012: 155). Merleau-Ponty contrasts this functional conception with the conception of the body as an object of some kind, as the “objective body”, and the relationship between these two conceptions is a complicated matter, to which we must now turn.

The ambiguous body Merleau-Ponty ends the Introduction to part I of Phenomenology of Perception with this enigmatic comment: And since the genesis of the objective body is but a moment in the constitution of the object, the body, by withdrawing from the objective world, will carry with it the intentional threads that unite it to its surroundings and that, in the end, will reveal to us the perceiving subject as well as the perceived world. (Merleau-Ponty 2012: 74) The challenge here is to understand how it can make sense to write of “the genesis of the objective body”, which presumably belongs to the objective world, and then, in the same sentence, to write also of the body’s “withdrawing from the objective world”. The solution must be that we are to recognize that there are two conceptions of the body, the objective body, which does belong to the objective world, and the phenomenal body, which so far from belonging to the objective world turns out to be the perceiving subject of that world. But quite what this solution involves, and how far Merleau-Ponty is willing to endorse a conception of the objective body at all requires further discussion.2 A good place to start is with “the ideal of objective thought” – The ideal of objective thought – the system of experience as a bundle of physico-mathematical correlations – is grounded upon my perception of the world as an individual in harmony with itself; and when science attempts to integrate my body into the relations of the objective world, it does so because it attempts, in its own way, to express the suturing of my phenomenal body onto the primordial world. (Merleau-Ponty 2012: 366) There is a complex line of thought here. Merleau-Ponty compares the ideal of objective thought with the way in which our ordinary perception of the world as “an individual in

177

Thomas Baldwin

harmony with itself ” has a manner of “suturing”, fitting, the phenomenal body onto “the primordial world” – which is the world of “primordial” experience that is “on this side of all traditions” (Merleau-Ponty 2012: note 7, 530). It is then this ordinary experience of ourselves as belonging within the world which the ideal of objective thought attempts to replicate by treating our experience as “a bundle of physico-mathematical correlations”. One can think of this ideal as a version of physicalism, the thesis that physics is the fundamental natural science; thus its way of “suturing” the phenomenal body onto the physical world is to maintain that perception and action are just “physico-mathematical correlations” whose explanation can be reduced to physical explanation. The central thesis of Part I of Phenomenology of Perception is that this application of the ideal of objective thought to the phenomenal body is a mistake. Merleau-Ponty expresses this claim in a striking endorsement of Descartes’s anti-materialist position: “It will never be made clear how signification and intentionality could inhabit molecular structures or cellular masses, and here Cartesianism is correct” (Merleau-Ponty 2012: 367). Thus, according to Merleau-Ponty, the physicalist’s objective conception of the body as “a chemical structure or a collection of tissues” which “is formed through a process of impoverishment” (Merleau-Ponty 2012: 367) fails to account adequately for central aspects of our being in the world such as our perceptual and motor powers and their inherent intentionality. He returns to this theme of “impoverishment” in a later comment about the objective body: The objective body is merely an impoverished image of the phenomenal body, and the problem of the relations between the soul (âme) and the body has nothing to do with the objective body, which has merely a conceptual existence, but rather has to do with the phenomenal body. (Merleau-Ponty 2012: 456) What is notable here is the claim that the objective body “has merely a conceptual existence”. For this would seem to imply that the conception of the objective body is just a fiction of the mistaken physicalist ideal of objective thought. But in Phenomenology of Perception Merleau-Ponty also writes of the objective body as if it were a genuine aspect of our embodiment which needs to be allowed for alongside the capacities which constitute the phenomenal body (see Merleau-Ponty 2012: 212, 241, 322). One idiom which attests to this dual conception of the body is that of the “ambiguity” of the body which is experienced when one brings one’s hands into direct contact, for this gives rise to “an ambiguous organization where the two hands can alternate between the functions of ‘touching’ and ‘touched’” (Merleau-Ponty 2012: 95). Merleau-Ponty describes these functions in a revealing way: in a case where one touches one’s right hand with one’s left hand while the right hand is itself touching an object, he writes: The right hand, as an object, is not the right hand that does the touching. The first is an intertwining of bones, muscles and flesh compressed into a point of space; the second shoots across space to reveal the external object in its place. (Merleau-Ponty 2012: 94) These roles of “touching” and “touched” are clearly those of what one might call the “phenomenal hand” and the “objective hand”. But he now goes on to write, concerning the

178

Maurice Merleau-Ponty

body, that this situation “prevents it from ever being an object”, which one might take to imply that the conception of the objective body is an illusion: Insofar as it sees or touches the world, my body can neither be seen nor touched. What prevents it from ever being an object or from being “completely constituted” is that my body is that by which there are objects. (Merleau-Ponty 2012: 94) The way to understand this passage, however, is to recognize that although Merleau-Ponty does not use the phrase “phenomenal body” here, this is what he is describing when he writes of the body “insofar as it sees or touches the world”. Hence when he goes on to write that “it” cannot ever be an object, the reference is just to “the phenomenal body”, so that there is here no claim concerning the objective body. After all, one could match the earlier passage with the following one: Insofar as it can be seen or touched, my body can neither see or touch. What prevents it from ever being a subject or from being “purely constituting” is that my body is in these ways an object. Here it would be clear that one was just writing of the objective body, and not asserting that, because it can be seen and touched, the body is in no way a subject. The conclusion to be drawn, therefore, is that the body’s ambiguity amounts to a dual aspect theory whereby the right hand, for example, can have both the aspect of touching and that of being touched. Although Merleau-Ponty takes it that these two aspects are mutually exclusive, there is no contradiction in his holding that they are aspects of one and the same thing since each obtains only relative to the relevant perspective – phenomenal or objective. In later writings Merleau-Ponty in fact rejects the assumption of mutual exclusiveness: in “The Philosopher and his Shadow” he writes “I touch myself touching; my body accomplishes ‘a sort of reflection’” (Merleau-Ponty 1964: 166).3 Similarly in The Visible and the Invisible Merleau-Ponty writes that the experience of a handshake with another person is one in which “I can feel myself touched as well and at the same time as touching” (Merleau-Ponty 1968: 142). Hence the conclusion to take is that which he affirms in the same book, that “my body is at once phenomenal body and objective body” (Merleau-Ponty 1968: 136). The question that now arises is how these two aspects of the human body, the phenomenal and the objective, are related. Since the phenomenal body comprises “a system of motor powers or perceptual powers” (Merleau-Ponty 2012: 155) and the objective body is an “intertwining of bones, muscles and flesh” (Merleau-Ponty 2012: 94), one suggestion might be that the “phenomenal” powers can be understood as products of the body’s objective constituents in the way in which the powers of familiar substances such as salt and water can be understood in terms of their physical constitution. But that is the physicalist hypothesis to which Merleau-Ponty is emphatically opposed. The obvious alternative to the physicalist’s “bottom-up” approach is, therefore, one which is, in some respects, “top-down” – i.e. takes it that the exercise of the perceptual and motor powers which are characteristic of the phenomenal body are fundamental elements of human life which are realized by the body’s objective constituents but not determined by them. The dualism here is not a dualism of substances, but of two aspects of human life. So the starting point for a consideration of this position is not Cartesian dualism, but Aristotelian hylomorphism, the dualism of form and matter as applied to the philosophy of mind. This is not the place for a detailed discussion 179

Thomas Baldwin

of the options and issues that confront positions of this kind, but the “naturalism” John McDowell presented in Mind and World is a good example of a position of this kind (McDowell 1994: 108–110), and I say more below about the way in which one can interpret Merleau-Ponty’s position as one of this kind.4

Bodily intentionality The main way in which Merleau-Ponty attempts to substantiate his position is by introducing two cases of people who have suffered serious injury, and showing that the nature of their injury is not adequately accounted for by a physicalist approach which relies on physical damage, including neural damage, but needs instead to be reframed in terms of a phenomenological approach which focuses on the ways in which their basic being in the world have been damaged by their physical injury. The first case concerns the phenomenon of a phantom limb, the often painful and troubling experience of those, who having lost a leg or an arm, feel as if they still had the limb that they know to be missing. Merleau-Ponty’s suggestion is that this phenomenon reveals a kind of bodily memory of an ability once possessed which is especially activated in situations in which the missing limb would have been used – for example, when someone who has lost an arm sees a plate of food that he would have reached for without thinking of it, this memory is felt as the ghostly presence of his missing arm. This account makes good sense, and one can see why Merleau-Ponty introduces it, for it suggests that the phantom limb phenomenon is best understood from within the phenomenological perspective of being in the world, that is being someone who remains “within the practical field that one had prior to the mutilation” (Merleau-Ponty 2012: 84). This account does not exclude a neurological account of the way in which motor abilities are entrenched by neural connections in the brain, such that when neural centers are activated by perception of situations suitable for moving a limb they activate proprioceptive sensors which give rise to the feeling of the limb even though it is now missing. On the contrary, Merleau-Ponty acknowledges that some such hypothesis is correct (Merleau-Ponty 2012: 88). Thus this case shows how Merleau-Ponty’s dual aspect approach to the body is supposed to work. In the course of his discussion of this case Merleau-Ponty introduces a further line of thought which is going to be central to his account of our being in the world, and especially agency. He starts from the thought that the motor abilities and predispositions which are manifested by the phenomenon of the phantom limb are typically states of which one is not conscious, and he then generalizes this point to suggest that what is revealed by this case is a web of sub-personal abilities and dispositions which organize much of our life beneath the level of personal existence: . . . we say that my organism – as a pre-personal adhesion to the general form of the world, as an anonymous and general existence – plays the role of an innate complex – beneath the level of my personal life. My organism is not like some inert thing, it itself sketches out the movement of existence. (Merleau-Ponty 2012: 86) One might think at first that the distinction Merleau-Ponty draws here between one’s “personal life” and the “anonymous”/“pre-personal” existence of one’s “organism” is an implication of the basic ambiguity of the phenomenal and objective aspects of human existence. But on reflection it is clear that this is exactly what Merleau-Ponty denies when he writes “My organism is not like some inert thing, it itself sketches out the movement of existence”. 180

Maurice Merleau-Ponty

That is, the basic perceptual and motor abilities which constitute our phenomenal body are features of one’s anonymous “organism”, the pre-personal being which operates only in the background. Nonetheless, he claims, these abilities, despite their sub-personal status, have a crucial role in enabling us to maintain our personal lives, for our personal freedom is founded upon our bodily organism. Hence we should replace the conception of agency as the control of a physical body by an abstract mind, and view it instead as the interplay between the pre-personal being in the world of our organism and a personal self which uses this being in the world to understand and change it: Taken concretely, man is not a psyche joined to an organism, but rather this back-andforth of existence that sometimes allows itself to exist as a body and sometimes carries itself into personal acts. (Merleau-Ponty 2012: 90). This is not a complete account of agency; but it indicates the direction of travel by which Merleau-Ponty will seek to elucidate its possibility. This line of thought is further developed through discussion of an Austrian soldier, Schneider, who suffered a brain injury during World War I and whose subsequent situation was described in detail by the psychiatric neurologists Adhémar Gelb and Kurt Goldstein. They describe one central aspect of Schneider’s disability as an inability to point to things, despite have the ability to grasp them. Thus he can undertake concrete movements, such as reaching out to grasp the things which he requires for his employment in making wallets – cutting pieces of leather and stitching them together, but he cannot perform abstract movements such as pointing to his nose. For Merleau-Ponty this disability reveals a disturbance in Schneider’s capacity for undertaking bodily movements which are not tied in to his current practical situation. Although Schneider can perform a military salute, he does so only by treating his situation as a whole as one in which a salute is required; he cannot simply put himself into an imaginary situation and act out what would be appropriate there. Merleau-Ponty summarizes this account of Schneider’s disability as follows: The normal person reckons with the possible, which thus acquires a sort of actuality without leaving behind its place as a possibility; for the patient, however, the field of the actual is limited to what is encountered in real contact or linked to these givens through an explicit deduction. (Merleau-Ponty 2012: 112) Thus the initial contrast between concrete and abstract movement is transformed into a contrast between Schneider’s limited repertoire of habitual actions based on his practical needs and a normal person’s open-ended capacity for actions that construct and explore novel possibilities that take us into new worlds. Merleau-Ponty describes how these new worlds then become “sedimented” in turn as “acquired worlds” (Merleau-Ponty 2012: 131) which provide a background for the construction of further possibilities, so that, for a normal person (but not Schneider), action has a significance which draws on their open-ended ways of enhancing their situation. Merleau-Ponty describes this significance as a kind of intentionality that operates at a deeper level than that of explicit thought: So let us say . . . . that the life of consciousness – epistemic life, the life of desire, or perceptual life – is underpinned by an “intentional arc” that projects around us our past, 181

Thomas Baldwin

our future, our human milieu, our physical situation, our ideological situation, and our moral situation, or rather, that ensures that we are situated within all these relationships. And this is what “goes limp” in the disorder. (Merleau-Ponty 2012: 137) Merleau-Ponty now uses this way of characterizing our situation to advance a central thesis of Phenomenology of Perception, that “motricity”, our capacity for bodily movement, is “original intentionality” (Merleau-Ponty 2012: 139), so that “Consciousness is originarily not an ‘I think that’, but rather an ‘I can’” (Merleau-Ponty 2012: 139). The claim here, that intentionality is not fundamentally a matter of mental representation, is one that I mentioned at the start of this chapter in the context of Merleau-Ponty’s critical discussion of Liepmann. In the following passage (to which Merleau-Ponty attaches his discussion of Liepmann as a footnote) Merleau-Ponty writes that Brentano’s famous “reference to an object” is to be understood as actually reaching out toward the object in question (which is presumably something the agent is perceiving), rather than as a representation of his future arm movement: The gesture of reaching one’s hand out toward an object contains a reference to the object, not as a representation, but as this highly determinate thing toward which we are thrown, next to which we are through anticipation, and which we haunt. Consciousness is being toward the thing through the intermediary of the body. (Merleau-Ponty 2012: 140)

Agency Toward the end of Phenomenology of Perception Merleau-Ponty returns to this distinction between the conception of the intentionality of consciousness as based on the “I can” of agency and the conception of it as based on the “I think” of mental representation. He now presents this distinction as one of the main conclusions of the book, emphasizing the difference between his new conception of bodily, or “operative” (as he now calls it), intentionality and the traditional conception of the “thetic” intentionality that is characteristic of judgment, and reaffirming the thesis which he had introduced through his discussions of phantom limbs and of Schneider’s case, that this operative intentionality is fundamental: We uncovered, beneath act or thetic intentionality – and in fact as its very condition of possibility – an operative intentionality already at work prior to every thesis and every judgment. (Merleau-Ponty 2012: 453) This “operative” intentionality lies at the core of agency as Merleau-Ponty conceives it. But it is important to consider first how it constitutes a kind of agency at all, for the way in which it is pre-personal might be thought to imply that the bodily movements required are no more cases of genuine agency than the movements of someone’s gut during digestion. As we have seen, in Phenomenology of Perception Merleau-Ponty used the term “organism” to contrast with “personal life” in the context of his discussion of the phantom limb phenomenon (Merleau-Ponty 2012: 86). So a good way to take the issue further is to look briefly at his

182

Maurice Merleau-Ponty

discussion of the behavior of “organisms”, as he calls them, – i.e. non-human animals, in The Structure of Behavior (Merleau-Ponty 1963). He argues here that what distinguishes organisms from complex physical systems is that the behavior of organisms is organized by reference to a milieu which is appropriate to the species of the organism (Merleau-Ponty 1963: 129–130). He uses the term milieu as a translation of the German term Umwelt (environment) whose use in this context comes from the work of the Estonian biologist Jacob von Uexküll who introduced the term to describe the meaningful aspects of an animal’s environment; and it is in effect von Uexküll’s position that Merleau-Ponty affirms in The Structure of Behavior.5 Merleau-Ponty explains this position in the following passage: We mean only that the reactions of an organism are understandable and predictable only if we conceive of them, not as muscular contractions which unfold in the body, but as acts which are addressed to a certain milieu, present or virtual: the act of taking a bait, of walking towards a goal, of running away from danger. The object of biology is evidently not to study all the reactions which can be obtained with a living body in any conditions whatsoever, but only those which are its reactions or, as one says, “adequate” reactions. (Merleau-Ponty 1963: 151) So what is distinctive about organisms is that their behavior constitutes “acts which are addressed to a certain milieu, present or virtual”. In developing this position Merleau-Ponty introduces the concept that, as we have seen already, is central to his account of human life in Phenomenology of Perception, that of the organism’s “phenomenal body”; and it is the organism’s phenomenal body which organizes its milieu: The gestures and the attitudes of the phenomenal body must have therefore a proper structure, an immanent signification; from the beginning the phenomenal body must be a center of actions which radiate over a certain “milieu.” (Merleau-Ponty 1963: 157). It is the final phrase here that is especially significant – the connection between the organism’s agency, as a “center for actions”, with its way of being in an environment (“milieu”). The message from these writings, therefore, confirms the thesis that because operative intentionality, be it human or animal, is a way of being in the world (Umwelt, milieu, or environment), behavior which manifests this kind of intentionality is a clear case of agency. As he puts it in his later lectures on Nature, life is “the opening of a field of action” (Merleau-Ponty 2003: 173). Although he does not directly address these issues in Phenomenology of Perception, Merleau-Ponty does introduce a line of thought that helps to develop the position further. In chapter III of the Introduction, he contrasts both causal explanations and reason-based explanations of depth perception with phenomenological explanations in term of “motives”, and one can appreciate from Merleau-Ponty’s discussion of this case what he has in mind when he describes “motivation” as an example of a non-thetic consciousness (Merleau-Ponty 2012: 50), “a sort of operative reason” (Merleau-Ponty 2012: 51). He also applies this conception of motivation to behavior, and the implication is that practical motives are often non-thetic, “operative reasons”; so where behavior is motivated by a subject’s situation in their world (milieu), it has its own rationality, the operative rationality of practical motives. And what is agency, if not the capacity for rational action?

183

Thomas Baldwin

This completes the case for Merleau-Ponty’s thesis that operative intentionality, the fundamental way in which perception and behavior constitute being in a world, suffices for agency. Before leaving the topic, however, there is one further issue to be settled. Merleau-Ponty’s thesis is not warranted simply by a verbal connection between talk of “intentionality” and “intentional action”; a more substantive connection is required, which I have suggested is to be found through his discussion of what it is for the life of an organism to be a way of being in a world. But there is a separate question one can now ask, namely whether Merleau-Ponty holds that it is a mark of agency that actions are, in the familiar sense, intentional – “agent-intentional”, as I shall call it. This is, of course, a thesis that is familiar from the writings of Davidson and many others – for example, Sartre.6 It has also been challenged, for example by O’Shaughnessy in the first edition of The Will.7 It is not sensible for me to attempt here to assess this thesis itself: there are many putative counterexamples to it which typically involve strange behavior that cannot be easily understood.8 But it is worth considering whether Merleau-Ponty himself subscribed to this thesis. So far as I’m aware, he never explicitly affirms it; but, equally, he does not explicitly reject it, though it has been argued that he holds that at least some actions are not intended (and not simply because there are some aspects of them which are not intended).9 But he certainly writes a good deal about the “practical intentions” of agents, and my own view is that he just takes it for granted that actions are agent-intentional because they are behavior which manifests the operative intentionality of a subject whose being is being in the world. For example, he writes: When I motion to my friend to approach, my intention is not a thought that I could have produced within myself in advance . . . I signal over there, where my friend is. (Merleau-Ponty 2012: 113) Thus an agent’s intentions are just the contents of the operative intentionality which informs their behavior. In this sense, therefore, all action is inherently intentional – including that of animals. Merleau-Ponty would endorse Wittgenstein’s comment: What is the natural expression of an intention? – Look at a cat when it stalks a bird; or a beast when it wants to escape. (Wittgenstein 1953: § 647)

The will By now it will be clear why Merleau-Ponty rejects the picture of us as “a machine surmounted by a will” (Merleau-Ponty 2012: 452). He has emphatically rejected the conception of the body as a machine, and replaced it with the conception of the body as a natural subject, the vehicle of our being in the world, which, as we have just seen, brings with it the capacity for agency. But what of the will? Merleau-Ponty does not provide an extended discussion of this topic; but in Phenomenology of Perception he alludes several times to aspects of human life which are “voluntary”, contrasting them with the “pre-personal life of consciousness” as rational and reflective, including judgment and practical deliberation (see Merleau-Ponty 2012: 216, 362, 460). This contrast brings out the need to clarify the relation between the pre-personal life of consciousness and the reflective, rational, personal life of someone capable of voluntary practical deliberation. Merleau-Ponty makes it clear that these two aspects of human life 184

Maurice Merleau-Ponty

belong together. Just as the perceptual and practical abilities of living organisms depend on the biochemical properties of their material structures, without being reducible to them, the cognitive and practical abilities of rational persons depend on the perceptual and physical powers of their body. In The Structure of Behavior Merleau-Ponty calls these relations of dependence “dialectic”, by which he seems to have in mind the enhancement of what was present at a lower level through its incorporation in some more complex whole. In Phenomenology of Perception Merleau-Ponty adds Husserl’s conception of Fundierung (founding) to his account of this process, so that he now describes it as a “dialectic between form and matter” whereby “form absorbs matter” in that there is a “perpetual taking up of fact and chance by a reason that neither exists in advance of this taking up, nor without it” (Merleau-Ponty 2012: 128–129), and this way of thinking about the matter connects with my earlier suggestion that the relation between phenomenal and objective body be thought of in terms of an Aristotelian relation between form and matter. As we have seen he takes it that the capacity for rational agency builds on the motives and values that inhere in our pre-personal being in the world. Indeed one can readily adapt his account of the relation between thetic and operative intentionality which I have cited earlier (Merleau-Ponty 2012: 453) to this case. It is as if – We uncovered beneath the capacity for explicitly rational action – and in fact as its very condition of possibility – a capacity for motivated action already at work prior to every thesis and every judgment. But a further point is suggested by the following passage: No more can aphonia be considered voluntary. The will presupposes a field of possibles among which I choose: here is Pierre, I can choose to speak to him or not. (Merleau-Ponty 2012: 165)10 The reference here to “a field of possibles” is reminiscent of Merleau-Ponty’s emphasis on “the capacity of orienting oneself in relation to the possible” as the “essence” of human life (Merleau-Ponty 1963: 175–176) and his comment in the context of his discussion of Schneider’s case that “The normal person reckons with the possible” (Merleau-Ponty 2012: 112). Thus one can take from these passages the suggestion that voluntary acts are characteristic of personal human life, whereas the world of non-human animals and of our pre-personal bodily existence is limited to what is familiar and actual. What now needs some explanation is what is required for this capacity to reckon with the possible. Merleau-Ponty gives this question little attention; but there is one suggestion which seems to me worth exploring briefly, namely that Merleau-Ponty’s account of language offers a way into the basis for a capacity which can then be used as a background for an account of voluntary action. Merleau-Ponty presents his account of language in Phenomenology of Perception, in Part I, Chapter VI: (“The Body as Expression, and Speech”). In the present context two claims are especially relevant: first, that “speech does not translate a ready-made thought; rather, speech accomplishes thought” (Merleau-Ponty 2012: 183); second, that this accomplishment rests on the “ultimate fact” that “speech gives rise a new sense (sens)” (Merleau-Ponty 2012: 200): Thus, we must recognize as an ultimate fact this open and indefinite power of signifying – that is, of simultaneously grasping and communicating a sense – by which man transcends 185

Thomas Baldwin

himself through his body and his speech toward a new behavior, toward others, or toward his own thought. (Merleau-Ponty 2012: 200) This “ultimate fact” is the fact that we are able to create new meanings through speech in which we use more or less familiar sounds (words) in new ways. So the fundamental relation between thought and language is not, Merleau-Ponty insists, that through learning language we learn to translate our own inner thoughts into sounds which make them available to others. Rather, in learning language we acquire the ability to express new meanings and thus to have thoughts which we could not otherwise have had. Merleau-Ponty emphasizes that speech is a type of behavior, a vocal “gesture” as he puts it (Merleau-Ponty 2012: 190). But when he writes that the speaker “transcends himself through his body and his speech toward a new behavior toward others or toward his own thought” he intimates that language is central to the transformation of the pre-personal consciousness of the human animal into the voluntary and rational life of a human person. In his account of this transformation Merleau-Ponty emphasizes the role of speech in giving meaning to novel possibilities and also providing the structure for an intersubjective world of shared meanings and culture: Speech is the excess of our existence beyond natural being. But the act of expression constitutes a linguistic and cultural world, it makes that which aimed beyond being fall back into it. This results in spoken speech, which enjoys the established significations as if they were an inherited fortune. On the basis of these acquisitions, other authentic acts of expression – those of the writer, the artist, and the philosopher – become possible. (Merleau-Ponty 2012: 203) Indeed Merleau-Ponty takes things further, remarking that because speech provides us with a way of refining our arguments and conclusion, it provides “a privileged place for Reason” (Merleau-Ponty 2012: 196). Similarly, he remarks that “speech establishes the concordance of myself with myself and of myself with others” (Merleau-Ponty 2012: 412) and this understanding of oneself and of others is central to reflective thought. So the development of language makes possible the cultural world we share with others and the rational reflective life that is characteristic of voluntary thought and action. To confirm this conclusion, however, it is important to check that it offers a way into the basis of a capacity to reckon with the possible which Merleau-Ponty identified as a presupposition of voluntary action and distinctive of human life in general. It is, I think, not difficult to satisfy oneself on this point. For it is central to Merleau-Ponty’s account of speech that we can use language to give a new sense to our words in a way which makes sense of new possibilities. So the connection between language and possibility is straightforward, as is the connection with practical deliberation. This is typically an assessment of the practical possibilities inherent in one’s situation: what language contributes is a way of making these possibilities explicit and thereby making it easier to reason concerning them. With language on hand, the motives (implicit operative reasons) for action characteristic of one’s pre-personal existence can be articulated by oneself as explicit reasons and assessed as such. Where they conflict, one will need to make a choice between them in the light of the “spontaneous valuations” (Merleau-Ponty 2012: 465) which enter into our experience of the world. So voluntary deliberation, so far from being a sham, as Sartre notoriously maintained,11 is an essential ingredient of one’s consciousness of oneself 186

Maurice Merleau-Ponty

as someone who can in principle give meaning to their life through their own deliberate acts, their own Sinngebung décisoire as Merleau-Ponty puts it. There is now one final question to be addressed, namely how it is that one’s decisions and other acts of will lead to bodily actions. It is clear that for Merleau-Ponty this relation is not a causal one; instead it is to be in some sense “rational”. But it is not obvious how decisions add to the reasons for action which have already been taken into account in making the decision. Merleau-Ponty does not, so far as I can see, address this issue; but there is a way of extending the appeal to language to provide a solution. It is that one should construe decisions as self-addressed instructions concerning how one should act. For just as instructions by others with some authority over one give one a reason for doing what one is told to do over and above one’s other reasons for doing it, a self-addressed instruction gives one a reason to do what one is telling oneself to do over and above the reasons on which the decision is based. So if decisions are conceived as self-addressed speech-acts of this kind, they will provide a reason of their own for doing what one has decided to do. Of course, our wills are not omnipotent; our decisions can be trumped by powerful temptations or anxieties. But these failures do not undermine the importance of this capacity for self-direction in the life of a rational, reflective, person. I proposed earlier that one might use Merleau-Ponty’s account of language to suggest how the emergence of language transforms the consciousness of a human animal into the “voluntary and rational life” of a human person by providing the basis for a reflective understanding of oneself. This final suggestion that acts of will be regarded as self-addressed instructions is an extension of this transformation into the domain of rational discourse, so that language does not just make practical reasoning possible but adds to our rationality by incorporating a recognition of the significance of the decisions which are the conclusions of our reasoning. It gives us a new kind of reason. Since this new capacity is one in which by acting in the way in which you have decided to act you affirm your own self, there is a sense in which “you are your act” – as Merleau-Ponty puts it in the conclusion of his Phenomenology of Perception.12

Related topics Chapters 11 (on Sartre), 17 (on Dreyfus), 21 (Hanna).

Notes 1 References to Phenomenology of Perception are to the translation by D. A. Landes, which I have occasionally modified. 2 Merleau-Ponty’s distinction between two conceptions of the body, phenomenal and objective, is influenced by Husserl’s distinction between Leib (the lived body) and Körper (the ‘corporeal’ body) – as in Husserl (1989: 352–354). But there are important differences between their positions, see Carmen (1999). 3 This paper is a tribute to Husserl, and, among other things, explores the conception of the lived body (Leib). The passage quoted from Merleau-Ponty’s paper is similar to Husserl’s remark “my Body as touched Body is something touching which is touched” (Husserl 1989: 155). On the theme of embodied reflection see also the conclusion of the seventh sketch “Man and Evolution: the Human Body” in Merleau-Ponty’s 1959–60 lectures on Nature (Merleau-Ponty 2003: 273). 4 One might seek to use the dual aspect theory developed and defended by Brian O’Shaughnessy in The Will as a model for thinking about Merleau-Ponty’s position (O’Shaughnessy 1980). Although there would be much to be learnt from a comparison between these two positions, despite the radical differences between their presumptions and methodologies, it would require a detailed exposition and examination of O’Shaughnessy’s theory which is not appropriate here.

187

Thomas Baldwin



188

15 PAUL RICŒUR A phenomenological hermeneutics of meaningful action Timo Helenius

Paul Ricœur’s (1913–2005) philosophy can be described as a hermeneutic phenomenology that extensively covers questions pertaining to action, agency, volition, intention, trying, reasons/causes, explanation, freedom, and responsibility. Ricœur’s philosophy of action is distinct from the analytic discourse, however, in that it generally sees the question of action not conceptually but hermeneutically as the question of meaningful action. In this sense Ricœur’s philosophy of action functions simultaneously at several levels; embodied action and practical action are tied in with the activity of interpretation that also calls for an analysis of its own. This second question, in short, is about the readability and understandability of action that Ricœur resolved to explore under the heading of reconfigurative emplotment. As in the case of meaningful texts, meaningful action exposes an actively engaged hermeneutic subject who narrates his or her situated being through interpretative action: “I see in the plots we invent the privileged means by which we re-configure our confused, unformed, and at the limit mute temporal experience” (Ricœur 1984: xi). Action that carries some significance or meaning indicates of an actively interpreting subject who is personally involved and engaged in the intersecting embodied, temporal, practical, social, institutional, linguistic, and cultural spheres of one unified experience of one’s life. Ricœur’s stance from early on is that “to exist is to act” (Ricœur 1966: 334). Such action does not only mean embodied or physical action whereas it is always – insofar as it is meaningful, that is, unceasingly hermeneutical – tied in with linguistic and expressive action; for this reason these both are interlinked in the realm of ethical and political, or practical, action. Without having the opportunity to discuss it here, it should therefore be noted that in Ricœur’s thought there is an integral expansion of the phenomenologico-hermeneutical theory of action in the direction of practical, ethical, and political action; this can be seen, in particular, in Oneself as Another (1990), The Just (1995), and Reflections on the Just (2001). Even though his moral and political philosophy has also been read to present a distinct realm of philosophical concern, in Ricœur’s own view all these “distinct centers of interest” (Ricœur 1995c: 444) are inseparable from his overall hermeneutic phenomenology of action that, in turn, concerns the agent or the actively engaged and situated hermeneutic subject Ricœur was ultimately after in his philosophy. In spite of his thoroughly continental manner of doing philosophy, Ricœur was also an informed commentator of analytical theories of action. His analyses might explicitly contest 189

Timo Helenius

those proposed by Anscombe, Davidson, Strawson, and many others, but his aim is rather to complement such theorizations that are already on the way of addressing the philosophical problems Ricœur himself was interested in discussing. Even though this essay will begin by Ricœur’s relatively late commentary on the Anscombe-Davidson thesis on intentional action, we will therefore also introduce Ricœur’s own hermeneutic phenomenology of “meaningful action” by paying special attention to his early phenomenological analysis of action as well as his relating semantic analyses of action. This introductory essay will argue that, while in many ways engaging itself with the analytic manner of pursuing philosophy, Ricœur’s philosophy of a “living, acting, and suffering subject” was consistently asking about meaningful action, or action as meaning-bearing, while elaborating a hermeneutic phenomenology of “I can” (Ricœur 1992: 112).

Ricœur’s criticism of Anscombe and Davidson In the field of Analytic philosophy, the philosophy of action relates to and draws further clarity from the philosophy of language and the philosophy of mind. In terms of the second, the point is that insofar as there is an agent that manifests itself in physical or corporeal action, there should also be a clarification regarding the relation between the agent’s mental states and her bodily movements. In turn, analytical philosophy of language draws attention to the inherent problems regarding the medium of language that is used for providing such clarifications. For example, giving an account of an agent’s action (“Tobias scored a goal”) can be ambiguous in the sense that it does not necessarily distinguish between the antecedent motivational mental conditions or intentions (willingness to score a goal), the act itself (the kicking of the ball), the causal chain of interlinked actions (that all constitute, for example, the leg to move in such a way that it hits the soccer ball), and the causal outcomes, results, or effects of that action (a broken toenail and the flying of the ball deep into the goal in such manner that the net swings back). Furthermore, a description of action may, justifiably, make a philosophical analysis complex by indicating a shared intention – or at least of a joint one – and related mutual action. In a rather loose reading, the phrase “as the best striker, Tobias scored a goal” would indicate such a state of affairs in which Tobias acted as a member of a soccer team that as a whole was intentionally acting toward scoring; that it was Tobias who scored a goal stands out merely as an insufficient or a partial description of this group effort and action. In order to understand action, we thereby also need to pay attention to the ascriptions given of actions; this expands to the question of explaining actions. All in all, this brief example shows that extended analyses of action are as relevant as are the more tightly focused questions pertaining to intention, reasons or causes, and so on. Ricœur’s understanding of action is not confined to such tight analytic mode of philosophizing, and yet he also found himself to be critically interested in that particular discourse concerning action. Ricœur began teaching courses at the University of Chicago in 1970, and this brought him in substantial contact with Analytic philosophy. Even though phenomenology and hermeneutics were then perceived to have nothing much to share with the analytic pursuits, Ricœur himself adopted a different kind of approach. “Far from treating it as an enemy” (Ricœur 1995a: 31), Ricœur recounts in his intellectual autobiography, Analytic philosophy complemented and thereby strengthened his own explorations. The reason for such appreciation was the analytic aspiration for precision and clarity to which our own example just alluded to. Ricœur’s application of these analytic resources was, however, critical and appropriative as it, for example, moved freely between the philosophy of action and the philosophy of language. In Ricœur’s words, “it was mainly in the philosophy of ordinary 190

Paul Ricœur

language that I found the most reliable support” (Ricœur 1995a: 32).The analyses provided by Austin, Searle, and others were easily relatable to Ricœur’s own corollary explorations of that time, but they also led Ricœur to a path that bore explicit fruit in his analyses of action in the much-read 1990 work Oneself as Another – a work that among other themes discusses Ricœur’s uneasiness concerning “an agentless semantics of action” (Ricœur 1992: 87). Ricœur’s main concern is that the analytic approach to action is too limited in its analyses that seem not to heed questions about the relation between action and its agent particularly as they would concern attribution, ascription, self-referentiality, practices, ethical and political prescriptive attitudes, and responsibility. “It is perhaps due to the very style of Analytic philosophy and to its almost exclusive preoccupation with description, that it ignores problems pertaining to attestation, which cannot be reduced to a criteriology suited to description” (Ricœur 1992: 72). Placing the focus on the questions of “what” and “why” – on something that occurs – the analytic theories do not sufficiently open up the question of the “who” of action. Put differently, Ricœur argues that these theories detach the self from action and do not thereby consider it as a real “someone”: “the agent of action will have little resemblance to a self who is responsible for both words and actions” (Ricœur 1992: 57). In his view, analytic theories are reductive in a sense that they tend to anonymize and impersonalize the engaged and situated acting self by turning it into “an ethically neutral agent, as free of blame as of praise” (Ricœur 1992: 57). Emphasizing the need to make a robust distinction between events that simply happen and actions that make something happen, Ricœur therefore argues that the narrative and the teleological aspects of action should not be cast aside insofar as action is to be of full-fledged philosophical concern. Approaching the question of “what” from the side of “why” does not suffice in the attempt at fleshing out the whole of the problematic of action that should also address the aspect of “who.” The lack of attention to this question is clear, Ricœur argues, in the discussion that follows from the work done by Elizabeth Anscombe and Donald Davidson. Even though Anscombe’s distinction between knowing-that (viz. event) and knowinghow (viz. action) brings out the aspect of immediately present knowledge “without observation” and even relates that to “practical knowledge,” Anscombe’s position is defective “by omission” in that it is “concentrated on the ‘what’ of action without thematizing its relation to the ‘who’” (Ricœur 1992: 62). According to Ricœur, this opens up the danger of collapsing human action back in to the realm of mere causal events, an outcome that would surely be considered undesirable. A similar kind of difficulty concerns the distinction between motives and causes that, according to the same Humean-based approach, should nevertheless be strictly held: “The internal, necessary (and in this sense, logical) connection characteristic of motivation is incompatible with the extrinsic, contingent (and, in this sense, empirical) connection of causality” (Ricœur 1992: 63). The critical point Ricœur wishes to make is that insofar as in “a logical chain of involvement” action is understood in a manner that “we move from wanting to trying to do and, finally, to doing” (Ricœur 1992: 63), such view does not in fact mutualize wanting and action (as one could expect in terms of motives) whereas they are cut off of each other. Wanting is treated “as an internal event, logically distinct from the action” (Ricœur 1992: 64). In short, the discussion concerning actions and their motives is not treated distinctly from the discussion concerning events and their causes; the attempt at trying to keep “the two universes of discourse” distinct from one another “has been unable to resist the assaults of a conceptual analysis; . . . the result has been the constant encroachment of one on the other to the point of rendering problematic the very principle of their dissociation” (Ricœur 1992: 65). Taking into account the agent – virtually absent in these theorizations – in form of the question of “who” would, in Ricœur’s view, help the philosophy of action by preventing it from collapsing into such assaults and encroachments. 191

Timo Helenius

The “conceptual impressionism” Anscombe undertakes in Intention is also defective in the sense that the very concept of intention is dealt with in Anscombe’s conceptual analysis in a manner that has “deliberately turned its back on phenomenology” (Ricœur 1992: 67). What Anscombe has resolved to address is not intentionality – as one could assume if one were to follow Brentano and Husserl – whereas her analysis, following the analytic approach, targets the question of a declared intention. Here, however, Ricœur points out a difficulty that results in from not grounding the analysis on the intentional structure of consciousness itself as “the surface grammar of declared intentions is not clearly defined” (Ricœur 1992: 68). In Anscombe’s theory, such confusion can only be resolved by strictly focusing on the notion of “intending to” in the sense of “a reason for acting,” meaning that a severe restriction in terms of the starting point of the analysis will also have to be adopted: “One therefore begins with the adverbial usage of the term ‘intention’,” that is, Anscombe’s conceptual analysis begins in a manner that leads it to explicitly focus on the question of “why” (and “what”) instead of also keeping the “who” in sight. Anscombe’s analysis “privileges the use that exemplifies in the least explicit way the relation of intention to the agent” (Ricœur 1992: 68). In Ricœur’s critical reading of Anscombe, her theory privileges the objective and causal and not the subjective and motivational side of action – as seen in “the inability to thematize attestation” (Ricœur 1992: 73) – resulting in the unintentional tendency to view actions as events. In Ricœur’s view, this tendency was extended in Davidson’s Actions and Events that argued for a logically and ontologically reductive “ontology of the impersonal event,” rendering “actions themselves a subclass of events” (Ricœur 1992: 74). Such approach precludes the inclusion of an analysis concerning the relation between an action and an agent. As Ricœur puts it, the severe defect that Davidson’s theory contains is the “failure to return action to the agent” (Ricœur 1992: 74). Ricœur’s observant criticism makes it clear that the main reason for such failure is Davidson’s reductive stance toward the teleological aspect of intentional action; “this descriptive feature is quickly subordinated to a causal conception of explanation” (Ricœur 1992: 74). Even though an action is distinct from other events due to intention, Davidson nevertheless treats actions in light of rigorous causal explanations. And to be even more precise, Davidson’s position “that rationalization is a species of causal explanation” (Davidson 2001: 3) results in a strong preference for event-causality in a manner that also effectively subsumes the notion of intention, thereby rendering it part of a causalist schema. In Davidson’s theory, intention is equated with the reason for doing that itself is a rationalization and as such a form of causal explanation. Put differently, Davidson holds that a description of action is a rationalization and therefore an explanation of action that relies on the causalist view. According to Ricœur, the crucial omission that results from this is to neglect the teleological view of explanation that would retain a clear connection with the agent’s motives. Using Charles Taylor’s argument for a teleological rather than a causal explanation of action, Ricœur extends his criticism of Davidson by demanding that “the fact for an event to be required to obtain a given end is a condition for the appearance of that event” (Ricœur 1992: 78). Put differently, Ricœur understands a teleological explanation to be “an explanation in which the order is itself a factor in the production of an event; it is a self-imposed order” (Ricœur 1992: 78). This means that – at the descriptive level – the agent’s motive for action counts toward and produces the action that can, by the virtue of this internal relation, also be considered as an explanation of this action: Classifying an action as intentional is determining by what type of law it is to be explained and, by the same token, ruling out a certain type of explanation. In other words,

192

Paul Ricœur

it is deciding on the form of the law that governs the action and at the same time ruling out the possibility that this is a mechanical law. (Ricœur 1992: 79) To Ricœur’s satisfaction, such model of teleological causation would push the notion of intention more toward the phenomenological manner of describing action that also allows for addressing the volitional and temporal dimensions of action – these both indicate of the need to ascribe action to an agent, thereby bringing intentionality and agency to the forefront. The major defect in Davidson’s theory, valuable in its own right in its own context, lies precisely in this omission: is not an ontology of events, founded on the sort of logical analysis of action phrases conducted with the rigor and subtlety of Davidson’s analysis, condemned to conceal the problematic of the agent as the possessor of his or her action? (Ricœur 1992: 85) In Ricœur’s view, approaching the question of action from the phenomenological vantage point would resolve the problems inherent in Davidson’s ontology of impersonal event that only allows for “an agentless semantics of action” (Ricœur 1992: 87).

Ricœur’s phenomenology of action We now move to providing a positive account of action the manner Ricœur proposes to have it. As is already clear, Ricœur’s thought is situated at an intersection of various philosophical traditions, later also including those of Analytic philosophy, that are not reconcilable with one another. The so-called continental tradition itself is a non-unified field of various approaches that, when brought together, expose their conflictual and dissonant characters. As Ricœur puts it himself, I have found myself from the start in a multicornered debate among various heritages: that of existential philosophy (Gabriel Marcel and Karl Jaspers), that of reflexive philosophy ( Jean Nabert), that of phenomenology (Husserl), and that of hermeneutics (Heidegger and Gadamer). (Ricœur 1995c: 445) In spite of openly acknowledging the difficulties regarding the ultimately unachievable harmonization of these various discourses and heritages, Ricœur nevertheless maintained that his philosophical pursuit is in general both phenomenological and hermeneutical as well as concerned about human existence. In one of his last interviews, Ricœur explained this dual effort by clarifying that the goal of his hermeneutic reflection was to reveal the capable, meaning actively able, human being: The ultimate purpose of hermeneutic reflection and attestation, as I see it, is to try to retrace the line of intentional capacity and action behind mere objects (which we tend to focus on exclusively in our natural attitude) so that we may recover the hidden truth of our operative acts, of being capable, of being un homme capable. So if hermeneutics is right, in the wake of Kant and Gadamer, to stress the finitude and limits of consciousness, it is

193

Timo Helenius

also wise to remind ourselves of the tacit potencies and acts of our lived existence. My bottom line is a phenomenology of being able [une phenomenologie du je peux]. (Ricœur and Kearney 2004: 167) In spite of a multitude of hermeneutical studies concerning language and linguisticality – and, in general, cultural situatedness and productivity – all these relate to the fundamental phenomenological pursuit of exposing the contours of human experience through the vantage point of being able or “I can.” Put differently, in Ricœur’s thought human action and linguisticality are organically related to each other and warrant complementary analyses that highlight the crucial aspects of human existence and of being able: I will say that the place of astonishment, and hence that of questioning, has been twofold from the very start. On the one hand, one can rightfully say that action has always been the organizing center of my philosophical reflection under a variety of headings: first, the voluntary and the involuntary, then desire and effort – raised to the ‘metaphysical’ level of the desire to be and the effort to exist – and, finally, the power of acting, in Oneself as Another; on the other hand, one can consider language to be the organizing focus of many investigations. (Ricœur 1995c: 444–445) As is clear, in this context the philosophical explorations into the “acts of our lived existence” are not confined to those focusing more specifically on intentional capacity and action, but the question of human intentionality should be regarded as implicitly present and relevant for the broader or refashioned scope of investigation. The previous section focused on Ricœur’s criticism of the Analytic philosophy of action in particular. In this section we will steer our exploration in the phenomenological direction on the basis of Ricœur’s own work on action. In spite of his willingness to engage in the analytical discussions and to draw support and further clarity from them, Ricœur remained deeply invested in the phenomenological tradition. Ricœur’s semantic analyses of action also manifest this point. As Ricœur puts it in his 1995 intellectual autobiography, the question concerning the semantics of action brought again forth his “initial interest in the phenomenology of the voluntary and the involuntary” (Ricœur 1995a: 31).The analytic theory of action that Ricœur begun to integrate in his hermeneutico-phenomenological explorations of human action in the early 1970s was clearly helpful as it provided “the complement of a logical semantics” that helped “to strengthen the linguistic semantics” he had pursued (Ricœur 1995a: 32). Such refashioned philosophical analysis of the action-discourse (and discourse-action), in spite of its indisputable inherent value in its own context, was nevertheless to serve a more long-standing philosophical interest. Put briefly, Ricœur’s substantial engagement with analytic action-theory provided him with the opportunity to re-examine and sharpen the question of the capable self, the “I can” – always checked by the “I cannot” as implied by effort and consent – or a subject’s “fundamental possibilities” (Ricœur 1966: 3). Focusing in his later work on intention rather than the will’s “project,” the conceptual and theoretical means may have thereby slightly changed, but the overall exploration would still indisputably bear resemblance to his earlier work. As Ricœur puts it elsewhere, “the philosophy of will was already at its inception a philosophy of action” (Ricœur 1995b: 346). In spite of its various reformulations in Ricœur’s work, it is the question of the living, acting, and suffering self that Ricœur pursued from his 1950 dissertation work Freedom and Nature even beyond Oneself as Another to his last work The Course of 194

Paul Ricœur

Recognition (2004). Understanding Ricœur’s early phenomenological explorations pertaining to this theme will thereby unlock the full breadth and depth of his philosophy of action. The preceding discussion reflects Ricœur’s 1970–1971 and 1971–1972 lecture courses in Louvain that kept their focus on the semantics of action. As the course title Le discours de l’action indicates, Ricœur’s angle of approach was unsurprisingly that of expressed experience, or, having a sense of human experience and action that is brought to language. In these lecture courses, published in French in 1977, Ricœur’s detailed attention nevertheless opens up as a full-fledged linguistic analysis in that it spans from the conceptual to the propositional and furthermore to the discursive level of analysis. In this course of interlinked analyses, Ricœur’s questions vary from those regarding the many possible ways to pursue the discourse of action (Parsons, Touraine, Aristotle, Moore, later Wittgenstein, Anscombe, Austin, Strawson, Husserl) to discussing the definitions given of the key concepts in that discourse (such as action, intention, reasons for acting, motivation, cause, desire, wanting, willing, choice, agent, agency, and responsibility1), to speech-act theory, performatives, and volition (Austin, Searle, Benveniste, Grice, Kenny), the philosophical problems of agency (agent-causation, motivation, intention, mental acts, explanation of action), and furthermore to the relation between linguistic analysis and phenomenology. It is this last one that we will now need to pay attention to in order to see how Ricœur’s semantics of action relates to his phenomenology of action. In spite of Ricœur’s insistence that regardless of their points of convergence – the one clarifies the essential structures of experience, and the other those of utterances – phenomenology remains distinct from linguistic analysis that, due its focus on language, has lost sight of the origins of the experience seeking expression. In short, “phenomenology operates at a different strategic level from that of linguistic analysis” (Ricœur 1977a: 120). This also means, however, that “there is no opposition between two adversary theories at the same level whereas there is a difference between two strategic levels; . . . phenomenology gives an ‘experiential’ foundation to utterances, and utterances give an ‘expression’ to experience” (Ricœur 1977a: 127).The view that phenomenology is about the fundamental structures of experience, and linguistic analysis about the manifestation of this experience, is why Ricœur consistently pressed the point that it is important to take into account the explanatory resources offered by Analytic philosophies of language and action in his own “linguistic phenomenology” (Ricœur 1977a: 14–15, 115) as these serve it by bringing further philosophical clarity. An indication of this importance is that Ricœur borrows the very term “linguistic phenomenology” from Austin who likewise stressed that a linguistic clarification should be in the service of analyzing experience – utterances are operative not only on the linguistic plane. Even though Ricœur in the end will retain the phenomenological manner of analysis to be the most fruitful, his engagement in the analytic discourse is valuable for these reasons. This does not mean, however, that the linguistic analyses would in any manner overwrite Ricœur’s ultimately phenomenological aim. This conviction is clearly stated in Le discours de l’action in that Ricœur’s aim was “to explore the contribution of language to the philosophy of action” in the sense of analyzing the “discourses in which human being speaks his doing (l’homme dit son faire)” (Ricœur 1977a: 3, 5). To understand what this statement means requires that we situate it in the context of Ricœur’s repeated claim – made in The Symbolism of Evil (1960), Freud and Philosophy (1965), and The Rule of Metaphor (1975) – that poetic expressivity, meaning innovative linguistic activity, “makes” human being that would otherwise remain mute and unreflectible: “Bachelard has taught us that the image is not a residue of impression, but an aura surrounding speech: “The poetic image places us at the origin of the speaking being.” 195

Timo Helenius

The poem gives birth to the image; the poetic “becomes a new being in our language, expressing us by making us what it expresses; in other words, it is at once a becoming of expression, and a becoming of our being. Here expression creates being (l’expression crée de l’être) . . . one would not be able to meditate in a zone that preceded language”. (Ricœur 1977b: 215. Cf. Ricœur 1969: 13; Ricœur 1970: 15–16) Put differently, the human experience of being is expressed and made evident in language. For this reason, any phenomenology that aims at exploring and exposing the fundamental structures of that experience will have to approach its task as a linguistic phenomenology. “Language,” Ricœur argues, in turn, by quoting Jean Ladrière, is “the structuring of the meaningful life” in that it is the objective correlate, or the noematic aspect, of intentional experience that phenomenology has set to explore (Ricœur 1977a: 6, 13–14). These convictions led Ricœur to be also interested in those analytic means that the ordinary language philosophy in general and the speech act theory in particular offer. In short, dire, c’est faire or “to speak is to do” (Ricœur 1977a: 8). This view finds its correlate in the phenomenological stance that “all consciousness is consciousness of ” something (Ricœur 1977a: 14, 122). In this sense it certainly holds that the philosophy of action has a linguistic and phenomenological “double constitution” as it ultimately is about l’action sensée, that is, about sensible or meaningful action (Ricœur 1977a: 17, 128). But as it comes to relating these linguistic analyses to phenomenology, phenomenology can always claim preference due to a certain hierarchy at the level of analyses: “At the level of foundation, phenomenology takes priority as it sets up the declaratory framework on the noematic framework; what one speaks is built on the meaning of experience” (Ricœur 1977a: 15, 121–122). Moreover, as Ricœur puts it in “Existence and Hermeneutics,” linguistic analysis always stops short in that it keeps its focus on language that itself, “as a signifying milieu, must be referred to existence” (Ricœur 1974: 16). In terms of Ricœur’s overall project, Le discours de l’action and the thematically similar analyses that followed this initial discussion with Analytic philosophy thereby merely “redo” (refaire) the more fundamental phenomenology of the voluntary and the involuntary – exposing the agent, or the active corporeal subject, through her actions – that Ricœur worked on in the 1940s and 1950s. We thereby arrive at Ricœur’s phenomenology of action; Ricœur himself attempts at giving an overview of it in Le discours de l’action. Drawing a distinction between Husserl’s phenomenology and that of his, Ricœur adopts the Merleau-Pontian view that stresses corporeity, or embodied being-in-the-world, and the actions of the will as the phenomenology of “I can” (Ricœur 1977a: 24, 31, 128–131). The key term here is “one’s own body,” le corps propre, as it is exposed as the “domain of motivation,” “the organ of voluntary motion,” and also as “the dispositional anteriority” or as that what remains “the absolute involuntary” (Ricœur 1977a: 129). In Ricœur’s view, this manner of framing phenomenology not only avoids the problems relating to Husserl’s notion of the transcendental ego but also constitutes the breaking point between phenomenology and linguistic analysis that, due to its methodical semantic focus, is unable to thematize the body. But to more properly understand why the notion of body is so prominent, we will have to comprehend its interrelatedness with the world that constitutes the extended and corporeal field of its experience and action. The world, Ricœur argues, is disclosed as the stage of one’s action; the subject, or the self, is in the action that is done, that is, in “the works done by me” (Ricœur 1977a: 126). It is to this at once hermeneutic and reflexive phenomenology of action that we will now turn. At this point we will need to place our focus again on Ricœur’s late statement that his “bottom line is a phenomenology of being able” (Ricœur and Kearney 2004: 167). This claim is 196

Paul Ricœur

materially substantiated in The Course of Recognition (2004) and Oneself as Another (1990) that both draw from Ricœur’s early phenomenological works and explicitly address the question of the living body. This phenomenology of action and embodied agency dates back to Freedom and Nature (1950), a work that would be more properly translated as The Voluntary and the Involuntary (the French title is Le volontaire et l’involontaire). Already in this early – and massively rich – work Ricœur stresses the idea that a human being finds himself in and through his capabilities that are manifested in objectifiable action: in projecting myself as the subject of an action, I affirm myself capable of that action . . . I feel myself capable ( je me sens capable), as an incarnate being situated in the world, of the action which I intend in general. (Ricœur 1966: 203) The critical point can be summarized by using Ricœur’s “Existence and Hermeneutics” that considers the Cartesian cogito as vain; even though it holds that the ego in a way posits itself in its intentional thought-action (“I exist insofar as I think”), the ego of ego cogito is nevertheless not immediately present to itself whereas it has to be reflectively “recaptured in the mirror of its objects, of its works, and, finally, of its acts” (Ricœur 1974: 17). The importance of Ricœur’s intentional use of the reflexive verbal form “je me sens” is thereby in that it expresses what takes place through that very expression: “reflection is nothing other than the appropriation of our act of existing by means of a critique applied to the works and the acts which are the signs of this act of existing” (Ricœur 1974: 17). In short, the shattered cogito – cogito brisé – can only be rediscovered “in the documents of its life” or recaptured “only in the expressions of life that objectify it” (Ricœur 1974: 18). This “documentation” or expressivity should now be understood in the broadest possible sense; in an essay that extends his phenomenology of the body, Ricœur goes as far as stressing “the value of sexuality as a language without words, as an organ of mutual recognition and personalization – in brief, as expression” (Ricœur 1964: 135).The “phenomenology of being able” thereby brings together those “various heritages” and “multicornered debates” of existential philosophy, reflexive philosophy, phenomenology, and hermeneutics Ricœur himself acknowledged to organically relating to (Ricœur 1995c: 445). To restate, in his later philosophy (more properly: in and after Time and Narrative, 1983–1985) Ricœur emphasizes the role of the active “reader,” or the role of an experiencing subject, who interprets his own being while expressing it in action – actions are “interpreters of conduct” – and consequently considers the subject’s own life as a text that opens a redescribed world (Ricœur 1991: 194–195, 1992: 152–163). The early analyses of human action in Freedom and Nature – as a subject’s action “in the world ‘through’ his body” (Ricœur 1966: 226) – are then reformulated in Ricœur’s later philosophy. “It is to the extent to which the entire world [of action] is a vast extension of our body as pure fact that it is itself the terminus of our consent” (Ricœur 1966: 343), Ricœur argues in his early work, implying that selfrecognition and self-understanding become possible only if one acts and brings his or her reappropriable “project” to the world while confronting it at the same time. According to Ricœur, the world is presented to a human subject “as horizon, as theater, and as a matter of my actions” (Ricœur 1966: 212). Such phrasing may give an indication why Ricœur’s phenomenology of active and embodied consciousness has also been seen valuable for enactivism that stresses active interaction between the embodied mind and the world (Dierckxsens 2018). The unavoidable difficulty nevertheless is that the world itself – as well as life – is not dependent on volitional acts whereas it presents itself as the involuntary that is indicated by 197

Timo Helenius

the experience of at once intellectual and muscular effort and consent. The shattered cogito seeks to realize and express its self amid the involuntary: “Moving and consenting means confronting reality with the entire body to seek expression and realization in it; [. . .] this is an engagement in being” (Ricœur 1966: 345). A subject gains a notion of him or herself as a capable person only in this unceasing and existentially grounding dialectic of effort and consent that is manifested in physical and practical action and then re-appropriated. My self is only available to me in the continuous reflective “reading” of it in a concrete, embodied, and objectified situation that I will have to confront. In sum, Ricœur’s hermeneutic and reflexive phenomenology of action/agency stresses the existential need to “affirm myself in my acts,” as it is the case that “I project my own self into the action to be done” (Ricœur 1966: 59).A “project,” to be clear, is nevertheless a mere intention of action whereas a practical act authenticates that intention, testing and verifying it in the corporeal world, thus taking on risk as well as choosing one’s own self. A practical act includes muscular movement, although it is difficult to describe how the will and the bodily movements are intertwined; they are chiastic. It thereby becomes clear that “the acid test of a philosophy of the will is indisputably the problem of muscular effort” (Ricœur 1966: 308). In short, “I ‘recognize’ the empty intention in the full act” (Ricœur 1966: 206). The step from hesitation to authentic choice, or from imposed existence to self-recognized human existence, is made in practical action. “I can” is an empty notion if not made manifest in action in ways that retains a connection between the subject and its object(s) by binding them together in a dynamic and tensional polarity. This results in the need to appropriate the objectified self in and through the action the self was capable of doing. Put differently, the self finds the “I” as the locus of personal identity that in itself is achievable only in reappropriation. Such reappropriable projecting implies that before action the “I” already commits itself – il s’engage as Ricœur puts it in Freedom and Nature – as well as binds itself (il se lie) to the object that results from the action to be done. The reflective moment relating to self-consciousness is crucial here as it exposes “the conditions of existence of a responsible subject” (Ricœur 1967: 8); it distinguishes the realm of human endeavors from those that remain unaware or senseless of the commitment and responsibility that go along with agency. As Ricœur puts it in “Philosophy of Will and Action,” there are in the world physical movements; we can speak of animal and human behavior; but the specificity of the human act can only be reached through a reflective method, in which a human being recognized himself as the author of his acts. (Ricœur 1967: 8–9) Actions stand for the subject that has a chance of beginning to understand his or her existence by reflectively reappropriating the actions in the whole context of his or her life; reflection is thereby “a reappropriation of our effort to exist” (Ricœur 1970: 45), that is, a recovering of the notion of “I” as the subject of all my actions. “The act of existing, the positing of the self,” Ricœur claims, is recovered, re-appropriated “in all the density of its works.” The positing of the self is therefore not “given” whereas “it is a [reflective] task.” Put differently, I do not understand myself from the beginning but only gradually begin to understand myself as a living, acting, and also suffering subject. We thereby arrive at the theme of “pre-reflexive affirmation” (Ricœur 1967: 19) that constitutes the most fundamental thought-action in the sense that all human action or

198

Paul Ricœur

meaningful action, including reflection, is grounded on this pre-reflexive imputation. This comment aims in a way to summarize Ricœur’s phenomenology of action as well as explain how it ultimately remained distinct from its Husserlian ground. In its reliance on Jean Nabert’s reflexive philosophy, Ricœur’s phenomenology contested the Husserlian notion of the self-positing subject or the ego cogito that would be immediately and transparently present to itself. This crucial observation also provides us with an explanation why Ricœur’s phenomenology turned out to be a hermeneutics; it contained a hermeneutical seed in itself from the very beginning. Put differently, the reflective self in search of itself can only find itself as a trace in its acts; “I affirm myself in my acts” (Ricœur 1966: 59).These actions that pertain to the physical but always also to the ethico-political, social, linguistic, and cultural world – which precedes the self as already there – rely, however, on an act that can be considered to provide the fundamental ground for the whole of Ricœur’s anthropology and also to that of action. I am thinking about Nabert’s notion of the originary affirmation, or, the originating pre-reflexive act of consciousness that is clearly reiterated in Ricœur’s later texts, such as Oneself as Another, that discuss “ascribing an action to an agent” (Ricœur 1992: 89) and, in particular, self-ascription. Insofar as we are after Ricœur’s philosophy of action and agency, we cannot set this originating pre-reflexive act aside that maintains and upholds the conditions of a fragile unity at the core of the shattered and self-narrating self. The immediately given or self-transparent Cartesian cogito is inconceivable as this cogito is de facto in a state of primordial conflict. The “living tension” between the infinitude of discourse and the finitude of perspective, the human disproportion as non-coincidence of self to self, makes a human being “a fragile mediation,” and thus also a task for himself (Ricœur 1986a: 140–141). Such task of retrieving the cogito calls for reflection, but this reflection, in turn, is based on an originating affirmation, the “vehemence of Yes” as Ricœur calls it in Fallible Man (1960). Although the originating affirmation is only the beginning of the dialectics of “existential difference” (Ricœur 1986a: 135–136) and the fragile synthesis of the self, it constitutes the most fundamental aspect of the “I can” in that it denotes the power of self-affirmation that, even as thoroughly fragile, is always at the bottom of human action as its condition: “We can understand reflexive judgment precisely by starting out with this prereflexive imputation of myself in my projects” (Ricœur 1966: 60).This fundamental, albeit always “mute,” act of “I can” is never directly accessible, and never submits to intellectual insight. As such, it nevertheless provides the basic condition for a unified experience that itself always remains in the making: It is clear that the entire initial implication of myself is not a conscious relation or an observation. I behave actively in relation to myself, I determine myself. Once again French usage throws light on the situation: to determine my conduct is to determine myself – se determiner. Prereflexive self-imputation is active not observational. (Ricœur 1966: 59) As such active self-imputation, the pre-reflexive aspect is nevertheless never detached from the reflectible acts of consciousness. This implies and reaffirms that there is a duality at the very root of cogito brisé; the self is always in the duality of productively projecting itself and finding itself as projected. We may also understand this duality in terms of the noeticnoematic consciousness; as the noema pertains to the noesis, the projected self pertains to the projecting subject. In other words, the projecting ego is only logically prior to the projected self as the projected self relies on the activity of the projecting self.

199

Timo Helenius

Ricœur’s hermeneutics of action In this section we will expand the exploration of Ricœur’s phenomenology of being able and relate it to his hermeneutics of language and action. In his Kluge Prize acceptance speech, given at the end of his life, Ricœur listed constitutive human powers. These are one’s temporal personal identity, that is, personal history, as well as the capacities to say, to act, and the preeminent capacity to recount. In a similar manner as in Oneself as Another and The Course of Recognition, Ricœur stated that these basic capacities or powers constitute “the primary foundation of humanity” as distinct from everything nonhuman. Saying, acting, and recounting can be understood, respectively, as the abilities to spontaneously produce (1) a reasoned discourse, (2) events in society and in nature, and (3) life narratives that have led up to self-identity. To these three it is possible to add the equally constitutive moral capacities of imputability and promising (Ricœur 2010: 22–23. Cf. Ricœur 1992: 22; Ricœur 2005: 89–110, 127–134). At the level of their execution, however, all these constitutive powers necessarily rely on a cultural system that utilizes symbolic communication. Most importantly, this also pertains to action as Ricœur points out in his Lectures on Ideology and Utopia (1986): We have to articulate our social existence in the same way that we have to articulate our perceptual existence. Just as models in scientific language allow us to see how things look, allow us to see things as this or that, in the same way our social templates articulate our social roles, articulate our position in society as this or that. And perhaps it is not possible to go behind or below this primitive structuration. The very flexibility of our biological existence makes necessary another kind of informational system, the cultural system. Because we have no genetic system of information for human behavior, we need a cultural system. No culture exists without such a system. The hypothesis, therefore, is that where human beings exist, a nonsymbolic mode of existence, and even less, a nonsymbolic kind of action, can no longer obtain. Action is immediately ruled by cultural patterns which provide templates or blueprints for the organization of social and psychological processes, perhaps just as genetic codes – I am not certain – provide such templates for the organization of organic processes. (Ricœur 1986b: 11–12) In sum, Ricœur argues both that nonsymbolic action is not human action, and that the cultural system, which patterns and guides the symbolically mediated being-here, is necessitated by the fact that life itself – and as a result also its social forms – would otherwise remain obscure and unorganized. Life and action gain meaning only if structured cultural-symbolically. This view relies on Ricœur’s earlier conviction that discourse itself is action through which the active and semantically innovative human being expresses itself (while at the same time covering it up by the very linguistic expression): “It is first or all and always in language that all ontic or ontological understanding arrives at its expression” (Ricœur 1974: 11). Ricœur’s “semantics of shown-yet-concealed” or “the semantics of multivocal expressions” (Ricœur 1974: 12) thereby focuses on paying philosophical attention to linguistic expressions as action-expressions or actively produced symbols of the human being. Furthermore, these symbols or innovative expressions of existence call for hermeneutic activity aimed at disclosing that existence by freeing it in understanding: “Interpretation, we will say, is the work of thought which consists in deciphering the hidden meaning in the apparent meaning, in unfolding the levels of meaning implied in the literal meaning” (Ricœur 1974: 12). 200

Paul Ricœur

Ricœur’s hermeneutic phenomenology, in other words, “starts from the symbols, and endeavors to promote the meaning, to form it, by a creative interpretation” (Ricœur 1969: 355). Such interpretation does not start in a vacuum but in medias res: “For it, the first task is not to begin but, from the midst of speech, to remember itself again; to remember with a view to beginning” (Ricœur 1969: 348–349). Action and language are therefore tied in with each other in that the human capacity to act is made manifest in discursive action that invites or “donates” (donner)2 further thought-action that, in turn, exposes the possibility of its originating ground. In discursive action an idea of reason (logos) encounters the activity of biophysical life (bios) in symbolic expressivity. At this point it should be clear that Ricœur’s interest in meaning ful action splits into two dynamically interrelated areas of examination: meaning as action and action as meaningful. Put differently, “symbolic action” can refer to meaning as action-manifesting and to action as meaning-bearing. What we will be now focusing on is the notion that for Ricœur language and action are distinct and yet interdependent. This is precisely the point Ricœur makes in his trilogy Time and Narrative by introducing the hermeneutic model of threefold mimesis or that of reconfiguration. In short, Ricœur’s main thesis is that “between the activity of narrating a story and the temporal character of human experience there exists a correlation that is not merely accidental but that presents a transcultural form of necessity” (Ricœur 1984: 52). This correlation, according to Ricœur, relates to the threefold mimetic operations of emplotment – as a plot is “an imitation of action” (Ricœur 1984: 54) – by which he means the prefigurative (mimesis1), configurative (mimesis2), and refigurative activities (mimesis3) that constitute “the entire arc of operations by which practical experience provides itself with works, authors, and readers” (Ricœur 1984: 53). Ricœur’s attention is specifically on the concrete activity “by which the textual configuration mediates between the prefiguration of the practical field [of experience and action] and its refiguration through the reception of the work” (Ricœur 1984: 53). The broad scope of meaning Ricœur gives to “authors,” “texts” or “works,” and “readers” is the key to understand that his exploration concerns the interweaved duality of meaning as action-manifesting and action as meaning-bearing. To briefly explain the importance of each of the three endlessly and circularly repeated moments of figuration – this is the unceasing hermeneutic activity of a self that Ricœur is ultimately after – we will need to add the following: threefold mimesis itself is fully about action. Mimesis1 or prefiguration is about “the preunderstanding of the world of action, its meaningful structures, its symbolic resources, and its temporal character” (Ricœur 1984: 54). As Ricœur puts it, “a semantics of action makes explicit this competence” (Ricœur 1984: 54) that, in turn, exposes the fact that there is a distinction between action and mere physical movement in that actions “imply goals,” “refer to motives,” and “have agents” that “can be held responsible for certain consequences of their actions,” but also that “to act is always to act ‘with’ others” (Ricœur 1984: 55). But this is only the phase of preunderstanding. Mimesis2 or “configuring activity” concerns the “organization of the events,” by which Ricœur means the arranging narrativization or textualization of the world of action. In short, “this configurational act” – which relies on the productive imagination 3 – “consists of ‘grasping together’ the detailed actions or what I have called the story’s incidents” (Ricœur 1984: 66). Due to this schematization, configuration mediates between the “preunderstanding” and the “postunderstanding of the order of action” (Ricœur 1984: 64–65). The mimetic arc is, however, only brought to its fullness in refiguration. Mimesis3, in Ricœur’s words, brings about the world of a living and acting subject as it marks “the intersection of the world configured by the [plot] and the world wherein real action occurs and unfolds its specific temporality” (Ricœur 1984: 71). Meaningful action is only constituted by the “act of reading” (Ricœur 1984: 76) – or the 201

Timo Helenius

fundamentally selective activity in the inexhaustible excess of meaning (Ricœur 1988: 169) – that rests on configuration and, even before that, on prefiguration. As is clear, the term “reading” is now understood to denote the subject’s reconfigurative hermeneutic activity that results in temporal understanding. Moreover, in the circular and continuous activity of “reading,” refiguration (post-understanding) fuses in with prefiguration (preunderstanding) that, as explained, gives grounds for the likewise unceasing activity of configurative emplotment (viz. the moment of objectification or “explanation”). In sum, our concrete action that includes labor in “the world of work”4 (praxis) only makes sense or is meaningful – which is, again, Ricœur’s marker for distinguishing action from mere physical movement or events – if we are at the same time creative “authors” and “readers” (poiesis), or continuously schematizing and interpreting active subjects, that is, agents. This correlation between language and action is further clarified in From Text to Action, a 1986 collection of Ricœur’s essays. As in the title of the book, the main thesis of the relevant essays is directly displayed in titles such as “The Model of the Text: Meaningful Action Considered as a Text” (1971). Opening his examination from the side of speech acts, “such as we find it in Searle and Austin,” Ricœur argues that their theory offers a model to understand the certain “intentional exteriorization” or materialization that takes place in a shift from a locutionary act to that of an illocutionary act, and furthermore to a perlocutionary act (Ricœur 1991: 146–147). But this movement from language to action is not a one-way street. Ricœur’s point is that all human action can be conceived as meaningful when it meets the condition of objectification that he correlates with “the fixing that occurs in writing” (Ricœur 1991: 151) – this enables the readability of action. Like discourse that, according to Ricœur, is (1) situated as an “instant of discourse,” (2) is self-referential, (3) actualizes the symbolic function of language, and in which (4) the exchange of messages, or communication, takes place, human action is (a) an event that leaves a “trace” or a record, (b) ascribable to an agent or agents, (c) can have (symbolic) importance beyond mere action-event and the social conditions of its production, and (d) is “open” to constant reinterpretation and reevaluation at the level of human praxis: “Human action, too, is opened to anybody who can read” (Ricœur 1991: 150–156). In short, an action is an appropriable Aus-sage, a “speaking out,” or an utterance of a human agent (Ricœur 1991: 146–147). The implied idea, reaffirming the need of approaching action hermeneutically, is that Ricœur finds there to be “an interesting link between the specific plurivocity of the text and the analogical plurivocity of human action” (Ricœur 1991: 160). Referring to Anscombe’s Intention and Melden’s Free Action, Ricœur thereby also makes the point that reason-giving and explaining action amounts to “an expression, or a phrase, that allows us to consider action as this or that” (Ricœur 1991: 160).All in all, the two distinct realms of text and action converge by virtue of the fact that both can be understood. Not merely a derived notion but the nexus of Ricœur’s thought opens up from this understandability; his concern is in the hermeneutical subject aiming to understand him or herself. Action plays a crucial role in this as, according to Ricœur, a subject understands him or herself as an agent in and through the action that arises from the motivational basis that is both construed and “distanced” in that same event of action: What seems to legitimate this extension of understanding the meaning of a text to understanding the meaning of an action is that in arguing about the meaning of an action I put my wants and my beliefs at a distance and submit them to a concrete dialectic of confrontation with opposite [sociocultural] points of view. This way of putting my action at a distance in order to make sense of my own motives paves the way for the kind 202

Paul Ricœur

of distanciation which occurs with what we called the social inscription of human action and to which we applied the metaphor of the “record.” The same actions that may be put into “records” and henceforth “recorded” may also be explained in different ways according to the plurivocity of the arguments applied to their motivational background. (Ricœur 1991: 161) All human action, in other words, becomes meaningful – as well as explainable in some manner – in objectifying the acting person’s wants and beliefs and giving them a shape of a “record” that can then be explained (in a sociocultural sphere) as a correlate of this motivational basis: “the texture of action is transposed into a cultural text” (Ricœur 1991: 195). Through this “putting at distance” as the very “fixing that occurs in writing,” and its corresponding modes of explaining, action also enables self-understanding. Ricœur maintains, in other words, that appropriation is self-interpretation, or understanding one’s self better through physical action that is hermeneutically ascribed and “explained” to one’s own self. Ricœur’s slightly later essay “Explanation and Understanding: On Some Remarkable Connections between the Theory of Texts, Action Theory, and the Theory of History” (1977) capitalizes on this discussion by further clarifying the relation between the agent and action. In short, Ricœur adopts the full-blooded approach to agency and action, that is, takes into account the agent of action as a situated self who desires, wants, acts, and interprets his or her own being. In spite of such emphasis on the acting self, Ricœur’s theory nevertheless oscillates between event-causation (viz. biophysical necessity) and agent-causation (viz. intentionality and volition) – or takes into account them both as Ricœur also forcefully argues in his dialogical exploration of the philosophy of mind in What Makes Us Think? (1998). From the vantage point of human experience, or that of phenomenology, the focus of interest is in a subject “who is ‘mental’ and ‘corporeal’” (Ricœur and Changeaux 2002: 28, 128–129, 162–178); the essay on explanation and understanding grounds this claim. Showing the limits of the Humean notion of causation by critically assessing the work of later Wittgenstein, Anscombe, and Austin, the main point Ricœur wants to make against “semantic and epistemological dualism” (Ricœur 1991: 135) – using both von Wright’s notion of a “closed system” and Danto’s notion of “basic action” as his guide – is that the theories whose outcome is dichotomous should be discarded in favor of a theory that espouses dialectical dynamism. It only produces an illusory dissolution of the philosophical problem of action to dichotomously set the discourse of explanation, causality, and factuality against that of understanding, motives, intentions, and reasons. As bios and logos are chiastically intertwined in human experience, human action too pertains to the two realms at once: “Human being is as it is precisely because it belongs both to the domain of causation and to that of motivation, hence to explanation and to understanding” (Ricœur 1991: 135). In other words, the proposed dynamism is enriching also in the sense that understanding “envelops explanation” whereas explanation “develops understanding analytically” (Ricœur 1991: 142). Put together, the theories of von Wright and Danto show that the relation between doing something immediately (basic action) and making something happen mediately (by doing something that is in my power) follows the lines of the causal analysis of closed systems. . . . these two elements – the course of things and human action – are intertwined in the notion of intervention in the course of things. (Ricœur 1991: 136–137. Cf. Ricœur 1977a: 26–30, 104–108) 203

Timo Helenius

According to Ricœur, this notion of intervention allows for a correlating notion of an agent, thereby bringing the agent and action theoretically together. Ricœur’s response to Hume and the theorists following his model of universal causation is, therefore, that “human action and physical causation are too interconnected in this entirely basic experience of the intervention of an agent in the course of things for one to abstract from the first term and raise the second to an absolute” (Ricœur 1991: 137). Even though Ricœur thereby most certainly harkens back to Aristotle’s view that reasons are causes, he also maintains the view that the agent is also affected by physical causality – but not in any deterministic sense that would preclude the possibility of intervention.5 A semantic approach to the question of causation reveals that the agent and the action-event are interrelated. It is on these grounds that Ricœur offers an extended discussion of action and agency in one of his most read works with which we opened this essay and with which we will also close it. In Oneself as Another (1990), Ricœur elaborates his ethically, practically, and ontologically concerned hermeneutic theory of the self in form of a theory of narrative identity. This manifold analysis is later echoed in The Course of Recognition and in Ricœur’s Kluge Prize acceptance speech that both enlisted the human powers Ricœur holds to be constitutive: the capacity to have a personal identity that is reconfigured in saying, acting, recounting, and in moral ascription. As Ricœur explains, the theme of narrative identity is approached in Oneself as Another by manner of four interlinked questions that nevertheless also demarcate distinct realms of philosophical exploration pertaining to the self. In spite of such willful fragmentation, however, it is clear from the beginning that for Ricœur the self is inherently active – ultimately so in its foundational act of self-attestation. This notion of the acting self that is “implied reflexively in its operations” is the thread that holds the work together through its distinct realms of analysis; “these studies together have as their thematic unity human action.”6 In sum, known for its theory of narrative identity, Oneself as Another unfolds as a sophisticated reflection on the interlink between action and agency. Ricœur begins his analyses of the self in Oneself as Another by first addressing the question “who is speaking?” – or, rather, “of whom does one speak in designating persons” – from “the twofold perspective of semantics and pragmatics” as discussed in the Analytic philosophy of language (Ricœur 1992: 16–17). As Ricœur’s discussion particularly draws from those of Strawson, Austin, and Searle – leading him again to address the speech act theory – it is understandable that the second question Oneself as Another deals with is agency. In short, “it is in speech acts that the agent of action designates himself or herself as the one who is acting” (Ricœur 1992: 17). Taking this particular question, “who is acting?,” as another way to approach the question of the ultimately hermeneutic self, Ricœur nevertheless again follows the resources provided by the Analytic philosophy of action in its asking about “what” and “why.” The hypothesis is that taking a detour through these two questions will prove to be helpful in terms of pointing out the need for the “who” of action/agency. Without the need to repeat Ricœur’s criticism of Anscombe and Davidson, it can be reminded that Ricœur wishes to draw attention to the agent “as the possessor of his or her action” (Ricœur 1992: 85). Aspiring to make a move from the semantics of action to the pragmatics of the agent, Ricœur thereby continues by discussing attribution as ascription, or “ascribing an action to an agent” (Ricœur 1992: 89), in the manner Strawson does in Individuals. In Ricœur’s hands, however, the issue is pushed toward the direction of Danto and von Wright, meaning that the discussion concerning “intervention” in Time and Narrative 1 and From Text to Action is re-enacted in Oneself as Another in a novel manner. At this point we are able to proceed with the two remaining questions Ricœur proposed to study in Oneself as Another. The following will demonstrate that in spite of his 204

Paul Ricœur

extended discussions on the relevant viewpoints and theories addressed in Analytic philosophy, Ricœur’s own philosophical pursuit is consistently phenomenological and hermeneutical. First, in spite of addressing analytical discussions and concerns, Ricœur nevertheless ultimately understands the semantics of action to clarify meaningful action. His conviction is that action, in distinction to mere physical movement, can and must be meaningful or meaning-bearing. In other words, Ricœur’s linguistic analyses as well as his relating hermeneutical dialectics of explanation-understanding most certainly were worthwhile pursuits themselves, but they ultimately served his “attempts to make the field of practice, and in general human action, the privileged place” of such hermeneutics (Ricœur 1995a: 32). Second, this hermeneutic opens a way to pursue an existentially attuned phenomenology, as can be seen in what follows in Oneself as Another. Ricœur’s questions regarding the speaking and acting subject constitute only the first half of the overall arch of the discussion that is aimed at exposing – or at least drawing the reader close to recognizing – the ethical, practical, and ontological reality of an “acting and suffering subject” by means of elaborating a “phenomenology of ‘I can’” (Ricœur 1992: 112). The third and fourth questions, “who is recounting about himself or herself?” and “who is the moral subject of imputation?,” (Ricœur 1992: 16) are, respectively, tied in with the questions of temporal personal identity and ethical ascription. Here the notion of meaningful action, now standing for an autonomous and yet situated agent, reappears as relevant. Defining narrative identity or “self hood” as ipseity or ipse-identity (in distinction from idem-identity that understands identity as either numerical or qualitative sameness), the discussion concerning the agent – or, rather, the “acting and suffering subject” as a phronimos – quickly expands into a phenomenologically attuned ethics concerning self-ascription or holding oneself “accountable for my actions before another” (Ricœur 1992: 165, 178).This move – that echoes Ricœur’s earlier convictions that “the description of action is the base on which an ethics could be construed” and “an ethics is also a politics” (Ricœur 1977a: 18–19) – would be quite extraordinary in the field of Analytic philosophy whereas for Ricœur it organically grows out from the preceding studies in action and agency. In terms of meaningful action, Ricœur holds, human agency comes explicit in the ethical intention of “aiming at the ‘good life’ with and for others, in just institutions” (Ricœur 1992: 172). This brief summarizing phrase of the ethico-politically and practically accentuating response to the question of “who,” which culminates the discussion and leads it to the relating ontological implications Ricœur addresses in the end of this work, conceals the fact that the discussion about such imputation itself is materially substantiating in terms of Ricœur’s philosophical work well beyond Oneself as Another.7 In form of a brief conclusion we can therefore state that Ricœur’s “bottom line” of “phenomenology of being able” is thereby merely a reiteration of his earlier aim of revealing phenomenologically “the fundamental human possibilities” (Ricœur 1966: 3; Ricœur and Kearney 2004: 167). In Ricœur’s thought, finding ourselves capable of being ethically responsible agents – at which level the self finds itself as an active ethico-political co-citizen – relies on the continuously exercised and tested capacity to act (physically, linguistically, mnemonically, narratively, and so on) that, in turn, relies on the reflective, self-ascribing, and embodied cogito, and, most fundamentally, on the pre-reflexive imputation. All these together constitute the “I can” Ricœur sought to clarify: It is necessary to uncover the most primitive possibility of myself, which I launch within myself in making up my mind. This is the easiest analysis because it still refers to our analysis of the project. In effect, for a responsible being, that is, a being who commits 205

Timo Helenius

himself in the project of an action which he at the same time recognizes as his, determining oneself is still one with determining his gesture in the world. We can thus search out what possibility of myself is simultaneous with the possibility of action opened up by the project. (Ricœur 1966: 63) As Ricœur’s early statement makes evident, his philosophy of a “living, acting, and suffering subject” was from its incipient stage asking about meaningful action, or, action as meaningful in pursuing a phenomenology of “I can” (Ricœur 1992: 112). Being invested in action bears significance as my action, recognized as pertaining to my self from the very beginning. But the wording here is intentionally indeterminate; Ricœur’s reference to the “easiest analysis” is meant to caution his readers about the fact that a pure analysis of the agent is never fully achievable. “To understand freedom is to understand precisely the history which we have held in suspension” (Ricœur 1966: 65). The self is never only for itself whereas it is always also with others in its situatedness that it has to unceasingly interpret.

Notes 1 The crucial role that the notion of responsibility has for Ricœur’s thought is also evident in his discussions concerning Analytic philosophy. Cf. Dierckxsens (2017: 578–584). 2 Ricœur (1969: 348, 355). The essence of Ricœur’s “wager” is captured with the phrase le symbole donne à penser. In short, the symbol gives or donates a “gift” (don) for thought by giving rise to thinking. 3 This context does not, unfortunately, allow us to address Ricœur’s philosophy of imagination that plays an inherently important role in his philosophy of action. For a brief commentary on the imagination, cf. Ricœur’s essay “Imagination in Discourse and Action.” Ricœur (1991: 168–187). 4 Ricœur’s hermeneutic and phenomenological theory of action extends to and covers the questions of human work in texts such as Fallible Man and Freud and Philosophy. For a concise discussion of praxis as labor in reference to Ricœur’s History and Truth, cf. Jervolino (1996: 70–76). 5 It should be noted that the term “cause” has a different signification for Hume and Aristotle. Cf. Hutto (2013: 63–64). 6 Ricœur (1992: 18–19). For a clear and concise exposition of the interrelatedness between narrated action and narrative identity, cf. Pellauer (2015). 7 Cf. Ricœur’s “little ethics” in studies seven, eight, and nine of Oneself as Another; see Ricœur (1992: 169–296).

References Davidson, D. (2001) Essays on Actions and Events: Philosophical Essays, Oxford: Clarendon Press. Dierckxsens, G. (2017) “Responsibility and the Physical Body: Paul Ricœur on Analytical Philosophy of Language, Cognitive Science, and the Task of Phenomenological Hermeneutics,” Philosophy Today 61/3, 573–593. Dierckxsens, G. (2018) “Ricœur’s Take on Embodied Cognition and Imagination: Enactivism in Light of Freedom and Nature,” in S. Davidson (ed.), A Companion to Ricœur’s Freedom and Nature, Lanham, Boulder, New York, London: Lexington Books, pp. 191–205. Hutto, D. (2013) “Still a Cause for Concern: Reasons, Causes, and Explanations,” in G. D’Oro and C. Sandis (eds.), Reasons and Causes: Causalism and Anti-Causalism in the Philosophy of Action, New York: Palgrave Macmillan, pp. 49–74. Jervolino, D. (1996) “Gadamer and Ricœur on the Hermeneutics of Praxis,” in R. Kearney (ed.), Paul Ricœur: The Hermeneutics of Action, London, Thousand Oaks, CA, New Delhi: SAGE Publications, pp. 63–79. Pellauer, D. (2015) “Narrated Action Grounds Narrative Identity,” in R.W.H. Savage (ed.), Paul Ricœur in the Age of Hermeneutical Reason: Poetics, Praxis, and Critique, Lanham, MD, Boulder, CO, New York, London: Lexington Books, pp. 69–83.

206

Paul Ricœur Ricœur, P. (1964) “Dimensions of Sexuality: Wonder, Eroticism, and Enigma,” Cross Currents 14/2, 133–141. Ricœur, P. (1966) Freedom and Nature: The Voluntary and the Involuntary, trans. by E.V. Kohák, Evanston, IL: Northwestern University Press. Ricœur, P. (1967) “Philosophy of Will and Action,” in E.W. Straus and R.M. Griffith (eds.), Phenomenology of Will and Action, Pittsburgh, PA: Duquesne University Press, pp. 7–33. Ricœur, P. (1969) The Symbolism of Evil, trans by E. Buchanan, Boston, MA: Beacon Press. Ricœur, P. (1970) Freud and Philosophy: An Essay on Interpretation, trans by D. Savage, New Haven, CT and London: Yale University Press. Ricœur, P. (1974) The Conflict of Interpretations: Essays in Hermeneutics, Evanston, IL: Northwestern University Press. Ricœur, P. (1977a) “Le discours de l’action,” in D. Tiffeneau (ed.), La semantique de l’action, Paris: Éditions du centre national de la recherche scientifique, pp. 1–137. Ricœur, P. (1977b) The Rule of Metaphor, trans by R. Czerny, K. McLaughlin and J. Costello, Toronto, Buffalo, London: University of Toronto Press. Ricœur, P. (1984) Time and Narrative 1, trans by K. McLaughlin and D. Pellauer, Chicago, IL: University of Chicago. Ricœur, P. (1986a) Fallible Man, trans by C.A. Kelbley, New York: Fordham University Press. Ricœur, P. (1986b) Lectures on Ideology and Utopia, New York: Columbia University Press. Ricœur, P. (1988) Time and Narrative 3, trans by K. Blamey and D. Pellauer, Chicago, IL and London: The University of Chicago Press. Ricœur, P. (1991) From Text to Action: Essays in Hermeneutics II, trans by K. Blamey and J.B. Thompson, Evanston, IL: Northwestern University Press. Ricœur, P. (1992) Oneself as Another, trans by K. Blamey, Chicago, IL and London: The University of Chicago Press. Ricœur, P. (1995a) “Intellectual Autobiography,” in L.E. Hahn (ed.), The Philosophy of Paul Ricœur, trans by K. Blamey, Chicago and La Salle, IL: Open Court, pp. 3–53. Ricœur, P. (1995b) “Reply to Charles E. Reagan,” in L.E. Hahn (ed.), The Philosophy of Paul Ricœur, Chicago and La Salle, IL: Open Court, pp. 346–347. Ricœur, P. (1995c) “Reply to David Stewart,” in L.E. Hahn (ed.), The Philosophy of Paul Ricœur, Chicago and La Salle, IL: Open Court, pp. 443–449. Ricœur, P. (2005) The Course of Recognition, trans by D. Pellauer, Cambridge, MA and London: Harvard University Press. Ricœur, P. (2010) “Asserting Personal Capacities and Pleading for Mutual Recognition (the Kluge Prize Acceptance Speech),” in B. Treanor and H.I. Venema (eds.), A Passion for the Possible: Thinking with Paul Ricœur, New York: Fordham University Press, pp. 22–26. Ricœur, P. and Changeaux, J.-P. (2002) What Makes Us Think?, trans by M.B. DeBevoise, Princeton, NJ and Oxford: Princeton University Press. Ricœur, P. and Kearney, R. (2004) “Part II: Interviews with Paul Ricœur,” in R. Kearney (ed.) On Paul Ricœur: the Owl of Minerva, Burligton, VT: Ashgate, pp. 115–169.

207

16 OPERARI SEQUITUR ESSE Hermann Schmitz’s attitudinal theory of agency, freedom, and responsibility Henning Nörenberg

According to a widespread account of human agency, the usual way to understand an action is to know what caused the conduct in question not in a purely physical sense, but rather in terms of motivation or reason (cf. Gallagher and Zahavi 2008: 157). In this perspective, a satisfactory account of why I am walking to the fridge would not entail many details about the connections between my neuronal and muscle activities. Instead, one would expect information about specific connections between some of my beliefs and desires on the basis of which, for instance, my intention of having a drink is formed: I φ in order to Γ. In this sense, actions are intentional. In many cases, however, I act before I have a chance to deliberately decide to act: If, upon approaching the bus stop, I see the bus pulling away, I might start running to catch it. If you stop me and ask, ‘Are you trying to catch the bus?’ my answer would be yes, that was my intention. (Gallagher and Zahavi 2008: 159) That is, this action, too, is intentional, although the intention is nothing separate from the action, but rather “in” it (cf. Searle 1983: 88) in the form of a pre-reflective awareness of what one is actually doing (cf. Gallagher and Zahavi 2008: 158, 160). Sometimes, however, matters are more complicated. Consider the following case: on September 8, 2015 at the Röszke collection point in Hungary, camerawoman Petra László was filmed kicking and tripping migrants who were fleeing the police. The video attracted widespread media attention. The majority of the viewers had no difficulties in identifying a blameworthy conduct and reacted with indignation. The incident eventually resulted in a charge with antisocial, violent behavior and caused László’s employer to fire her. Many viewers of the video in question ascribed a racist motivation to her. In a public apology, László acknowledged that what she did was “bad” (cf. Srinivas 2015). That is, in a certain respect, she had no difficulties in identifying a blameworthy conduct either. However, she denied any racist motivation and described her conduct as a panic reaction in the face of a large number of migrants running in her direction rather than a full-blown intentional action: “I started to panic and as I re-watch the film, it seems as if it was not even me” (cf. ibid.). In a later interview, however, she stated she was “trying to help the police.”1 Apart from the 208

Hermann Schmitz

obvious attempt to regain interpretational sovereignty, the different accounts László tries to give (no genuine intention, just panic reaction; helping the police) may also point to a serious difficulty with making unambiguous sense of what she actually did in that moment. In this and many other, less extreme cases, there seems to be a difficulty with understanding an action, a difficulty that is absent in cases such as trying to catch the bus. It seems that this difficulty cannot be resolved in terms of a more fine-grained account of intentional action, e.g. by distinguishing a pre-reflective sense of what one is doing from unintentional consequences of this doing. Rather, László’s remark that “it seems as it was not even me,” her insistence that “I’m not a heartless, child-kicking racist” (cf. Holdsworth 2015) as well as the accusations in social media she attempts to reject with her apology revolve around another way of understanding an action. This understanding sets out with a question of the following kind: “What kind of person must I be that I could actually do this?” (cf. Scheler 1921: 24; Schmitz 1973: 585). Rather than on particular intentions, the focus of this way of understanding an action is on the person, the agent or, more precisely, on the relevant disposition, stance, or fundamental attitude with which the agents act. An interesting version of such an attitudinal theory of agency can be found in the work of the German phenomenologist Hermann Schmitz. Schmitz belongs to the postwargeneration of philosophers such as Jürgen Habermas and Karl-Otto Apel (both of whom he studied with under Erich Rothacker) that was highly motivated to critically understand what had happened before and under Nazi rule. The focus of Schmitz’s work, however, is not so much on failures of public reason, but on questions of personhood and bodily existence. According to Schmitz, these latter aspects are crucially involved in shaping the background against which something emerges as a good reason or as demanding one’s respect. Schmitz’s theory of attitude (Gesinnung)2 as the fundamental way in which the individual engages with the world and which cannot be exhausted in terms of intentions is part of his comprehensive investigations concerning personhood, subjectivity, and human freedom. A first version of that theory was formulated in Schmitz’s 1973 book Der Rechtsraum (literally: The Space of Law), but the argument has been slightly revised in his 2007 book Freiheit (Freedom). A condensed English version of the argument can be found in the final chapter of New Phenomenology (Schmitz 2019: 123–133). It does not only draw on Heideggerian phenomenology but also on arguments from German Idealism, existential philosophy, and P.F. Strawson’s descriptive metaphysics. According to Schmitz’s theory, a person’s actions are grounded in that person’s character (operari sequitur esse), agency is conceived in terms of “self-efficacy” (Selbstwirksamkeit), and human freedom and responsibility have their appropriate place in the attitude rather than in willing or intending. In the following, I will flesh out these claims by sketching Schmitz’s attitudinal theory of agency. The first section aims to delineate the concept of self-efficacy in the context of human agency. Then, I introduce the notion of “subjective factuality” (subjektive Tatsächlichkeit) which is crucial for Schmitz’s argument that the attitude of a person satisfies the conceptual conditions for self-efficacy. Finally, I present and discuss the notion of attitude and attempt to clarify the sense in which the attitude of a person is supposed to be a form of self-efficacy.

Agency as self-efficacy: conceptual considerations Schmitz develops his notion of agency in the context of his theory of human freedom and moral responsibility. Freedom, in this sense, is a necessary as well as sufficient, non-trivial 3 condition of moral responsibility (Schmitz 2007: 22). The most central trait of freedom is what Schmitz calls “independent initiative” (unabhängige Initiative), that is, the capacity to 209

Henning Nörenberg

bring or to refrain from bringing a state of affairs into being by oneself, to make or to abstain from making it a fact (Schmitz 2007: 29). The term “to bring something into being” intends a specific form of causation and covers a broad spectrum from moving one’s finger to realizing a plan (Schmitz 2007: 41). The relevant capacity is independent in as much it is, in a particular sense, not determined by any objective fact, but only effected by itself. The existence of such an initiative is not only relevant to our normal practices of holding each other responsible but also to an even more fundamental condition of human life: the practically lived confidence in there being something that is actually depending on oneself. According to Schmitz, the loss or absence of such confidence would tangibly impair a person’s willingness to make an effort and result in ennui (Schmitz 2007: 23). Before he begins to argue for the existence of a person’s independent initiative, Schmitz suggests a set of conceptual conditions that any candidate for the initiative in question would need to satisfy. To begin with, the independent initiative must carefully be distinguished from other forms of causation: the color of my hair or the size of my nose may be triggering causes of instances of attraction or repulsion in other people, but none of these causes would count as cases of free, independent initiative or agency. That is, being a distinct form of causation, the initiative in question must be related to the agent in a sufficiently intimate manner (cf. Schmitz 2007: 48). Only if this is the case, our practices of holding each other responsible as well as our practically lived confidence that something is up to us are ultimately more than wishful thinking. Whereas authors such as Sartre or Arendt would more or less identify the agent with that initiative (cf. Sartre 1956: 457; Arendt 1998: 177), Schmitz resists that conclusion, since it would imply conceptual problems with regard to one’s responsibility for what one fails to do (Schmitz 2007: 47). More generally and in terms of ontology, he suggests conceiving of the relevant cause as a particular fact rather than a particular object. Causation, according to Schmitz, is a relation between facts. Thus, he looks for a sufficiently intimate property of the agent. According to these conceptual specifications, the independent initiative is formally a fact that is its own cause (causa sui), i.e. a peculiar form of self-efficacy (Schmitz 2007: 48). Schmitz attempts to mount this form as an ontological alternative to determinism on the one hand and indeterminism on the other hand. Schmitz takes determinism to amount to the claim that every non-analytic fact F has a sufficient cause being a fact distinct from F, while indeterminism consists in the claim that there are non-analytic facts that have no cause in another fact (Schmitz 2007: 48–49.). In this respect, self-efficacy means, ontologically, that there is at least one non-analytic fact that is its own cause (Schmitz 2007: 49). Moreover, Schmitz insists that we have experiential access to this form of self-efficacy, which makes it distinct from other forms of causation (Schmitz 2007: 47). Schmitz further delineates the notion of independent initiative as self-efficacy by distinguishing it from related notions such as choosing, the opportunity to do otherwise or control over one’s own behavior. For starters, the independent initiative is distinct from making one’s choice in as much as this would entail confining oneself to an option in the conviction of having more than one option (Schmitz 2007: 33–39). Such a conviction, however, is absent in cases of spontaneous actions that we nevertheless consider to be worthy of moral praise or blame. Schmitz mentions spontaneous acts of bravery and cowardice in dangerous situations, but one might also think of the László case mentioned in the introduction. Cases of negligence, too, are subject to our practices of holding each other responsible although the person held responsible lacked the conviction of having other options in the crucial moment. 210

Hermann Schmitz

Although initiative and choice are distinct from one another, Schmitz also assumes a close connection between them (Schmitz 2007: 23). That is, the conviction of having several options is one of the possible backdrops against which one’s initiative may play out so that in this case such playing out amounts to making one’s choice. In turn, a choice that was essentially determined or sufficiently caused by something else than one’s independent initiative would be anything but a choice (cf. Schmitz 2007: 35). Thus, choosing depends on initiative. Moreover, in as much as one confines oneself to one of several options in the light of reasons, choice and the independent initiative enabling it are also relevant to rational conduct. Another concept that is not identical with but dependent on there being an independent initiative is the actual opportunity to do otherwise: someone has the opportunity to do otherwise, if he or she in a given situation is able to choose independently from real – i.e. not only imaginary – options. And since choosing is not an essential trait of the initiative in question, the opportunity to do otherwise isn’t either (Schmitz 2007: 34ff.). Schmitz even holds that one’s independent initiative does not necessarily amount to control over one’s own behavior. In this sense, even a person obeying to an irresistible drive could be free. In such scenarios, freedom as independent initiative would have to be located in one’s approval, in one’s consent to the inevitable. Schmitz cites the example of an addict giving up the futile fight against his drive and allowing himself to surrender. Pushing the intended notions of freedom and independent initiative to an extreme, Schmitz also considers amor fati as a form of freedom: a falling person is free if she does not resist but resigns herself to her fate (Schmitz 2007: 32–33). With this claim, Schmitz probably parts with the usual accounts of our sense of agency, since the latter seem to adhere to some form of control as a necessary element of agency.

The notion of subjective factuality Having conceived of self-efficacy in terms of there being at least one non-analytic fact that is its own cause and having ruled out notions such as choosing or self-control for explicating the causal power in question, Schmitz needs to show how this concept actually applies to human reality. Therefore, he mounts a complex argument that is supposed to establish that a b c

one’s attitude (Gesinnung) by which one involves oneself with the world is the sufficiently intimate property the having of which is its own cause, precisely such an attitude is the appropriate target of our normal practices of praising and blaming, and there is no reason to deny a certain independence of that attitude.

In short, causal power on one’s own initiative consists in the attitude one invests in one’s affective involvement with the world and, thus, in the way one minds what is going on (cf. Schmitz 2007: 71). The first steps of the argument introduce Schmitz’s notions of subjective factuality (subjektive Tatsächlichkeit), affective involvement (affektives Betroffensein), and attitude (Gesinnung). As these three notions are highly interrelated, it is a matter of taste with which to begin. Usually, Schmitz begins with the relevance of subjective factuality and I will do the same here. Under the notion of subjective factuality Schmitz addresses central insights pertaining to the phenomenology of pre-reflective self-consciousness. Schmitz, however, articulates his notion in critical distance to what he takes to be the implications of more mainstream 211

Henning Nörenberg

concepts of subjectivity and self-consciousness. In his view, subjective factuality is logically as well as experientially prior to the self. But what is subjective factuality supposed to mean? Objective facts such as “It is raining,” “Caesar crossed the Rubicon,” or “Martin loved Hannah” can, in principle, be predicated by anyone who has sufficient knowledge and articulation. In contrast, subjective facts such as “I’m in love with Sally,” “I’m sorry for my mistake,” or “I hope Jim will get by” entail the nuance of a first-personal character, i.e. the “mineness” (Meinhaftigkeit) of what is going on and what could, in principle, be predicated only by the relevant person in their own name (cf. Schmitz 1969: 27–77, 1973: 527 ff., 2007: 55 ff.; Schmitz et al. 2011: 248–249).4 “Mineness” refers to the fact that it is irrevocably oneself and no other who is involved in the fact in question in the stated way. The relevant nuance cannot be substituted by any amount of more detailed objective facts about oneself (exact date and place of birth). None of these facts could add any reference to oneself – while the latter is what the nuance of “mineness” is all about (Schmitz et al. 2011: 248). Thus, subjective facts cannot simply be reduced to neutral or objective ones. Schmitz often illustrates this with the example of one’s own sadness which is supposed to function as a stand-in for any fact that could only be predicated by the respective person in her own name: If, for instance, I sincerely say “I am sad”, this contains more than saying that Hermann Schmitz, disregarding the fact that I am him, is sad, i.e. the actual affectivity and intensity of personal involvement. This surplus persists despite the complete equivalence of the propositional contents of both facts. In both cases, the same person is determined by means of the same fact including the affective involvement that is a necessary part of sadness. Nevertheless, the subjective fact is richer than the neutral one. This difference does not concern the propositional content, but the factuality, which is richer in the case of subjective facts than in the case of the pale objective ones. The difference lies not in a merely private inner mode of presentation as against an external one of objective facts. This is plain to view in that the personal pronoun “I” as well as its equivalents cannot simply be replaced by a proper name, as in the case of neutral facts, if we are reporting on facts of affective involvement. (Schmitz et al. 2011: 249) In contemporary discussions, “mineness” has also been broadly investigated as a primitive form of self-awareness implicated in experience (cf. Zahavi 1999). In Schmitz, this form of self-awareness is called “affective involvement” and is systematically linked with a theory of the feeling body (Schmitz et al. 2011: 248–249). The body, in this sense, is not the physical body, that is, for instance, the body from the point of view of a third person observing my arm movement, but rather the feeling body that is felt in various instances of contraction, constriction, tension, relaxation, pulsation, and expansion. Schmitz argues that the sense of “mineness” – along with the sense of reality – is first and foremost provided by contracting and constricting tendencies such as fright or pain, for they present one’s situation in terms of a narrow here and now which also implicates a “me.” On the other hand, expansive tendencies involved, e.g. in joy entail at least an echo of an constricting tendency in so far as they entail a sense of getting away from the latter (cf. Schmitz et al. 2011: 249–250).

The notion of attitude Affective involvement, Schmitz says, is Janus-faced: on the one hand, it entails the aspect of passive experience, i.e. one’s being touched by or exposedness to what is going on, 212

Hermann Schmitz

e.g. hunger, pain, exuberant joy, or attraction. On the other hand, it also entails the rather active aspect of resonating with, responding to, caring for, engaging with what is going on (Schmitz 2007: 61; Schmitz et al. 2011: 254). Human beings can undergo hunger grumpily, patiently, wailingly, or aggressively (Schmitz 2007: 69). They can engage their pain in a stoic or a whiny manner. They may resonate approvingly with the joy that seizes them or they may not allow themselves to go along with that feeling unselfconsciously (Schmitz et al. 2011: 254). Except for extreme cases, which will be addressed further below, both the passive and the active side cannot be separated from one another, although their connection may be more loose or more tight. Similar observations can be found in the late Husserl’s notes on affectivity (Affektion) that have recently been published: in affectivity, there is not only an impact on the subject; the latter is also provoked to a prompt and more or less spontaneous response in terms of turning toward the relevant issue, to get involved or concerned with it in this or that way (Husserl 2014: 34ff.). In a certain sense, both, impact and response, are one (Husserl 2014: 40). Husserl, too, emphasizes that the response is the way in which “I am active” (Husserl 2014: 41), grounding it in the notion of “interest” (Husserl 2014: 40, 44) which describes the way in which we are “living into the world” (Husserl 2014: 76). More generally, this active aspect of engaging in one’s experiences is not alien to phenomenologists. For instance, Scheler calls that aspect “functional quality of feeling” and illustrates it with the differences between suffering, bearing, or enjoying pain (Scheler 1954: 270). The early Heidegger calls it “enactment sense” (Vollzugssinn) of the relevant phenomenon (Heidegger 1995: 63). Sartre (1956: 456f.) addresses the relevant aspect under the name “projet originel” and argues that the individual’s fundamental project shapes the very facticity of the situation – a thought that, as we will see below, Schmitz seeks to reformulate within his own account (cf. Schmitz 1973: 566–570). In this context, however, one may also think of Arendt’s notion of acting as the way in which we “insert” ourselves into the human world (Arendt 1998: 176f.). In the context of his own investigation, Schmitz calls the active aspect of affective involvement the individual’s “attitude” (Gesinnung) in order to relate the intended phenomenon to what he takes to be our normal sense of responsibility (Schmitz 2007: 69). Simultaneously, he shifts the scope of the phenomenon in question from transcendental or other forms of constitution to a peculiar, experientially accessible form of causation. In short, the attitude is the phenomenon that satisfies the conceptual conditions for an independent initiative outlined earlier. Schmitz argues that our normal practices of praising and blaming ultimately refer to the agent’s attitude rather than to their intentions or decisions. This is most obvious in cases of spontaneous acts of bravery or cowardice as well as in cases of negligence. In all these cases we may admit that the agent was not in the position to carefully reflect on what they should (not) do, although we insist to praise or blame the agent. Schmitz cites the drastic example of a mother nodding off nearby a lake and later finding her child fallen into the water (Schmitz 2007: 34). Strictly speaking, she did nothing wrong, rather she failed to do something, namely attending to her child. Likewise, she did not choose the wrong option, for she did not choose between the alternative to either nod off or be vigilant in order to prevent the child from drowning. But very likely she would blame herself in bitter remorse: “How could I have been so unmindful?” That is, what is blameworthy is her attitude conceived as the way in which she was engaged in the situation. According to Schmitz, we need more than the actual occurrence of the relevant reactive attitudes in us (cf. Strawson 2008) or the intersubjectively projectible patterns of rationality 213

Henning Nörenberg

they are forming (cf. Helm 2017) in order to account for the appropriateness of our practices of blaming and praising in the cases just mentioned. Schmitz uses his phenomenological theory of the intertwinement of passive and active aspects of experience to argue that we are actually right in praising or blaming another person’s or our own attitude. This is because it does make a difference how we are engaging the situation in which we find ourselves – a difference that under certain circumstances can become morally relevant – and because the relevant engaging constitutes a fact sufficiently intimately related to ourselves. A person’s attitude constitutes a considerable part of her character or, in Arendt’s words, who that person is (cf. Arendt 1998: 178). Whether we tend to undergo hunger and pain wailingly or patiently, or whether we are or are not prone to indulge in schadenfreude, for instance, is a matter of our character and, in the given context, it is defining us more intimately than, say, facts about the color of our hair. Other relevant aspects of a person’s character, which under certain circumstances may also compensate, outbalance, or even conceal the attitude, consist in that person’s “countenance” (Fassung) which is related to the ancient notions of ethos and habitus as well as in the person’s bodily dispositions (Schmitz 2007: 113). These and other factors are involved in the formation of intentions and volitional acts such as concentrating one’s attention, mental calculation, or bodily movement (Schmitz 2007: 121). However, Schmitz holds that among these factors, the attitude is most relevant in terms of freedom and responsibility: a person’s causal power, i.e. their power to bring something into being by their own independent initiative, is not so much a matter of intending something, but rather the attitude as the way in which the person is engaged in what affects them (Schmitz 2007: 71). Obviously, this does not exclude that the way in which a person engages the situation could and should be accompanied by her reflection and deliberation. However, it is not necessary that a person can choose her attitude or is in full control of it. While Schmitz acknowledges the kinship between his notion of attitude and Sartre’s projet originel, he hastens to state that he is more skeptical than Sartre regarding the actual purview of self-efficacy and chances to change one’s own attitude overnight (Schmitz 1973: 568–569). This is not to deny that a human subject can reinvent itself, but such a feat seems to involve long and arduous processes of self-formation. This, at least, seems to be the general thrust of related concepts such as the Greek notion of daimon or the “intelligible character” introduced by German Idealists (Schmitz 1973: 581f.). Apart from a few backdoors for choosing one’s own character before birth or in the non-empirical, intelligible realm which would be rather difficult to defend these days, the notions in question amount to the idea that one’s character is by and large inaccessible5 to oneself and is only revealed in one’s actions: Operari sequitur esse, as Schopenhauer quotes Thomas Aquinas (cf. Schopenhauer 1972: 693).

Attitude as a form of self-efficacy However, if a person’s attitude is supposed to be that person’s independent initiative by which she exercises causal power in the world, and if that attitude is by and large inaccessible to the person, in what sense, then, is it supposed to be independent and self-efficacious? The remaining steps of the argument aim to clarify precisely this. The next step is supposed to establish the following: subjective factuality, i.e. the subjectivity of facts relevant to the first-personal character of experience, is an accomplishment of precisely that active aspect of affective involvement called “attitude” (Schmitz 2007: 69). In order to show this, Schmitz refers to extreme experiences in which the subject is so overwhelmed that her active response becomes inoperative and the subject turns into an

214

Hermann Schmitz

indifferent observer of what is going on. As an illustration, Schmitz quotes a report from Erwin Bälz who had witnessed an earthquake in Tokyo. Bälz reports a sudden and total alteration: All higher affective life had gone out, all compassion for others, all concernment with possible harm, even the interest in the endangered friends and relatives and in my own life were gone, though my mind was completely clear [. . .]. I stood there and considered all these disastrous events around me having that cool concentration of someone monitoring a fascinating experiment. (Bälz 1901: 718; cf. Schmitz 2007: 66; own transl.) Schmitz interprets this kind of experience in terms of an absence of “mineness” in the sense of subjective factuality.6 Thus, subjective factuality goes hand in hand with the active aspect of affective involvement. Elsewhere, Schmitz also argues that the way in which one engages pain corresponds to the intensity by which that pain is experienced as oppressive, constricting, and, according to Schmitz’s theory of subjective factuality, self-involving (Schmitz et al. 2011: 16). If this is correct, the hypothesis of there being an intimate relation between subjective factuality, bodily sense of constriction, and attitude as the active engagement with the situation is further supported and the attitude is a sine qua non for “mineness,” i.e. it causes (bewirkt) subjective factuality. Now, the final step of the argument applies these ideas in order to show that the attitude satisfies the conceptual conditions for self-efficacy. This is how it is supposed to work: the attitude brings about the subjectivity (though usually not the content) of facts; yet the existence of one’s own attitude and its respective character are subjective facts themselves (Schmitz 2007: 69–70). These latter facts, however, are even trivially and causally self-sufficient. Thus, the attitude is self-sufficient and causes its own subjectivity. That is, it causes itself as a subjective fact and, thus, constitutes the causa sui (self-efficacy) that was looked for (Schmitz 2007: 70). Thus, human freedom is the freedom of one’s attitude, and therefore a person is morally responsible for her attitude precisely by virtue of her attitude (ibid.). This final step of the argument has the following structure: P1: The existence and quality of one’s attitude are subjective facts (especially in the context

of praising and blaming). P2: No subjective fact can have an objective fact as its sufficient cause. P3: One’s attitude is the sufficient cause of the subjectivity of any fact. C: Therefore, the attitude is the sufficient cause of itself as a subjective fact.

Initially, this may sound like a formal trick, but unpacking the premises will reveal that the argument is built on Schmitz’s phenomenological theory of subjective factuality and affective involvement. To begin with, the existence of one’s own attitude and its respective character are subjective facts (P1) insofar as one cannot help to acknowledge that these facts entail self-reference, i.e. that they are intimate facts about oneself. We can clearly see that this acknowledgment of self-reference is relevant in the drastic example of the negligent mother blaming herself (“How could I have been so unmindful?”), but there are, of course, less dramatic cases of acknowledging that this or that is one of one’s own character traits. Arguably, even a statement such as “as I re-watch the film, it seems as it was not even me” in the given context should not be interpreted as the speaker’s inability to acknowledge that something

215

Henning Nörenberg

about herself is revealed, but rather as the speaker’s (current) inability to smoothly integrate the revealed trait into the image she has of herself. Furthermore, Schmitz explains that the claim about the attitude accomplishing its own subjectivity would not be valid if subjective factuality were a mere add-on to otherwise plainly objective facts (Schmitz 2019: 129–130). Although he does of course not deny the existence of objective facts, Schmitz insists that there is no route that would lead from objective facts to subjective ones without the subject’s active involvement with these facts – which is precisely that subject’s attitude. In this sense, no subjective fact can have an objective one as its sufficient cause (P2). Schmitz illustrates this claim with the case of Phileas Gage (Schmitz 2019: 130): after an accident causing a severe lesion of his brain, Gage’s character – or, more precisely, his attitude – was fundamentally altered. There is no doubt that the lesion in question caused a particular objective fact, i.e. the fact of there being an altered attitude in the complex of states and events usually referred to by “Phileas Gage.” However, according to Schmitz, this is not the case for the subjective, i.e. self-involving or mineness-implicating fact that could have been predicated by no one else than by Gage himself by saying, for instance: “I, Phineas Gage, used to be a man of reliable stance, but now this has been thoroughly shaken” (ibid.). The nuance of mineness entailed in this fact about Gage cannot be accomplished by any enhanced precision of plainly objective facts, but requires the active part of affective involvement to be operative – even though this part has, objectively speaking, fundamentally changed in character. Thus (P3), it is the attitude that is the sufficient cause – not for the objective content of the fact, but for the nuance of mineness in virtue of which it is a subjectively meaningful fact (Schmitz 2007: 69–70). If this is correct, then the attitude is at least one suitable candidate satisfying the conceptual conditions for self-efficacy and is therefore a key element of human agency.

Concluding remarks Schmitz’s theory of attitude or Gesinnung as heart of human agency is built on his doctrines of self-consciousness as affective involvement and subjective factuality. In comparison to Sartre’s account, for instance, his theory is more modest with respect to the actual purview of self-efficacy. Although we have seen that there may be examples for the attitude’s effect on the that and how of experience, e.g. the subject’s engaging their pain, Schmitz hesitates to generalize such examples. Consider, for instance, Sartre’s hyperbolical claim that “[i]f I am mobilized in a war, this war is my war; it is in my image and I deserve it [. . .], everything takes place as if I bore the entire responsibility for this war” (Sartre 1956: 554). The attitudinal theory of free agency provides valuable conceptual tools for making sense of a claim such as “this war is my war,” namely by referring to the link between “mineness” and active self-involvement. However, whether it really would have been down to me to prevent, stop, or modify that war as an objective fact in any meaningful way cannot be sufficiently established within the framework presented. In this respect, Schmitz restricts himself to explaining the attitude’s role in choosing and intending. However, he also offers further arguments that are supposed to defeat the determinist’s claim to a thoroughgoing causal explanation of the lifeworld in order to make room for the lived confidence that there really is something that actually is up to us (cf. Schmitz 2007: 75–108). Although I cannot go further into detail, the references to Arendt I made throughout this chapter suggest that there is also high potential for mutual elucidation of her account of human action and Schmitz’s attitudinal theory of self-efficacy. That is, for instance, the theory of attitude provides us with valuable conceptual tools when it comes to making sense 216

Hermann Schmitz

of the “disclosure of the agent in speech and action” (Arendt 1998: 175), while Arendt’s discourse can help us understand further implications of that theory. Moreover, the attitudinal theory of agency is compatible with more recent accounts of “unreflective freedom” in important respects. Rietveld (2013), for instance, conceives of unreflective action such as running to catch the bus as a form of “allowing” oneself to be responsive to relevant situational affordances (see also Dreyfus 2007). The fact that an individual responds to these affordances rather than others is, at least in part, related to the fundamental way in which they engages the world. Whereas the “softy,” if I may say so, modifying one of Sartre’s favorite examples, may not find anything in the situation that motivates him to run for the bus (“it is futile, anyway”), a more agonal character may find that it is worth a try. We have also seen how Schmitz links the agent’s attitude as their independent initiative with choosing, rational conduct and the opportunity to do otherwise. However, with regard to Schmitz’s attempt to establish the attitude as an instance of self-efficacy, as it has been stated, a couple of open questions remain. For instance, his shift from transcendental constitution to a peculiar form of causation and, in line with this, the relation between “mineness” in terms of minimal self hood and “mineness” in terms of subjective factuality surely deserve further discussion. Furthermore, the way in which one’s attitude is experientially accessible to oneself should be fleshed out in more detail. However, even if we hesitate to follow Schmitz’s argument for attitudinal self-efficacy in every respect, the phenomenon of attitude is, arguably, a key element of human agency.

Related topics Chapters 11 (on Sartre), 14 (on Arendt), 17 (on Dreyfus).

Notes 1 Cf. Toporkova (2015). The English translation of the quoted sentence can be found in RT, October 21, 2015; https://www.rt.com/news/319316-reporter-hungary-refugees-facebook/ [last accessed on 19 April 2018]. 2 Linguistic annotation: unfortunately, Schmitz’s technical term Gesinnung has not been translated into English in a consistent manner. The relevant concept is referred to as “attitude” (Schmitz et al. 2011: 254), but also as “stance” (Schmitz 2019: 129). 3 “Non-trivial” is supposed to mean that the condition in question is not simply logically implied in the concept of moral responsibility (cf. Schmitz 2007: 22). 4 It is important to stress that subjective factuality is not simply an effect of that predication. Schmitz regards such predications as a good didactical vehicle to explain the difference between objective and subjective facts. However, he does not see any reason why subjective facts should not apply to babies and animals just because they do not have the linguistic capacities to state them. 5 In this context, “inaccessible” does not mean “experientially inaccessible,” but rather “hard to influence at will.” 6 This is probably one of the most crucial points at which Schmitz’s notion of subjectivity diverges from the one suggested by Zahavi (1999). As far as I can see, the latter would insist that, despite the alteration Bälz describes, the first-personal givenness of experience remains in place in some minimal form.

References Arendt, H. (1998) The Human Condition, Chicago, MA: Chicago University Press. Bälz, E. (1901) “Über Emotionslähmung,” Allgemeine Zeitschrift für Psychiatrie und psychiatrisch- gerichtliche Medizin 58, 717–721.

217

Henning Nörenberg Dreyfus, H.L. (2007) “The Return of the Myth of the Mental,” Inquiry 50/4, 352–365. Gallagher, S. and Zahavi, D. (2008) The Phenomenological Mind: An Introduction to Philosophy of Mind and Cognitive Science, London: Routledge. Heidegger, M. (1995) Phänomenologie des religiösen Lebens, Frankfurt am Main: Klostermann. Helm, B.W. (2017) Communities of Respect: Grounding Responsibility, Authority, and Dignity, Oxford: Oxford University Press. Holdsworth, N. (2015) “Camerawoman Who Attacked Refugees: ‘I’m Not a Heartless, Child-Kicking Racist,’  ” Hollywood Reporter, https://www.hollywoodreporter.com/news/camerawoman-whoattacked-refugees-regrets-822110 [last accessed on 19 April 2018]. Husserl, E. (2014) Grenzprobleme der Phänomenologie: Analysen des Unbewusstseins und der Instinkte. Metaphysik. Späte Ethik. (Husserliana XLII), New York: Springer. Rietveld, E. (2013) “Affordances and Unreflective Freedom,” in R.T. Jensen and D. Moran (eds.), The Phenomenology of Embodied Subjectivity, New York: Springer, pp. 21–42. Sartre, J.-P. (1956) Being and Nothingness: An Essay on Phenomenological Ontology, New York: Philosophical Library. Scheler, M. (1921) Vom Ewigen im Menschen, Leipzig: Der Neue Geist. Scheler, M. (1954) Der Formalismus in der Ethik und die materiale Wertethik (Gesammelte Werke 2), Bern: Francke. Schmitz, H. (1969) Der Gefühlsraum (System der Philosophie III-2), Bonn: Bouvier. Schmitz, H. (1973) Der Rechtsraum (System der Philosophie III-3), Bonn: Bouvier. Schmitz, H. (2007) Freiheit, Freiburg/München: Alber. Schmitz, H. (2019) New Phenomenology: A Brief Introduction, Milan: Mimesis International. Schmitz, H., Müllan, R.O. and Slaby, J. (2011) “Emotions Outside the Box: The New Phenomenology of Feeling and Corporeality,” Phenomenology and the Cognitive Sciences 10, 241–259. Schopenhauer, A. (1972) Die Welt als Wille und Vorstellung: Band 2, Wiesbaden: Brockhaus. Searle, J.R. (1983) Intentionality: An Essay in the Philosophy of Mind, Cambridge, MA: Cambridge University Press. Srinivas, R. (2015) “Petra Laszlo Apologizes: Fired Hungarian Camerawoman Apologizes for Kicking, Tripping Migrants,” Inquisitr, http://www.inquisitr.com/2410707/petra-laszlo-apologises/ [last accessed on 19 April 2018]. Strawson, P.F. (2008) Freedom and Resentment and Other Essays, London: Routledge. Toporkova, I. (2015) “Петра Лазло: ‘Мы собираемся подать в суд на Facebook,’” Izvestia, https:// iz.ru/news/593528#ixzz3p8X42aZv [last accessed on 19 April 2018]. Zahavi, D. (1999) Self-Awareness and Alterity: A Phenomenological Investigation, Evanston, IL: Northwestern University Press.

218

17 HUBERT DREYFUS Skillful coping and the nature of everyday expertise Justin F. White

Introduction Hubert Dreyfus’s work in the phenomenology of agency is distinctive for the privileged and central position he gives to our ability to navigate the everyday world. Drawing on the existential-phenomenological tradition—particularly the work of Martin Heidegger and Maurice Merleau-Ponty—Dreyfus characterizes skillful embodied engagement with the world (skillful coping) as the paradigmatic instance of human intelligence and agency. He uses the notion of skillful coping to push against the emphasis on deliberation he finds in the traditional view of human agency, in which an intention to perform some action is formed as the result of deliberation involving desires and beliefs about how to best satisfy those desires. As he sees it, the traditional view relies on an overly intellectualized conception of humans, leading to an overly intellectual conception of action. Instead of starting with a highly intellectual and deliberative conception of the human mind and then working to understand how that mind leads to action, Dreyfus takes skillful coping—what we do when we carry on conversations, drive cars, write papers, dribble basketballs, and chop onions—as paradigmatic of human agency and seeks to understand the type of mind involved. If skillful coping—also described as absorbed, immediate, everyday, and embodied—is “simultaneously the highest and most basic form of engagement with the world” (Wrathall 2014: 4), there are various implications for the nature of human action and intelligence.1 One of Dreyfus’s central claims is that the complex understanding involved in skillfully performing apparently ordinary tasks need not be highly reflective. Indeed, Dreyfus thinks, the fluid coping distinctive of expert performance is often marked by a lack of deliberation and even a lack of conscious awareness of the process in which she is involved (Dreyfus 2014: 34). Highly deliberative action is not a sign of expertise, but of its lack. The expert driver no longer needs to deliberate about when to shift gears or change lanes and instead fluidly responds to the feel of the engine (with changing gears) and to the flow of traffic (with shifting lanes). Dreyfus claims that the same is true of experts in paradigmatically intellectually demanding pursuits—such as playing chess, performing Liszt’s “La Campanella,” interpreting complex scientific results, and crafting detailed furniture. What is distinctive of these experts is less their complex mental representations or virtuosity in deliberation than it is their ability to respond to the highly particularized significances or affordances of a concrete 219

Justin F. White

situation. Dreyfus connects the ability to respond to the particular significances of a concrete situation with typically existential concerns, such as authenticity. On his reading of Being and Time, the practical world is articulated for the authentic individual with fine-grained particularities not apparent to the inauthentic individual. The authentic individual thus can more skillfully navigate the world of practical concern by responding to the fine-grained and precise particularities she encounters than is possible for the inauthentic individual who finds herself in a situation articulated in terms of general features. The way Dreyfus connects skillful coping with authenticity exemplifies his syncretic approach in which he synthesizes and, arguably, extends the work of thinkers such as Heidegger and Merleau-Ponty. It is through his engagement with Heidegger and Merleau-Ponty that Dreyfus develops his account of skillful coping as a critique of the philosophical tradition, which he sees as intellectualistic and deliberative in a way that distorts our understanding of human existence. This chapter is organized around key themes in the account of skillful coping he contrasts with the traditional view of human intelligence and agency: the relationship between mind and body (and mind and world) and the importance and the nature of the context of human existence, the nature of learning and skill development, the relationship between flow, deliberation, and what the engaged agent perceives, as well the relationship between expertise and authenticity. The concluding section examines one worry about Dreyfus’s characterization of expertise within the skillful coping paradigm, particularly in light of his emphasis of the nondeliberative nature of skillful coping. Specifically, does Dreyfus’s rejection of the tradition’s emphasis on deliberation lead him to excessively minimize the role of deliberation as one develops and expands expertise? That is, how can one reconcile the skillful, absorbed coping Dreyfus emphasizes with the process of deliberately and intentionally learning new skills or of extending the realm in which one can skillfully cope? Dreyfus’s emphasis of the nondeliberative nature of skillful coping could seem to problematically push deliberative practice out of the picture. My suggestion is that there is space for such deliberate learning and skill development for experts within the framework Dreyfus proposes. To flesh out the place of deliberate learning, I draw, appropriately, on both Dreyfus and Merleau-Ponty to suggest how intentionally (and deliberately) trying to live in a certain way—whether acquiring new skills or expanding or refining one’s practical framework—is not at odds with but is ultimately crucial for the expertise Dreyfus has in mind, that of absorbed, fluid, and skillful coping in a practical world.

A note on historical influences As noted earlier, Dreyfus’s views on skillful coping often arise through his engagement with key figures in the existentialist tradition and, perhaps to a fault, he typically credits his insights to these philosophers. As Mark Wrathall (2014: 2) puts it, “In his published works, Dreyfus tends (rather too modestly, in my opinion) to attribute his insights to other philosophers—primarily Heidegger and Merleau-Ponty.” His tendency to attribute insights to others and the extent to which the version of Heidegger received in Anglo-American analytic philosophy is Dreyfus’s Heidegger has led some to use the term “Dreidegger” to refer to the fusion of Dreyfus and Heidegger, a fusion some think distorts Heidegger’s views (see Kelly 2005: 15). The same could perhaps be said of his interpretation of Merleau-Ponty and others. That said, Dreyfus’s engagement with these figures, and the subsequent work of his students, has played an important role in the now flourishing scholarship on Heidegger and Merleau-Ponty. 220

Hubert Dreyfus

But whether Dreyfus is primarily interpreting these figures, finding support for his views in them, or some combination of the two, Dreyfus’s views are often presented as readings of Heidegger or Merleau-Ponty and as criticisms of what he sees to be the prevailing (and mistaken) philosophical views—whether the views are Edmund Husserl’s, John Searle’s, or John McDowell’s. Through his reading of Heidegger, Dreyfus finds, for example, emphasis on practice over reflection (with more deliberative thought coming to the fore primarily in cases of practical breakdown) and on the way in which our human purposes shape the practical context, in addition to an account of the holistic nature of significance. In Merleau-Ponty, he finds, among other things, an emphasis on the thoroughly embodied nature of perception and action, as well as the notion of maximal grip, which he uses to show how one can attend to the particularities of a situation without robust (or perhaps any) conceptual representation. More broadly, it is through his engagement with Heidegger and Merleau-Ponty that Dreyfus develops his account of skillful coping.

Skillful coping Dreyfus thinks the philosophical tradition as a whole has long relied on an overly intellectualized account of human existence. By focusing on deliberative rationality as what is distinctive of human agency, we either overlook or distort the skillful but not always intellectual way we navigate the world. By focusing on the ubiquitous phenomenon of skillful coping, Dreyfus aims to understand the nature of human existence on its own terms rather than starting with some prior theoretical commitment—such as the assumption that humans are distinctive for a highly conceptual and intellectualized rationality—and then trying to fit human existence into a framework provided by that commitment. He thinks that careful analysis of everyday skillful actions yields a rich view of what is distinctive of human existence, including human intelligence and agency. Indeed, he thinks that typical conceptions of human intelligence and agency arise from analyzing human existence that has stalled. It is as if we have attempted to understand the nature of a functioning car engine by lifting up the hood on a car that has broken down, instead of looking at a functioning engine. Rather than illuminating human existence in action, we end up making claims about the nature of human existence when it is not functioning smoothly because something has gone wrong. In his descriptions of the skillful activity of experts, activities such as chess and athletic expertise figure prominently, but he also attributes robust expertise to everyday skills such as driving a car, understanding and speaking a language, and knowing how close to stand to others in a conversation. Looking carefully at the phenomenon of expertise—or engaged skillful coping—will illuminate key features of Dreyfus’s account of agency. One prominent element of Dreyfus’s account is the idea that our everyday being-inthe-world is typically unreflective and nonconceptual, even if our intellectual, conceptual capacities offer us unique possibilities. As he puts it: “What makes us special, then, isn’t that, unlike animals, we can respond directly to the conceptual structure of our environment; it’s that, unlike animals, we can transform our unthinking non conceptual engagement, and thereby encounter new, thinkable structures” (Dreyfus 2014: 123). On Dreyfus’s account, these “conceptual structures are not implicit in our involved experience any more than reasons for our actions are implicit in our expert coping, or than the detached attitude is implicit in the engaged one” (123). Our uniqueness depends instead on our ability to step back from our everyday “unthinking non conceptual engagement” with the world and think of it abstractly. 221

Justin F. White

But even if the capacity to step back from our everyday engagement with the world, to become reflectively self-aware, is crucial to our uniqueness as persons, it is relatively uncontroversial to claim that we are not always reflectively self-aware, especially with regard to the processes guiding our actions. However, imperfect reflective self-awareness can be compatible with the traditional picture: one could hold that the conceptual structures and deliberative processes are at play; we are simply not fully aware of them. On the traditional view, the expert has a more robust mental representation than the non-expert and more effectively deliberates about the considerations involved in the various relevant scenarios. Dreyfus, however, pushes back by suggesting that reading mental representation or unnoticed deliberation into expertise derives more from a commitment to the deliberative paradigm of human action than from careful attention to the nature of expertise.2 Although in this piece I typically adopt Dreyfus’s characterizations of notions such as concept, representation, reflectiveness, and deliberation, it should be noted that his characterizations of these and associated notions and his related criticisms of what he calls “the traditional picture” are controversial and lie at the heart of many criticisms of Dreyfus’s views. For example, the way Dreyfus (2007) equates “direct and unreflective” with “nonconceptual and nonminded” lies at the heart of his debates with Searle and McDowell.3 It is also one of Sacha Golob’s (2014) key points in his criticisms of the dominant work on Heidegger, work exemplified by Dreyfus and those influenced by him, such as Mark Wrathall and Taylor Carman. People like Searle, McDowell, and Golob criticize Dreyfus for relying on an overly intellectualistic and detached notion of concepts and representations, while Dreyfus consistently presses that the nature of skillful coping is nonconceptual and need not involve the representational content involved in what he sees as the traditional view. For one example, in “What Is Conceptually Articulated Understanding,” Joseph Rouse (2013) develops the distinction between descriptive and normative accounts of conceptually articulated content in order to pinpoint the disagreement (or sometimes, lack of disagreement) between Dreyfus and those he criticizes (and who criticize him). On a descriptive account of conceptually articulated content, conceptual content is “something actually present or operative in specific performances by concept users” (Rouse 2013: 250). Normative accounts, by contrast, “identify the conceptual domain with those performances and capacities that are appropriately assessed according to rational norms” (2013: 250). Dreyfus’s views could challenge the descriptive account but not the normative account (2013: 252). Because an adequate analysis of these issues would require (indeed, has required!) a chapter or book of its own, suffice it to say that the debate about the precise nature of concepts, representational content, mindedness, and the like, as well as their respective places in human existence—particularly when it comes to skillful, embodied coping—has been lively and is ongoing. In this chapter, when I refer to concepts, conceptuality, representations, representational, and the like, I typically adopt Dreyfus’s usage and, temporarily, bracket questions of whether his understanding of those notions is justified either philosophically or as an interpretation of historical figures such as Heidegger. Let us turn now to Dreyfus’s account of skillful coping and to what he thinks we can learn about human existence through careful attention to skillful coping.

Being-in-the-world, mind and world, and context Dreyfus thinks a careful analysis of skillful coping reveals an agent intertwined with the world in a way that undermines the common divide between mind and world. For one, Dreyfus thinks the “mind” in the mind/world distinction is too narrow to capture the way 222

Hubert Dreyfus

an agent relates to the world: the entity relating to the world is the whole human being and not just the mind. We are embodied beings and our bodies not only tie us inextricably with the world but shape the way we perceive and engage with the world. But Dreyfus suggests further that thinking of it as a “‘relation’ [between mind and world] is misleading, since it suggests the coming together of two separate entities—human being and world—whereas . . . mind and world [are] inseparable” (Dreyfus 2014: 170). Put another way, “[in] our most basic way of being—i.e. as skillful copers—we are not minds at all but one with the world” (Dreyfus 2014: 259). Dreyfus draws on Heidegger’s notion of being-in-the-world to explain the unitary nature of human existence, in which we find a human agent always already enmeshed in a practically significant world. The world solicits her to act in particular ways depending on her aims or practical commitments. The embodied coper responds directly to the affordances in the world—“food affords eating, doors afford going in and out, floors afford walking on” (Dreyfus 2014: 116).4 When looking for something to chop onions, the chef ’s knife on the counter shows up as “for chopping,” but if clearing kitchen clutter, the same knife could show up as “to be put away.” On this picture, the grounds for what we do lie with the world as much as the agent. The natures of knives and onions, say, play crucial roles in the way we use them. Dreyfus (2014: 95) uses the example of the difference between dribbling a basketball and a mime trying to mimic the act of dribbling. The act of dribbling is grounded by the contours of the basketball, the give of the basketball, and the amount of bounce of the surface (to say nothing of the particular details in an actual game) in a way that requires a sensitivity to the specificity of the situation that even the most skilled mime cannot re-create. For example, the ball makes an impression on my skin when it bounces up and creates friction that cannot be reproduced without the actual ball.5 The extent of the world’s influence becomes clearer if we consider the bounce of different surfaces (asphalt, sidewalk, or hardwood), whether the surface is level or slanted, the way temperature affects the ball’s springiness and my dexterity, and so forth. Within the holistic conception of agency Dreyfus proposes, my whole self is involved in agency and the contours of agency depend on a larger whole. In short, the human agent is more than a mind and the agent is intertwined with the world. The way the world appears depends on my aims, skills, and dispositions, and the way I press into possibilities depends on subtle features of the world. In contrast to the traditional account of human action, which assumes “a clear distinction can be drawn between an agent and the world within which the agent acts” (Wrathall 2014: 2), Dreyfus’s account proposes that agent and world are inextricably intertwined in a relationship with porous boundaries. One reason for the intertwinement is that the world, as we primarily experience it, depends on agential involvement—on aims, concerns, commitments, and cares. Even at the level of perception, we first see the world in terms of affordances. When we are tired on a hike or arriving late to a lecture, we are more likely to see the world primarily in terms of whether it affords sitting—to take a rest when hiking or to find a spot in the crowded lecture hall. But conversely, the shape of our aims, concerns, and commitments also depends on the stuff we find around us. The roles and pursuits available, and the forms they take, depend on a broader context. In some cases, the dependence of roles on context is obvious: becoming a computer scientist wasn’t an option for someone living in 14th-century London. But the dependence relation can be less obvious: does it mean something different to be a parent or spouse than it did a few centuries (perhaps even a few decades) ago? Although we have a hand in determining the type of parent or spouse we will be, the shapes of these identities are not (entirely) up to us. These identities or, on Dreyfus’s reading of Heidegger, “for-thesake-of-whichs” depend on other agents. But even if the way the world appears depends on 223

Justin F. White

having some for-the-sake-of-which, Dreyfus insists that engagement with the world need not be conceptual or reflective.6 Skillfully coping as a parent does not depend on consciously thinking of oneself as a parent. When I skillfully cope in the world, I can see or feel what the situation calls for, or affords, and respond effectively without representing my goal to myself and without the situation being conceptually articulated. Of course, the environment and what it calls for depend on my roles, identities, and dispositions. But even if those give me a foothold in the world, I need not consciously think of myself in terms of those features of my existence.7 In fact, for Dreyfus, consciously thinking of myself as a parent (or reflective self-awareness) is a sign that my engagement with the world has come up short. It is when I don’t know how to respond in a situation that I stop and ask, “As a parent, what should I do here?” There are two key points here: (1) the affordances we perceive in our practical engagement with the world depend on our roles and identities, but (2) the nature of these affordances and our experiences that make them possible need not be conceptually rich. The nature of affordances offers a way to understand how what is given can be meaningful but free of concepts (2014: 116). The world can be meaningfully structured for agents without the presence of concepts. In skillful coping, one responds to the particularities of a situation without the mediation of representation or concepts: “skillful coping does not require a mental representation of its goal at all. It can be purposive without the agent entertaining a purpose” (Dreyfus 2014: 84). But if skillful coping need not involve concepts or representation, what does it require? And what distinguishes experts from less-skilled performers, if it is not a richer conceptual framework and the ability to act on that framework?

Rules, skill acquisition, and expertise Hubert Dreyfus and his brother Stuart tackle this question in their analysis of skill acquisition and expert performance. They start with the novice and then move to advanced beginner, competence, proficiency, and expertise. The early stages of skill acquisition (novice and advanced beginner) are characterized by (often painfully) conscious rule-following and careful deliberation about the best action to pursue. The learning driver learns the appropriate distance (in terms of some unit of measurement) to maintain between her car and other vehicles; the beginning chess player learns context-free values of the various pieces and generic strategies to develop certain pieces first and, perhaps, to castle by the sixth move. At the level of competence, things start to change. The competent performer consciously chooses “a plan, goal, or perspective which organizes the situation” (Dreyfus 2014: 32). The particular plan or goal chosen determines the contours of the practical situation, making different elements stand “out as more or less important with respect to the plan, while other irrelevant elements are forgotten” (2014: 32). If one chooses to aggressively attack the opponent’s king, for example, the board and pieces will appear in light of that goal. Opportunities to advance her queen or rooks, or to fork, could stand out, as could obstacles to her plan. With a lot of experience, she may transition from competence to proficiency, coming to spontaneously see multiple possibilities in a situation. Instead of consciously choosing a plan (as a competent performer does), she intuitively sees what a situation requires. The proficient chess player can read the board and know that she should attack her opponent’s king, even if she then needs to deliberate about the best method of attack. As one progresses from proficiency to expertise, she spontaneously sees not only what a situation calls for her to do—say, attack the king—but also the best (or at least better) ways to do it. The situation could also call for action that defies the generic rules she initially learned—for example, to sacrifice a piece of value for positional advantage. However, the expert can see what to do without conscious awareness of why that path 224

Hubert Dreyfus

stands out. And it is not that the expert is simply unaware of the rules underlying her decision process. Instead, through vast experience, the expert has great situational understanding that is not grounded in (and is not reducible to) rules. In the early stages of skill acquisition, she relied on rules. But not only does she no longer rely on those rules, but her situational understanding “def[ies] complete verbal description” (2014: 35).8 The brothers Dreyfus suggest that the expert’s reliance more on pattern recognition than on an advanced understanding of rules explains why early attempts to develop computers with genuine chess expertise had fallen short of master level chess playing. These computers were programmed with the rules of chess, and with strategy rules offered by expert chess players (2014: 38). But if the Dreyfus account is right, the focus on rules explains why early programs were marginally competent by normal chess standards.9 In contrast to the traditional account according to which the expert better understands and incorporates relevant rules into her actions, the Dreyfuses claim not only that the expert is unaware of the rules guiding her actions, but that rules no longer underlie her actions. They refer approvingly to Euthyphro’s refusal to give a rule-based account of piety and his focus on examples of piety. For the expert, this refusal is more than mere refusal; an expert’s understanding is distorted if reduced to rules, however complex or intricate. The expert started with rules and could learn new rules along the way, but her expertise is not reducible to rules.10 As Wrathall concludes from Dreyfus’s analysis, “The phenomenology of skill acquisition . . . shows that agents who possess different levels of skill are, in a very real sense, engaged in different kinds of activities” (Wrathall 2014: 4). On the traditional model of skill acquisition, the same rules govern the actions of the beginner and the expert alike—the expert simply has a better understanding of the relevant rules and is better able to apply them in new situations (Wrathall 2014: 6). On Dreyfus’s account of skill acquisition, however, the expert no longer applies rules, but instead recognizes and responds to highly particularized situations (see Dreyfus 2014: 34–43). Thus, Dreyfus thinks, “the basis of expert coping may well be the sort of features that the expert could not be aware of and would not be able to think” (2014: 120). While skillfully coping, the expert responds to felt tensions in the practical gestalt, an ability that depends on highly particularized motor skills but need not involve a rich conceptual repertoire (see Dreyfus 2014: 81, 150–152). The possibility that experts could be unaware of the features they are responding to could help explain why interviews with expert performers about their performances can be disappointing. If we expect answers that show a complex understanding of advanced rules, it can be a letdown experts offer up the same rules that a beginner would, apparently falling back on clichés: “If one asks the experts for rules one will, in effect, force the expert to regress to the level of a beginner and state the rules he still remembers but no longer uses” (Dreyfus 2014: 36). So even if their answers are sophisticated, there is reason to be suspicious whether their description accurately captures what was going on. Because expertise is not rule based, when an expert is forced to give the reasons that led to the action, his account will necessarily be a retroactive rationalization that shows at best that the expert can retrieve from memory the general principles and tactical rules he once followed as a competent performer. (2014: 113)

Flow, deliberation, and agential perception By taking the fluid action of such expert performers as paradigmatic of human action and intelligence, Dreyfus pushes against the deliberative approach to action. Because flow is 225

Justin F. White

characterized not by deliberation “but by the withdrawal of deliberative activity,” the expert is not demurring or avoiding giving a genuine answer when she describes her performance with basic rules that seem inadequate to the complexity and sophistication of her actions (Wrathall 2014: 3). When in the midst of peak performance, she is not rapidly deliberating among rules—simple or complex—and so any attempt to recapture what she was doing in deliberative terms is destined to disappoint. However, her failure to articulate (or even to be guided by) complex rules is only a failure if we are antecedently committed to the paradigm of deliberative action, a paradigm prone to read deliberation into what is better understood as skillful but spontaneous action (see Dreyfus 2014: 191). On Dreyfus’s picture, deliberation about what to do and about what rules to follow is a sign of the limits of one’s expertise: When a master has to deliberate in chess or in any skill domain, it’s because there has been some sort of disturbance that has disrupted his intuitive response. Perhaps the situation is so unusual that no immediate response is called forth. Or several responses are solicited with equal pull. (Dreyfus 2014: 118) But if the expert performer in flow does not deliberate, it could seem that it turns expert action into something that is merely instinctual or a matter of reflex. And Dreyfus thinks there is something similar to instinctive or reflexive behavior, but it also involves what he calls primordial intentionality or purposiveness. When Dreyfus says above that humans can relate to the world in purposive ways without entertaining a purpose, he means they need not represent to themselves the state of affairs they are aiming to achieve or even what they are doing when they perform the action.11 We can find ourselves arriving to the office after commuting on “autopilot,” for example, without representing to ourselves the end of arriving at the office or the specific route we are taking. What Dreyfus’s characterization of primordial intentionality and the way it pervades much of our daily lives suggests, however, is that some instances of “autopilot” are more sophisticated than they might first appear. The master chess player and the expert tennis player need not represent their goals to themselves when playing chess or tennis, respectively. Searle replies that in such cases, some broad intentional action is represented—the tennis player wants to win the point or the chess master wants to checkmate the opponent (or put pressure on the opponent’s king) (Dreyfus 2014: 85). Under this represented intention, Searle thinks, the expert need not form fine-grained intentions to, say, turn her racket to a specific angle or to advance her bishop in a certain way. Dreyfus admits that there is something right in Searle’s response that the action is intentional (and represented to the agent) on some level, but he worries that the broad intention Searle describes overlooks the fine-grained nature of what the expert is doing. In trying to win a point, she could also rush the net, hit a cross-court backhand or slice the ball at a specific angle, and so on. And it is even more fine-grained than that. The novice and the expert could both be aware of different sequences that could win a point. They could even state rules about how to hold a racket in order to get a certain spin on the ball. But when the expert tries to win a point, she responds to the subtle situational forces at play by hitting the ball with a specific kind of spin, by hitting it sharply instead of slightly cross-court, and so forth. She intentionally does these things, but they happen too quickly for deliberation or conscious representation. As David Foster Wallace puts it in an essay on Roger Federer: The upshot is that pro tennis involves intervals of time too brief for deliberate action. Temporally, we’re more in the operative range of reflexes, purely physical reactions that 226

Hubert Dreyfus

bypass conscious thought. And yet an effective return of serve depends on a large set of decisions and physical adjustments that are a whole lot more involved and intentional than blinking, jumping when startled, etc. (Wallace 2012: 23) The expert’s skillful response to the situation is not reducible to rules, but it is not mere reflex either.12 As motor intentionality, it is intentional but without propositional success conditions (Dreyfus 2014: 94). So how, exactly, does the expert experience the world she navigates? On Dreyfus’s account, the agent initially experiences the world in terms of affordances (as “for-whats,” not “whats”), as “for sitting” (chairs) and “for writing” (pens) (Dreyfus 2014: 255). Because expertise involves coming to see (or perhaps feel) the affordances of a situation and developing the skills necessary to respond well, one could skillfully respond to a situation without consciously representing some aim or intention. One who is very familiar with an activity or situation and has trained until skilled responses are a matter of muscle memory can see what the situation invites, or feel the tension in the current practical gestalt, and respond in a skillful way without being consciously aware of what one sees, why one feels the tension, or of the precise way one is responding. Dreyfus uses the example of instructor pilots to support his point: Instructor pilots teach beginning pilots a rule for determining the order in which to scan their instruments—a rule the instructor pilots were taught and, as far as they know, still use. At one point, however, Air Force psychologists studied the eye movements of the instructors during simulated flight and found, to everyone’s surprise, that the instructor pilots were not following the rule they were teaching; in fact, their eye movements varied from situation to situation and did not seem to follow any rule at all. The instructor pilots had no idea of the way they were scanning their instruments and so could not have represented to themselves what they were doing. (Dreyfus 2014: 85) The expert pilots were unaware of the way they varied the order in which they scanned their instruments depending on the situation. With many hours of training, they had moved beyond the rules they were taught—and that they continued to teach—but they were unaware that they had done so and were thus unable to articulate new rules in their place. And more, given the situational-dependent nature of their actions, there did not seem to be any articulable rule guiding their actions. The expert, then, sees the affordances of the situation and responds well without being fully aware of the nuances of what she is doing. Mastery, on this picture, “requires a rich perceptual repertoire—the ability to respond to subtle differences in the appearance of perhaps hundreds of thousands of situations—but it requires no conceptual repertoire at all” (Dreyfus 2014: 119). Dreyfus follows J.J. Gibson in claiming that our “pickup of affordances as high-order invariants in the optic array, and neural net considerations as to how the brain might detect such invariants, suggest that expertise does not require concepts” (2014: 119). Extensive training and experience allow experts to perceive and skillfully respond to affordances of which they may not be consciously aware.

Expertise and authenticity As mentioned earlier, one distinctive element of Dreyfus’s account of expertise is the way he interweaves his account of skillful coping with existential themes such as authenticity. 227

Justin F. White

Drawing on the notion of leveling in Heidegger’s discussion of das Man and inauthenticity in Being and Time, Dreyfus claims that when an individual becomes authentic (taking ownership for herself ), the world of her practical engagement can appear in its fine-grained particularities and not merely as a general situation. The situation of the authentic agent reveals itself to her in its concrete specificity and not as some generic context. In Heidegger’s discussion of das Man (sometimes left untranslated or, if translated, typically translated as “the one” or “the they”), Dreyfus finds a description of what a general, rule-based understanding of a situation would be. Unlike the expert who sees the particularities of the situation in ways that cannot be captured by rules or general characterizations, the understanding of das Man is not sensitive to the concrete particularities of the situation. By taking ownership of herself, the expert no longer depends on the general rules of das Man and, instead, comes to see and respond to the situation in its specificity. Describing ethical expertise, Dreyfus writes, Caring in its purest form is not ordinary loving; it is doing spontaneously whatever the situation demands. As we have seen, even if two situations were identical in every respect, two ethical experts with different histories would not necessarily respond in the same way. Each person must simply respond as well as he or she can to each unique situation with nothing but experience-based intuition as guide. Heidegger captures this ethical skill in his notion of authentic care as a response to the unique, as opposed to the general, situation [Being and Time, 346]. (Dreyfus 2014: 199) These claims about ethical expertise are representative of claims about expertise in other arenas. In some cases, authenticity and expertise are so intertwined that it can be hard to disentangle the order of priority. For example, one can only be a radical innovator, or a world disclosing master (which Dreyfus characterizes as the highest level of skillful activity), when one moves beyond conformity with received practices and, instead, opens up new possibilities for ways of being and novel ways to take up inherited practices in some activity. For example, when Dick Fosbury introduced what is now known as the “Fosbury flop” at the 1968 Olympic Games, he revolutionized the high jump. Although Fosbury’s new approach also led him to be successful to typical standards—he did, after all, win the gold medal—world disclosure may depart from typical standards of excellence. But world disclosure reveals that we, as humans, are essentially world disclosers, disclosing new ways of being. The richness and influence of Dreyfus’s work likely stems in part from the way he draws on and makes connections between different (sometimes apparently disparate) thinkers and traditions. However, even if this syncretic approach is often a virtue, it is not without its worries. Despite the creativity of the resulting views, it is not always clear that the pieces fit together as seamlessly as Dreyfus suggests. For example, does the connection between the existentialist notion of authenticity and skillful, expert coping work the way Dreyfus proposes? One way to explain this connection is to see the transition to expertise happening as a proficient performer recognizes the inadequacy of rules to the complexity and specificity of the practical situations that arise in skillful coping—whether playing chess, flying jets, or performing surgery. This recognition could lead her to take responsibility for her absorbed coping in a way that allows her to move beyond the rules she initially learned and instead attend to the particularities of the situation. This is only a sketch and thus is underdeveloped, but one could worry that even though self-ownership and acting for one’s own reasons are

228

Hubert Dreyfus

crucial to Heidegger’s account of authenticity, it is unclear if the skillful coping in Dreyfus’s picture can do the existential heavy lifting associated with Heidegger’s notion of authenticity and its entanglement with anxiety, facing one’s death, and the like.

Skillful coping and the ongoing pursuit of expertise In what remains, we will examine the process of learning new skills and expanding one’s expertise, in light of Dreyfus’s conception of expertise as the ability to continue in absorbed coping. As we have seen, Dreyfus goes to great lengths to minimize the role of concepts, representation, and deliberation in expertise and skillful coping. Sometimes he even suggests a different standard of expertise. Expertise is typically thought to involve excellence or success in some activity—consistently winning tennis or chess matches, for example. By contrast, Dreyfus (2000) proposes that being an expert does not always mean that one is very good at the relevant activity—whether chess, tennis, driving, or carrying on a conversation. On Dreyfus’s account, someone who learns tennis by chopping at the ball and never learns a forehand stroke can become an expert tennis player insofar as she can continue to respond intuitively with the best possible chop. On the standard story, expertise is determined with reference to the aim of achieving success in the standard sense—say, winning the point or the match; on Dreyfus’s conception, however, the aim according to which one is an expert is that of “ongoing coping as an end in itself ” (Dreyfus 2014: 245 n24). For Dreyfus, expertise involves fluidly coping without pausing to act deliberately, even if that fluid coping sometimes falls short of success in the standard sense. There is something attractive in Dreyfus’s expansive understanding of expertise, expertise as maintaining one’s comportment without needing to pause to deliberate about what to do. For example, persisting in some activity despite apparent limitations (such as the lack of a forehand) could allow one to develop novel responses to a range of situations in order to allow fluid coping to continue. The chopping tennis player will likely expand the range of situations in which one could chop. However, there is also something potentially restrictive in Dreyfus’s expansiveness. We find ourselves in a sort of dilemma. If the aim of expertise is ongoing coping (without pausing to deliberate) as an end in itself, the expert could develop novel approaches in one’s current paradigm in order to maintain fluid coping. The chopper could find innovative ways to chop in a tennis match, for example. The worry in this direction, however, is that while the novel approaches developed by the chopping tennis player may allow her to continue to play tennis fluidly without pausing to deliberate, the novel approaches she develops may be less attentive and well-suited to the particularities of the situation. In the other direction, however, although expanding one’s skills would likely allow one to respond more effectively to a wider range of situations that arise in the context of a tennis match, Dreyfus is right that learning new skills and new ways to do things will likely disrupt fluid coping—learning a forehand or learning to dribble with one’s left hand can be slow, deliberate processes involving far-from-fluid coping. Over time and with lots of practice, new skills can become part of one’s absorbed coping, thus expanding the range of situations in which one can fluidly cope. But there would be a period in which one’s performance falls short of the aim of ongoing coping. Dreyfus’s explanation is that when learning new skills, the expert’s behavior is not that of an expert: “When we are following the advice of a coach, for example, our behavior regresses to mere competence. It is only after much practice, and after abandoning monitoring and letting ourselves be drawn back into full involvement in our activity, that we can

229

Justin F. White

regain our expertise. The resulting expert coping returns to being direct and unreflective, which I take to be the same as being nonconceptual and nonminded” (Dreyfus 2007: 355). Expertise thus seems to be incompatible with a certain kind of deliberate learning. If Dreyfus’s aim is to shift the understanding of expertise away from success in an activity and toward ongoing coping, the resulting picture is more than unintuitive. In order to avoid the deliberative element of skill acquisition, Dreyfus’s account could problematically devalue the learning process. My worry is less about Dreyfus’s expanding expertise to include the chopper than it is about the risk of mischaracterizing the person who pauses and deliberately reinvents herself in her pursuit of ever greater expertise. Although Dreyfus is surely right that one who continues to chop at the ball (and thus maintains fluid coping) instead of mastering a forehand may reveal new chopping possibilities in tennis, insofar as she seeks to continue to exist as an expert, that individual is also restricted to the chopping paradigm. One who takes the time to learn a forehand may experience less-than-fluid coping for a time, but it is in the interest of ever great fluid coping possibilities. Dreyfus uses Bill Russell as an example of someone who exhibits the fluid coping of expert performance (see Dreyfus and Kelly 2011). Russell could skillfully cope in an apparently endless array of situations. And when he was in the zone, he was less worried about who won the game than about maintaining the magic of the situation (Russell and Branch 1979: 157; Wrathall and Londen 2017). But here Dreyfus seems to be helping himself to Russell’s expertise in the standard sense. If we compare the chopping tennis player to Russell, Russell’s ability to skillfully cope in a variety of situations and respond fluidly to what the situation invites is the result of extensive training and skill, in sharp contrast to the chopping tennis player’s relatively limited ability to play tennis. There is something worrisome if, as Wrathall and Londen suggest in their characterization of Dreyfus, the aim of maintaining the fluidity of the comportment “would be undermined were she to play a forehand shot because, never having mastered the forehand, this would force her to act deliberately, and draw her away from her optimal gestalt” (2017, MS). Admittedly, there is something appealing, even beautiful, in the focus on fluid action and the way in which our expertise extends to all sorts of mundane activities. But expertise that avoids deliberation can also be stunting. To be sure, adding an effective forehand to one’s repertoire likely requires (perhaps painful) learning and unlearning, writing and rewriting muscle memory through intensive training and practice. But it also, in time, allows for the phronesis and world-disclosive mastery that Dreyfus emphasizes.13 Unless supplemented with typical standards of skill, Dreyfus’s alternate and expansive notion of expertise ultimately can limit one’s ability to fluidly cope in a way that is genuinely sensitive to the affordances of the concrete situation. The other side of Dreyfus’s emphasis on absorbed, skillful coping is the way he downplays the importance of conscious, deliberate trying. His shift offers an important corrective to the highly intellectualistic and rationalistic picture of human agency that he saw prevailing. But if taken to the extreme, his account could end up privileging agents who continue to do what they already do well above those who seek to develop new skills or to alter their approach to some existing skill, since learning new skills and refining skills both can involve periods of deliberative and relatively unskilled behavior. But because this type of learning seems to be an expression of one’s expertise, this could seem to pose a problem for Dreyfus’s account. My proposal is that in order to better capture the whole of human agency—including expert agency—we would do well to appreciate the potential fluidity between skillful coping and deliberative aspiration, planning, and learning. However, I think we can do that within a Dreyfussian framework. In what remains, I draw on Dreyfus’s own writings and 230

Hubert Dreyfus

(appropriately for an essay on Dreyfus) on passages from Merleau-Ponty’s Phenomenology of Perception to briefly sketch an account of the relationship between intentional, conscious skill development and the absorbed coping Dreyfus emphasizes throughout his work. In Phenomenology of Perception, Merleau-Ponty writes: Taken concretely, man is [ . . . ] this back-and-forth of existence that sometimes allows itself to exist as a body and sometimes carries itself into personal acts. Psychological motives and bodily events can overlap because there is no single movement in a living body that is an absolute accident with regard to psychical intentions and no single psychical act that has not found at least its germ or its general outline in physiological dispositions . . . [Through] an imperceptible shift, an organic process opens up into a human behavior, an instinctive act turns back on itself and becomes an emotion, or, inversely, a human act becomes dormant and is continued absentmindedly as a reflex. (Merleau-Ponty 2012: 90) On Merleau-Ponty’s account, the human agent exhibits two different ways of existing.14 Sometimes she exists as a body with motor understanding, an existence characterized by habitual actions and reflexes, some of which were once intentional actions but have now sunk into muscle memory; others she exists in a way that involves what Merleau-Ponty calls “personal acts,” an existence characterized by self-conscious planning, assessment, and deliberation about what to do. Instead of being at odds, however, these ways of existing are intertwined in both directions. In Dreyfus’s terminology, our absorbed skillful coping (Merleau-Ponty’s motor intentionality) is shaped by our deliberation (Merleau-Ponty’s personal acts) and our deliberative agency depends on our skillful bodily coping. In “The Primacy of Phenomenology,” Dreyfus gestures at the entanglement of these two ways of being. Of Merleau-Ponty, he writes: He calls the phenomenon of absorbed coping motor intentionality, and claims that it is our basic form of intentionality, missed by those who suppose that in all comportment the agent’s movements must be guided by what the agent (consciously or unconsciously) is trying to achieve. (Dreyfus 2014: 150) A little later, however, he adds: “Absorbed coping turns out to be occasioned by trying, but the skilled activity itself is caused not by trying, but by a response to a gestalt tension” (2014: 159). Conscious trying, or what Merleau-Ponty calls personal acts, can occasion sensitivity to a particular context or gestalt in ways that, depending on one’s skills, initiate either absorbed coping or the learning process, or perhaps both in what Merleau-Ponty describes as “this back-andforth of existence.” In short, our deliberately seeking to develop new skills or to learn new ways to respond to unfamiliar situations can allow for expanded possibilities for skillful coping and can help us see or feel new ways in which our already acquired skills can be used. Our consciously committing ourselves to particular acts, paths, and projects influences the way the world is polarized and shapes the gestalt in which we find ourselves. Dreyfus’s work shows how we overlook much of human agency if we too quickly identify human agency with deliberative processes, self-conscious intentions, and reflectively self-aware experience of that agency. In his zeal to push in the other direction, however, Dreyfus may end up underappreciating the conscious, intentional, and deliberative aspects of human agency. My suggestion is that by more fully incorporating Merleau-Ponty’s notion of the two ways 231

Justin F. White

of existing into to Dreyfus’s account, we can better account for the whole of human agency, including the process of aspiration and the way in which we can consciously seek to be different, sometimes better than we currently are.

Related topics See Chapters 8 (on Heidegger) and 13 (on Merleau-Ponty).

Notes 1 For examples, see skillful (Dreyfus 2014: 84, 86, 88, 89, 94, 98, 116); absorbed (81, 94, 95, 96, 99, 101, 147–151); immediate (84); everyday (81, 90; 94, 96 107, 116); embodied (94, 105, 240). 2 In this vein, Crowell (2013: 257–259) describes the coping Dreyfus focuses on as “mindless.” He develops this term from Dreyfus (1991: 3): “At the foundation of Heidegger’s new approach is a phenomenology of ‘mindless’ everyday coping skills as the basis of all intelligibility.” 3 For one example of Dreyfus’s and McDowell’s exchange, see Dreyfus (2007) and McDowell (2007). 4 The term “affordance” comes from Gibson’s (1986). 5 Merleau-Ponty (1945/2012: 143–148) makes similar claims in his analysis of habit in Phenomenology of Perception. He suggests that the way a skilled organist quickly adjusts to an organ with additional or fewer keyboards highlights the difference between the bodily space we inhabit and objective space. 6 As noted earlier, there is significant disagreement about what is involved in human experience being conceptually articulated, representational, reflective, and the like. 7 Merleau-Ponty (1945/2012: 459) claims that our particular characteristics “are the price we pay, without even thinking about it, for being in the world.” 8 Dreyfus (2014: 35) thinks that when something goes wrong for the expert, “[h]e detachedly calculates his actions even more poorly than does the competent performer since he has forgotten many of the guiding rules that he knew and used when competent, and his performance suddenly becomes halting, uncertain, and even inappropriate.” 9 Of course, computers (or, increasingly, chess computer programs) now routinely beat expert chess players, but it was a number of years before Gary Kasparov, then World Chess Champion, was defeated by IBM’s Deep Blue in 1997. Although the eventual inaccuracy of Dreyfus’s pessimistic predictions has been seen by some as a simple repudiation of Dreyfus’s criticisms of artificial intelligence, the issues are more complex than that. For one, Dreyfus’s criticisms of early artificial intelligence neither earned him many friends in the AI community nor helped him find a receptive audience. As a result, the response to Dreyfus’s losing to MacHack, a chess playing computer, in 1967 led to more satisfaction than was probably merited for the computer beating an amateur level chess player. As Daniel Crevier (1993) puts it, “[T]ime has proven the accuracy and perceptiveness of some of Dreyfus’s comments. Had he formulated them less aggressively, constructive actions they suggested might have been taken earlier. In fact, it was almost uncanny for a non-expert in computer science to anticipate as early as 1965 the difficulties AI would run into, and to point out why!” For his part, Dreyfus recounts his version of the exchanges in various places, including in Mind over Machine (1986) and What Computers Still Can’t Do (1992). In his more recent, “Why Heideggerian AI Failed and How Fixing It Would Require Making It More Heideggerian” (Chapter 12 in Skillful Coping), Dreyfus analyzes and critiques more recent AI developments that have tried, sometimes explicitly, to address the worries Dreyfus raises in his earlier work. 10 In his work on expert performance, Anders Ericsson (2016) suggests that experts have more detailed and sophisticated conceptual maps than ordinary performers (see Peak: Secrets from the New Science of Expertise). Although this may appear to be in conflict with Dreyfus’s position, much of this depends on how we interpret Ericsson’s notion of conceptual maps. 11 Again, Dreyfus’s characterization of representation is a point of contention between Dreyfus and Searle. For more on Dreyfus’s view, see “Heidegger’s Critique of the Husserl/Searle Account of Intentionality” and “The Primacy of Phenomenology over Logical Analysis” (Dreyfus 2014).

232

Hubert Dreyfus



233

Justin F. White

York: Routledge, 2013) is a rich collection of papers devoted to the debate between Dreyfus and McDowell (including essays from both Dreyfus and McDowell). E. Fridland (2014), “They’ve Lost Control: Reflections on Skill,” Synthese 191, 2729–2750, and M. Brownstein (2014), “Rationalizing Flow: Agency in Skilled Unreflective Action,” Philosophical Studies 168, 545–568, both examine the nature of skill. Fridland is critical of Dreyfus’s position; Brownstein is sympathetic to Dreyfus’s approach.

234

18 LIFE IS AN ADVENTURE László Tengelyi’s phenomenology of action Tobias Keiling

László Tengelyi was a Hungarian philosopher who was educated and initially taught at Eötvös-Loránd-University in Budapest before becoming chair of philosophy at Bergische Universität Wuppertal, Germany, in 2001. Tengelyi published extensively in Hungarian, German, French, and English. From a focus on Critical philosophy and Phenomenology in his dissertation and early writings, his interests broadened to themes in ancient and medieval philosophy. In 2011, Tengelyi intervened in political debate in his home country, calling attention to state-led media campaigns against fellow philosophers such as Ágnes Heller. Shortly after completing his major World and Infinity: On the problem of a Phenomenological Metaphysics (Welt und Unendlichkeit. Zum Problem phänomenologischer Metaphysik, Tengelyi 2014), Tengelyi died unexpectedly at the age of 60. World and Infinity is the longest and most ambitious of Tengelyi’s works, tracing problems in phenomenology back to authors such as Aristotle, Plotinus, Duns Scotus, and Suárez and engaging with Post-Kantian German Philosophy. His perhaps best-known book is an exhaustive exposition of currents in contemporary French phenomenology, written in German with the translator Hans-Dieter Gondek (Gondek and Tengelyi 2011). “Action and Self hood,” Tengelyi’s contribution to the Oxford Handbook of Contemporary Phenomenology (2012a), also published in earlier German and French versions (2007a, 2009a), offers a concise presentation of his views in the philosophy of action. Although the central objective of World and Infinity, published a few years later, is to give a new, genuinely phenomenological account of the metaphysics of infinity, the book also includes a discussion of the metaphysics of agency that advances over the earlier account by providing a more elaborate discussion of human freedom. I will concentrate on these two texts, leaving aside numerous other writings, in particular Tengelyi’s much earlier book-length treatment of “life-history” (Tengelyi 1998, 2004). It is defining Tengleyi’s approach that rather than focusing the moment an action is initiated, a phenomenological account of action should take into view a much larger share of the agent’s experience. His approach also stands out for its attempt to relate two problems, namely the metaphysical puzzle regarding the compatibility between human freedom and causal determinacy and the question of personal and, specifically, narrative identity. In Tengelyi’s view, a phenomenological account of agency reveals these problems as closely connected: “the agent’s experience of his or her action . . . supplies the missing link between 235

Tobias Keiling

acting and recounting” (Tengelyi 2012a: 273). To make good on this claim, Tengelyi proposes to begin with a specific aspect of agency, namely the fact that actions have unintended consequences. As the consequences of an action unfold, however, the agent’s involvement shifts from intentional activity to a more passive stance: in the perspective of its emplotment in a story to be told, action considered in its totality will prove to be a mixture of activity and passivity. . . . Our actions are not only that which we perform but also that which happens to us. (Tengelyi 2012a: 265) Both with regard to the meaning of one’s actions and with regard to the constitution of the (narrative) self, Tengelyi’s central intuition is that the defining elements in the constitution of the meaning of action and self are experienced more as something that happens to the agent rather than something the agent deliberately intends and actively pursues. It is for this reason that the problem of unintended consequences is central to Tengelyi’s view. This allows for a straightforward exposition: I will begin by introducing his example case for agency and its implications for first-person involvement in agency in the first section. I will refer to this example throughout the next sections, presenting the increasingly complex discussions of unintended consequences in different contexts: Tengelyi’s critique of Davidson’s semantic theory (Section ‘From action to its agent. Tengelyi on Davidson’), the role of freedom and determinacy in action (Sections ‘Freedom and necessity: the meaning of unintended consequences’ and ‘Grounding grounds: on different ways to be free’), and the constitution of narrative identity (Section ‘My practice, my story, and I: narrative identity’). I end with a critical remark and indicate how his understanding of action should be further developed (Section ‘Conclusion: a case for truth’). Specifically, I will argue that Tengelyi’s account of human freedom calls for a priority of a normative account of the constitution of meaning over a narrative one. I develop such an account by drawing from Heidegger’s treatment of authenticity and truth, hoping to contribute to what Römer (2017: 129) calls a development of a phenomenological account of agency “in the direction of . . . a practical metaphysics.”

What is it like to be an agent? The case of Oedipus Tengelyi’s choice of example is characteristic for his approach. Unlike authors who orient their discussion to simple cases like intentional movement, Tengelyi chooses a particularly complex example: the story of Oedipus. Although fictional, Tengelyi takes it as genuine expression of the first-person experience of action, revealing the basic traits of human agency. Tengelyi provides no explicit justification why it is permissible to turn to an example from drama, but he gives context to this idea at the beginning of his contribution to the Contemporary Phenomenology handbook by citing a passage from Aristotle’s Poetics (1450a3): “ ‘the expression of action is a fable‘ or even ‘a plot’ [estin . . . tes . . . praxeos o mythos he mimesis]” (Tengelyi 2012a: 266). This quote not only establishes the conceptual link between action and narration that Tengelyi will explore in detail. It also implies that a fictional narrative, by virtue of having a plot, may just as well provide insight about the nature of action as a factual account. For phenomenology as study of the structures inherent in meaning (Crowell 2001, 2013: 9–30; Zahavi 2017: 18), the problem of unintended consequences comes into view through its implications for understanding what is defining the meaning or sense (sense in French, 236

László Tengelyi

Sinn or Bedeutung in German) of an action. This is how Tengelyi approaches the case of Oedipus. It reveals the “highly significant fact that the meaning or sense of an action does not remain unaffected by its unintended consequences, which therefore . . . must be acknowledged as subsequently revealed shreds of meaning that belong to the original deed” (Tengelyi 2012a: 266). The possibility of a dramatic change in the experienced meaning of an action is the most conspicuous feature of the example: At the notorious crossroads, Oedipus intends to merely kill an arrogant stranger, yet in retrospect, his deed is revealed as the killing of his father. Now, this description of this action also belongs to the meaning or sense of what really happened at that crossroads. . . . An action can thus detach itself from the intention of the agent. Nonetheless, it remains a deed of its originator [Urheber]. Can it be ascribed to him as an intentional action at all, its unintentional consequences are also ascribed to him. . . . Through his act of revenge, Oedipus ipso facto becomes a patricide. Here, the agent experiences his action in suffering its unwanted consequences. (Tengelyi 2014: 368–369) As an example of how agency is experienced, Tengelyi’s example highlights this dramatic reversal from an experience of active involvement in a deed considered by the agent to be just to an experience of passively suffering the consequences of the very same action, now perceived as morally reprehensible. To capture the different moments of first-person involvement in action, Tengelyi typically refers not to an intention (intention, Absicht) to act but to the broader notion of the “experience” (expérience, Erfahrung, Tengelyi 2012a: 271 and passim, Tengelyi 2007a, 2009a) of action. This is more informative than it might seem: emphasizing a moment of Husserl’s analysis, Tengelyi sees experience as defined by a form of coherence or agreement (Übereinstimmung). Only in contrast to this “basis of agreement” can diverging moments be recognized and expectations can be disappointed (Husserl 1970: 212). Following Hegel and Gadamer, Tengelyi takes the notion of experience to imply both negativity and conflict (Widerstreit, Tengelyi 2007b: 322). The story of Oedipus is a case in point. After setting up the discussion, I will return to the implication of this notion of experience in the concluding section. Another implication of this example is the specific nature of the limitation to knowing what one does. This limitation is to be distinguished from the question whether or to what degree actions involve conscious deliberation or propositional thought. In terms of the debate between Hubert Dreyfus and John McDowell (Schear 2003), whether his actions were ‘mindless’ or ‘minded’ is not at issue in the case of Oedipus: although Oedipus realizes he has been ‘blind’ at the time at which they were initiated, his actions have been the ‘minded’ result of deliberation and conscious choice. The example rather draws attention to another way in which the knowledge of one’s doing is restricted: at the moment an action is initiated, its meaning is to a large extent indeterminate and will remain “unstable, unsettled, and subject to alteration” (Tengelyi 2012a: 271) as it evolves. This evolvement, Tengelyi stresses, has nothing to do with first-person awareness or “with any progress of knowledge; it results solely from the unfolding of the—often unintended—consequences of the original action” (Tengelyi 2012a: 271). Offering the alternative of ‘mindless’ or ‘minded’ involvement therefore does not identify this defining feature of action. As Tengelyi emphasizes, the “temporal gap” (Tengelyi 2012a: 270) separating consequences from the original action may be considerable. Nonetheless, it is “undoubtedly” true that “even the consequences of an action that emerge considerably later than the original 237

Tobias Keiling

action enter into the sense or meaning of this action” (Tengelyi 2012a: 270). To bring out the historical scope of the process in which the meaning of action is constituted, Tengelyi (2012a: 270–271) gives an example from the history of philosophy: the consequences of Fichte’s appropriation of Kant’s Critical philosophy in his early writings cannot be gauged by looking at the work of these two authors alone. The accumulated effects of this action can only be measured with a view to German Idealism and its impact as a whole. The same structural feature thus defines the historical development of Post-Kantian philosophy and the story of Oedipus. Both cases reveal the essential incompleteness of any action’s meaning. Although this feature is much easier to recognize from the third-person perspective of a historian of philosophy or the author of a drama, the specific indeterminacy and incompleteness in the constitution of an action’s meaning is also defining first-person involvement in it. An agent’s experience of her deeds takes shapes within structures of intelligibility that are incongruent with and typically larger than the scope of the agent’s intentions and the consequences she expects. As Tengelyi aims to show, these structures are essentially narrative: the subject is not merely involved in her actions as an initiator of her intentional actions but also as the person living through the experience of her deeds. This makes her a subject of the experience of action and a possible narrator. The account of the self underlying this view will be discussed in Section ‘My practice, my story, and I: narrative identity.’ With regard to the meaning of individual actions, it follows from this idea that the link between different descriptions of actions is genuinely narrative: “it is the articulated relation between different descriptions of one and the same action that provides this action with a specifically narrative intelligibility” (Tengelyi 2012a: 273). Consider our example. Both ‘I killed a stranger on the roadside’ and ‘I killed my father’ are possible articulations of Oedipus’ first-person experience as laid out in the fictional account. As Oedipus never intended what the latter proposition describes, it provides an example of the notion that narrative intelligibility does not stop short of unintended consequences. As Tengelyi writes, narration is also capable of making actions “intelligible as events that have happened to us rather than as initiatives we have taken” (Tengelyi 2012a: 265–266). That consequences “supervene on . . . the initiatives that are taken intentionally” (Tengelyi 2012a: 266), however, does not mean that actions are properly described as events without any reference to intentions. As Tengelyi writes, it cannot be solely from the point of view of an external spectator that the relation between an action and its unintended consequences is established . . . it is only from the agent’s point of view that the relation between the original action and its unintended consequences becomes properly perceptible. For, unlike natural causality, this relation does not simply connect different events in the world; rather, it links the entire chain of events with the agent as well. It is, indeed, the experience gained—or to be gained—by the agent of the consequences of his or her action that establishes this relation. (Tengelyi 2012a: 271) Note that Tengelyi here makes a claim regarding both the methodological and the metaphysical role of the agent: the “link” between the agent and her actions is not merely “properly perceived” but also “established” by the agent through the way in which she “relates first-personally to the consequences of his or her action”; it is an “experiential relation . . . established by the agent from within” (Tengelyi 2012a: 271). If this is so, the metaphysical task of a theory of action is to make sense of this privilege of the agent building on a phenomenology of what it is like to act “from within.” But it is unclear what in the experience 238

László Tengelyi

of agency corresponds to that metaphysical authority: intention is a necessary but not sufficient element, and so is the subject’s role as a possible narrator of her actions. The first cannot make sense of unintended consequences, while the role of a narrator does not necessarily draw from the first-person perspective. What then defines the subject’s involvement as an agent? In “Action and Self hood,” Tengelyi gives an implicit answer to that question when indicating that establishing a metaphysical relation between an agent and her actions is an exercise of freedom. Because the paper focuses on narrativity, Tengelyi here merely indicates that it follows from this notion that human freedom must not be identified with “noumenal freedom”: if it were true that, as Tengelyi quotes Kant from the second Critique, action “does not recognize any temporal difference” (Tengelyi 2012a: 271), it would be impossible to make sense of the temporally extended constitution and evolvement of the meaning of action. As we will see in Section ‘Freedom and necessity: the meaning of unintended consequences,’ Tengelyi elaborates this argument in World and Infinity, where he gives a positive account of freedom as “temporally determined capacity” (Tengelyi 2014: 384). A good way to understand what motivates this move is to consider Tengelyi’s critique of Davidson.

From action to its agent. Tengelyi on Davidson Tengelyi sees a “phenomenology of experience in the life-world” as necessary supplement to “a Davidsonian ontology of impersonal events” (Tengelyi 2012a: 272). Although it helps to account for unintended consequences, Tengelyi argues that Davidson’s analysis miscontrues the ontological implications of the phenomenology of agency. The aim of this section is to show what drives this argument. For Davidson, unlike many canonical and recent phenomenological authors, the problem of unintended consequences is an important problem. Tengelyi takes over Davidson’s semantic framing of this problem, namely as a problem of how to properly describe action. In this framework, an action is properly grasped as such in a true description in intentional language. Tengelyi rephrases Davidson when defining agency in this vein: “the agent is the originator of all actions that can be ascribed to him under at least one description in which what he does is understood to be intentional” (Davidson 2001: 46; Tengelyi 2012a: 269). In this way, Tengelyi aims to integrate into a phenomenological account an idea first formulated by Anscombe (2000: 37–47), which Davidson calls “the conclusion on which our considerations all converge”: there can be “a welter of related descriptions [corresponding] to a single descriptum” (Davidson 2001: 53). Tengelyi emphasizes this point: Davidson’s “formula indicates that each new description of an action is a re-description of the original action” (Tengelyi 2012a: 270). Davidson’s definition allows dealing with unintended consequences in the way Tengelyi already anticipated: when an action is experienced not as one’s intentional doing but as an event the agent suffers, this change in the sense or meaning of the action amounts to a subsequent addition or modification of its true descriptions. What is unclear, however, is whether a semantic theory can express what Tengelyi called the “experiential relation” (Tengelyi 2012a: 271) between an agent and her actions and its metaphysical import. Since Tengelyi holds that it is possible to recast intentions within a semantic framework, a promising approach to integrate first-person involvement into Davidson’s theory may be to include narratability as well. If “the plot of a story to be told about an action results necessarily from a connection between two or more different descriptions of it” (Tengelyi 2012a: 273), it should be possible to recast narrative connections in this way. 239

Tobias Keiling

To see how successful this strategy might be, consider the following: ‘Oedipus kills his father although he does not intend to kill his father.’ Combining the two propositions in such a way provides a kind of miniature narrative. It reveals that there is not merely semantic continuity (words means the same in both propositions) and alethic continuity between these descriptions (they are both true) but also a genuinely narrative one. This continuity is brought out by the although, functioning as a kind of narrative connector. Nonetheless, connecting propositions in this way is insufficient to grasp what Tengelyi had identified as defining features in the constitution of the meaning of action. Borrowing a term from Ernst Tugendhat, Tengelyi emphasizes that every narration must be based not only on a connection between descriptions of action but on an “experiential way” (Erfahrungsweg, Tugendhat 1979: 25) connecting the agent’s experiences of actions. “It is the agent’s experience of her action that supplies the missing link in the connection between acting and recounting” (Tengelyi 2012a: 272). This idea restricts how descriptions of an action can be thought of as being linked in a narration. Thus in our example: even if it is true to say ‘Oedipus kills his father although he does not intend to kill his father,’ according to the story, there has never been a first-person correlate of that proposition in Oedipus’ experience. Giving voice to his experience requires the past tense: ‘I killed my father although I did not intend to kill my father.’ A more elaborate narrative would bring out the temporal extension of the different experiences: ‘I had intended to kill a stranger, then I learned. . ., now I know. . . etc.’ Only because these tensed, first-personal statements have in the time passed turned out to be true can the story of Oedipus be told by way of connecting third-personal propositions. But doing so reveals not only temporal extension but also indeterminacy and incompleteness as defining the process in which the meaning of an action is established, and these are the constitutive elements in the agent’s experience of action: “every agent takes at least the risk of experiencing something unforeseeable and unexpected turns that result from his or her action” (Tengelyi 2012a: 274). Tengelyi’s charge against Davidson’s theory is that it requires that the meaning of action be stable and determinate, which is not the way Oedipus experiences his actions. As Tengelyi remarks, Davidson “remains a follower of Quine” in that he “relegates the decision regarding the existence of beings to the logical semantics of our linguistic expressions” (Tengelyi 2012a: 268). Although first-personal involvement can be mirrored to some extent within the semantic framework, it cannot reach the ontologically defining element in the agent’s first-personal involvement, the experience Tengelyi refers to as “risking” unintended consequences. Tengelyi (2012a: 271–272) refers to Ricœur (1992: 73–87) as voicing a similar criticism, when he accuses Anscombe and Davidson of addressing the ‘what’ and the ‘why’ of action, but not the ‘who.’ Tengelyi reiterates this point here but identifies not narrative intelligibility as such, but the specific incompleteness and indeterminacy of action as feature defining the ‘who’ of the action. Despite Tengelyi’s initial reliance on Davidson’s treatment of unintended consequences, that it is incapable of expressing the defining feature in the experience of action motivates him to “draw some consequences that remain alien to the Davidsonian ontology of action” (Tengelyi 2012a: 271). These consequences concern the explanatorily basic and metaphysically defining level in the constitution of the meaning of action. Tengelyi concurs with the assessment that the problem of unintended consequences threatens a clear metaphysical distinction between intentional actions and events. “Intentional actions as such” (Tengelyi 2012a: 270) form no coherent set. But, Tengelyi argues, Davidson is moving too quickly from the observation that the “actions justly ascribed to a particular agent” do form “a determinate class of events

240

László Tengelyi

among other events in the world” (Tengelyi 2012a: 270) to the claim that agency can be given an “expression [that] is itself purely extensional” (Davidson 2001: 46). That defining action is possible only relative to a specific agent is a claim that “is less innocent than it may initially appear” (Tengelyi 2012a: 270). Assimilating actions to a “class of events” (Davidson 2001: 46) in the world fails to heed the difference between the consequences of actions and the consequences of events, as they emerge both within first-person experience and in semantics. This difference comes out most clearly with regard to how the meaning of an action continues to matter to the agent over time in a way other events do not. To understand the point Tengelyi wants to make, consider the contrast between the eruption of a volcano and man-made climate change. Because the first is a natural event, its causal effects, Tengelyi writes, “do not enter into the meaning or sense of the original event . . . The eruption of a volcano remains what it is, regardless of the more or less devastating effects brought about by it” (Tengelyi 2012a: 270). The effects of climate change, by contrast, are unintended consequence of human actions. It is a defining trait in their meaning that they are the results of having taken the risk of an unforeseeable outcome and of unexpected turns. You can regret having contributed to climate change, but you cannot regret the eruption of a volcano. Although both cases can be described as events having causally determined effects, in the latter case, this description does not link the causal effects to the agent having, perhaps unknowingly, run the risk of unintended consequences. What defines actions in contrast to natural events is that, should unintended consequences occur, they enter into the meaning or the sense of the original action . . . The difference between the two cases clearly shows that we are far from interpreting the consequences of actions in the same way as the causal effects of natural events. (Tengelyi 2012a: 270) Although Davidson (2001: 54), too, observed that the criterion for taking an event to be an action is “whether we can attribute its effects to a person,” he failed to recognize that this rules out defining the metaphysical role of the agent in causal terms alone. As Tengelyi stresses, he aims for a “holistic view of action . . . according to which a single description never exhausts the content of a deed” (Tengelyi 2012a: 273). This holds true not only for descriptions in intentional language (as Davidson argues as well) but also for descriptions in causal terms. Although Tengelyi does not put it this way, he apparently rejects the claim that “the primary reason for an action is its cause” (Davidson 2001: 4). I will return to this point in Sections ‘Freedom and necessity: the meaning of unintended consequences’ and ‘Conclusion: a case for truth’. In Tengelyi’s view, Davidson’s theory is therefore incomplete not because it may not be extended to mirror aspects of first-personal involvement such as intention or narrativity but because it cannot give an account of temporal extension, indeterminacy, and incompleteness, which are the defining features of the meaning of action. Given Tengelyi’s brief remarks on human freedom in “Action and Self hood,” this disagreement is not merely methodological and not merely concerns the role of first-personal accounts. If Davidson’s theory cannot make sense of what it means to risk unintended consequences, it cannot connect to a metaphysical account of human freedom that builds on this phenomenology of agency. In the development of Tengelyi’s own theory, an account of freedom is what the treatment of action in World and Infinity adds to the earlier work on narrativity. Given the emphasis on freedom involving an “experiential relation,” as discussed in Section ‘What is it like to be an agent? The

241

Tobias Keiling

case of Oedipus,’ one might expect Tengelyi to draw from first-personal accounts as, for instance, Kriegel (2015: 205–244) does. Tengelyi remains closer to his earlier work, however, continuing the discussion of Oedipus, focusing the problem of unintended consequences, and tying in numerous authors from the history of philosophy.

Freedom and necessity: the meaning of unintended consequences To describe the metaphysical implications of unintended consequences, Tengelyi begins with a comment Aristotle makes in the Nicomachean Ethics (114b23, 1114b32–1115a1). Regarding dispositions (hexeis), Aristotle holds that the individual is only “co-originating” (synaitios) her dispositions. Only the initiative to act, the “beginning” (arche) of the action is in our control. Tengelyi follows Ricoeur (1992: 133) in extending this notion to actions in general: humans are not in full control of their deeds but only their “co-originator” (Miturheber, Tengelyi 2014: 370). With regard to the problem of unintended consequences, this allows Tengelyi to distinguish degrees of causal involvement: “although the freedom of action is the cause [Ursache] of the intentional action, it is only the partial cause [Mitursache] of the entire action, once the unintended consequences have entered into it” (Tengelyi 2014: 372). As the action unfolds, the agent’s spontaneity becomes less and less relevant and other determining factors contribute more and more to its meaning. This has an important negative consequence for any account of freedom: “The reason why the notion of a noumenal freedom, in the Kantian sense of the word, has to be rejected is precisely its inability to account for this complicity with reality” (Tengelyi 2012a: 282). In the chapter on action in World and Infinity, Tengelyi aims to spell out an alternative view. The larger metaphysical system Tengelyi develops starts from a number of basic experiential givens with far-reaching ramifications. Taking up a Husserlian notion, Tengelyi calls these “primordial facts” (Husserl 1973: 385; Urtatsachen, Tengelyi 2014: 14, 180–227), counting the ego, the world, intersubjectivity, and history among them. Not susceptible to proof or further analysis, these basic or primordial facts have been revealed in the history of philosophy. But this is not where their authority derives from. Rather, a phenomenological treatment of classical metaphysical problems reveals them as being indispensable for making sense of our experience. Although contingent in origin, all experience implies their “factual necessity” ( faktische Notwendigkeit, Tengelyi 2014: 189; Husserl 2014: 83). Rather than designating them categories of judgment or Heideggerian existentials, Tengelyi coins the new expression “experientials” (Experientialien, Tengelyi 2014: 197) to describe the status of primordial facts as phenomenological categories. In Kantian terms, experientials, despite being “categories of experience” (Tengelyi 2014: 198) are akin to principles of reflective rather than determinative judgment (Römer 2020). Tengelyi emphasizes that in contrast to Kant’s critical system, the “open transcendentalism” first developed by Husserl does no longer assume the “coherence of experience . . . as a necessary consequence of self-consciousness” (Tengelyi 2014: 359). Rather, “all search for rules, order, and unity within the ‘rhapsody of perceptions’ ” is found in a transcendental reflection of experience and thus “delegated to the competence of reflective judgment” (Tengelyi 2014: 359). Although Tengelyi presents this notion of freedom as primordial fact as transformation of Kant’s compatibilist position, recasting it within this system of a phenomenological metaphysics leads to a number of shifts from this position. Most importantly, Tengelyi denies a separation of a nexus of free determination initiated through the spontaneity of the subject on the one hand and the causal, mechanical order of 242

László Tengelyi

the world on the other. Rather than as somehow emerging on two parallel planes at once, experience shows that action is to be defined as the ongoing intersection of spontaneity with the causal order. The reasoning in World and Infinity here elaborates the argument from the earlier essay on action: while the idea of a parallelism of two separate domains of freedom and nature reduces freedom to a “merely intelligible, noumenal idea” (Tengelyi 2014: 372), that there are unintended consequences to an action shows that only assuming the constant interaction between the causal order of nature and human spontaneity can yield a plausible notion of human action. Therefore, any “parallelism” of freedom and necessity is to be rejected: The moment we initiate an action, there is not yet a parallel: only once the action has already been initiated is the causal mechanism of the world influenced in such a way as to cause the consequences of the action that come to change sense and meaning of the original action. Once these mechanisms have been set in motion, however, there can no longer be a parallel: now, events are already taking place that have not been intended in the original initiative to act. The initiative to act and the causal mechanisms of the world are constantly meshing; they never run parallel to each other. (Tengelyi 2014: 372) Tengelyi further shifts away from the Kantian picture by introducing the idea that there are other forms of determinacy (Determination, Bestimmung) than causal determination. Specifically, Tengelyi draws from the work of the Marburg neo-Kantian Nicolai Hartmann (1882–1950). Hartmann defines freedom of action as the superposition of different “types of determination” (Determinationstypen), that of causal determination (“causal nexus,” Kausalnexus) and that of volitional determination in the teleological historical process (“teleological nexus,” Finalnexus) in particular. As Hartmann says, “freedom is only possible where, in one world, at least two types of determination are superimposed one upon the other” (Hartmann 1932: 64). The teleological nexus is needed for freedom of choice not to remain merely negative (Hartmann 1932: 48). As Tengelyi emphasizes, Hartmann here follows Hegel’s notion of a “cunning of reason” (Hartmann 1960: 474; Hegel 2010: 663), “the idea that freedom begins where reason makes natural causality serve its purposes by using a natural object as a means to achieve its own end” (Tengelyi 2014: 373). Although Hartmann conceives the causal nexus as a “totality,” this totality is underdetermined, “never absolutely closed; it does not prevent the addition of new determining elements—if there be such” (Hartmann 1932: 55). Intentional action and the teleological development of its meaning come to fill the gaps in the causal order without thereby disrupting causal efficiency: “the process is never broken by such additions, it is only diverted” (Hartmann 1932: 55). In Tengelyi’s system, this idea of causality as an underdetermined totality anticipates the phenomenological account of infinity developed in later chapters of his book (Tengelyi 2014: 435–556). Tengelyi substantially modifies Hartmann’s account, however, by rejecting the claim that there is a hierarchy between the different forms of determination (Tengelyi 2012b). Although affirming the idea of a “redirectability of causal chains” (Tengelyi 2014: 375), Tengelyi denies that the aggregate of free choices, which forms the teleological nexus, supervenes on the only partial determination of action in the causal nexus. Experience shows that causal and volitional determination can and do in fact come to conflict. Again, the fate of Oedipus is a case in point: because killing his father was not what Oedipus intended, it is not the result of informed choice; rather than its meaning being exhausted by its being 243

Tobias Keiling

located in the teleological nexus, at least one description defines it as primarily causally determined event. As Tengelyi writes, “the phenomenon of unintended consequences shows that natural causality can always come to resist the teleological power of control and disposal. Natural causality often interferes with our free initiatives” (Tengelyi 2014: 375). This interference explains the phenomenology of agency identified in the introductory section as defining the case of Oedipus: the break with the specific intelligibility provided by the final nexus leads to Oedipus’ experience of a radical reversal of the meaning of his deeds. In Tengelyi’s understanding, Oedipus here falls back on the intelligibility provided by the causal nexus. This discussion of the case of Oedipus is notably different from that of both Hegel and Schelling. Where Hegel finds in Sophocles’ tragedy an eventual “reconciliation of the powers animating action” (Hegel 1975: 1215) and Schelling an “equilibrium between . . . freedom and necessity” (Schelling 1989: 255), Tengelyi sees a resurgence of natural causality and an unresolvable conflict between different forms of determination. In the earlier paper on narrativity, Tengelyi comments on and explicitly rejects Schelling’s interpretation of Oedipus Rex. It is central for this reading that, when Oedipus decides to punish himself after suffering the unintended consequences of his deeds, the teleological order is restored and eventually prevails. As that happens, Tengelyi summarizes, “freedom recuperates the unintended consequences of its actions from necessity (or fate) by assuming responsibility for them” (Tengelyi 2012a: 281). Tengelyi acknowledges that Schelling’s reading presents “a serious attempt to resolve the fundamental dilemma of the freedom of action” (Tengelyi 2012a: 281). Yet when Schelling describes Oedipus’ deeds paradoxically as “unavoidable transgression” (Schelling 1989: 254), he restates rather than dissolves the conflict between fate or necessity on the one hand and the realm of freedom on the other. Rather than as its confirmation sub specie aeternitatis, Tengelyi interprets the case of Oedipus as undermining the idea of a supervenience of the teleological order. The idea of an “equilibrium between freedom and necessity” must, for a “finite being” at least, remain of “speculative character” (Tengelyi 2012a: 281). It is no trait of the phenomenology of agency. These remarks highlight Tengelyi’s divergence from the compatibilist position: if a “conflict” between freedom and causality can always emerge, teleological processes do not simply fill the gaps of indeterminacy left in the causal nexus. Although constantly “meshing,” the different types of determination remain “heterogeneous” (Tengelyi 2014: 375). Rather than as threatening the subject’s freedom or as revealing an antinomy, however, Tengelyi takes this conflict as a confirmation of the metaphysical outlook phenomenology advocates: The possibility of such conflicts speaks most clearly for the claim that the teleology of action can be understood as an experiential category that—similar to other experientials—expresses a tendency of coherence [Einstimmigkeitstendenz]. The fact of a conflict between the two types of determination does not contradict this view but rather lays its ground. For where there is but a tendency towards coherence, naturally there is conflict as well. (Tengelyi 2014: 375) The negative point to take away from the example of Oedipus, therefore, is that a phenomenological account of merely factual necessities must not make the assumption that there is a single form of homogenous coherence that will eventually come to fully determine the meaning of an action. Again drawing from Husserl, Tengelyi labels the strongest form of determination to be identified in experience an Einstimmigkeitstendenz (Tengelyi 244

László Tengelyi

2014: 194–200, 303–365, 375), a precarious tendency toward coherence or harmony. Even the notion of a “world” does not designate a determinate totality with necessary ontological structures. In a phenomenological understanding, “world” merely “gathers in the form of a single concept the tendency-character of all tendencies towards coherence [Tendenzcharakter aller Einstimmigkeitstendenzen]” (Tengelyi 2014: 547). Modifying the metaphysical picture underlying the account of human freedom has the important implication that, pace Schelling, Hegel, and Hartmann, the coherence generated by intentions does not eventually assume the form of a teleological order that fully determines the meaning of actions over the course of history. Rather than being a hierarchy of different forms of determinacy, the problem of unintended consequences reveals that freedom must be understood as in both continuity with and contrast to several heterogeneous forms of determinacy. The aim of the next section is to show how Tengelyi develops what may be labeled a plural-compatibilist position by drawing from Heidegger’s philosophy.

Grounding grounds: on different ways to be free Recall from the above discussion that, even on a historical scale, the meaning of action remains indeterminate. While the earlier “Action and Self hood” merely discussed this indeterminacy with respect to an already initiated action, its evolvement and possible narration, the view laid out in World and Infinity (2014: 375–392) brings this indeterminacy into view as a potential for future action: central to Tengelyi’s discussion of freedom is our “capacity to initiate an action [Selbstanfangenkönnen]” (Tengelyi 2014: 384). Importantly, a full account of this capacity requires that indeterminacy not be regarded as the mere “redirectability” of causal chains. As Tengelyi now highlights, several forms of relative indeterminacy can be identified in the experience of agency. In particular, indeterminacy is also found with regard to the moral evaluation of action. To bring out that there is an indeterminacy relative to moral choice, Tengelyi again turns to a number of authors. He finds the strongest expression of this idea in Schelling’s account of human freedom as the possibility of choosing between good and evil, which Tengelyi takes to reveal a “strong indeterminism” (Tengelyi 2014: 200) in the moral determination of acts. Similar in Hartmann: although the teleological order eventually supervenes on moral indeterminacy, Tengelyi concurs with Hartmann’s idea that moral choice is marked by a genuine “indeterminism of values” (Hartmann 1932: 225). In contrast to these views, however, moral determination for Tengelyi has the specific form of an immediate demand or appeal (Anspruch). As Tengelyi writes, it belongs to the “mode of determination specific to appeals of ought [Determinationsweise verschiedener Sollensansprüche] to leave different options open, or rather: to open them up at all” (Tengelyi 2014: 381). Importantly, however, Tengelyi does not assume a quasi-Platonic domain of value, as Hartmann does, but moves, somewhat quickly, from a Neo-Kantian value ethics to an ethics of responsivity in the wake of Levinas. Tengelyi describes the situation in which the agent is confronted with moral choice as that, in the words of Bernhard Waldenfels (2016: 356), of a “conflict of simultaneous appeals,” where the content of these appeals is not further determined. Here’s Tengelyi’s own account: The leeway of action is indeed a space of open alternatives. But the possibilities for action offered within this space are not equivalent. They urge the agent, appeal to him and demand to be realized. They each lay claim to determine his will. Thus, they enter into conflict. (Tengelyi 2014: 381) 245

Tobias Keiling

Although I cannot discuss the plausibility of this moral phenomenology here, it is clear that the moral form of determination Tengelyi eventually opts for is markedly different from that of causality already in that its main form is not that of laws but of situational appeals to action. With regard to the metaphysical underpinnings of this moral phenomenology, this implies that the indeterminacy manifesting itself varies according to the form of determinacy with which it contrasts. This leads to the decisive implication that the agent is confronted with both kinds of relative determination and their corresponding forms of indeterminacy at once. She finds herself not only partially determined by the “relative strengths [Kräfteverhältnis]” (Tengelyi 2014: 380) of different causal forces but also within the “field of appeals [Anspruchsfeld]” (Tengelyi 2014: 383) defined by the different moral implications of her options for action. Having identified these different forms of determinacy and corresponding indeterminacy, Tengelyi returns to the story of Oedipus. Once we see human freedom as temporally extended and abiding capacity to initiate action and heed the heterogeneity of different forms of determinacy, a more complex account becomes possible. Rather than seeing it as embodying a single tragic conflict between freedom and necessity, as did the authors discussed in the last section, Tengelyi’s account aims to make the story intelligible as a series of decisions with which Oedipus reacts to the evolvement of his original deed. In this way, a more fine-grained account becomes possible: although we cannot always remain masters of our entire deed, we can, within certain limits, remain capable of reacting to the experience of unintended consequences. We are free to change and correct our initial plan for action or to even alter it as a whole. This idea, that we can still interact with a particular event of action [Handlungsgeschehen], even if we cannot control it, essentially belongs to our idea of being able to initiate an action [Selbstanfangenkönnen]. When referring to this idea, we actually always mean an always again renewable capacity. (Tengelyi 2014: 384) As Tengelyi highlights, thus drawing out the consequences of the temporal situatedness of human freedom “looks more harmless than it actually is” (Tengelyi 2014: 384). For in order to react to the experience of unintended consequences, we must be able to recognize, distinguish, and make use of the different forms of determinacy and their corresponding forms of indeterminacy. Being able to first intentionally cause, then suffer from, and then react to unintended consequences means to have the capacity to not only initiate, but also “deviate from the line of determinacy within the teleology of action that we are following at that moment,” choosing to “follow another route” (Tengelyi 2014: 384). If the case of Oedipus involves both the nexus of natural causality and the teleological nexus as two heterogeneous forms of determinations, as discussed in the last section, reacting to unintended consequences cannot be explained in reference to a teleological account of action alone. It also means to have the capacity to isolate a merely causally determined line of determination and find in it indications for how to transform the teleology of action we are currently following. These capacities point to a deeper meaning of freedom. Only a living being that can not only initiate an action but can also take distance from the teleology of action and react to natural causality as it interferes with this teleology is acquainted with something like natural causality and the teleology of action at all. (Tengelyi 2014: 384–385) 246

László Tengelyi

This is the most complex description of the case of Oedipus that Tengelyi gives. It again raises the standard for what a phenomenological account of human freedom must cover: it must not only make sense of the experience of both active and passive involvement in action. It must also show what it means to recognize, distinguish, and make use of the different ways in which the space of possible action is partly determinate, partly indeterminate. A finegrained account of the temporal and motivational sequence of choices involved in dealing with unintended consequences reveals that free choice requires being able to recognize different forms of determinacy in order to then choose, within the domain of the pertinent indeterminacies, what to do now. The agent does more than function as a cogwheel, as it were, in the meshing of different forms of determinacy. Tengelyi sees having identified this “deeper meaning of freedom” as “turning point” (Tengelyi 2014: 384) of his entire theory. In particular, Tengelyi draws the conclusion that freedom should be identified neither with spontaneity (with respect to the causal order) nor with moral choice (in responding to the conflicting moral appeals). Rather, an account of freedom must reach deeper than these views by describing the very “openness” (Offenheit) for interacting with the different forms of determination at all. Shifting from Hartmann’s notion of “determinations” (Bestimmungen) to Heidegger’s talk of “reasons” or “grounds” (Gründe), Tengelyi calls this freedom an “openness for relations of grounding in the world as such [Gründungszusammenhänge in der Welt überhaupt]” (Tengelyi 2014: 385). In the concluding passages of his treatment of action, Tengelyi presents Heidegger’s account of freedom from “The Essence of Ground” and the 1928 Leibniz-lecture. Tengelyi fully endorses this view. I conclude this section by outlining three major aspects before returning to the question of narrative identity. One aspect of Heidegger’s view on freedom is that the concept of a reason or ground is based on a prior interpretation of the situation in which an agent finds herself. Importantly, this interpretation is itself an exercise of freedom in the deeper sense. As Tengelyi’s puts this point, summarizing Heidegger in his own idiom, “a justification . . . is only possibly within a certain world projection [Weltenwurf ] and in reference to an each-time different being-in-theworld as factual being-captivated by entities [Eingenommenheit im Seienden]” (Tengelyi 2014: 386). Heidegger calls this basic exposure and the subject’s response to it “transcendence” (Transzendenz), emphasizing that this is what he means by freedom (Heidegger 1984: 185). Because freedom qua transcendence is what allows a subject to first of all establish an interpretation of what it means to give a reason, Heidegger in a recurring phrase calls it the “ground of ground” (Grund des Grundes) (Heidegger 1984: 214, 1998: 134). This phrase aims to bring out that the idea of freedom as transcendence is deeper because it is explanatorily more basic than notions of freedom such as that of spontaneity relative to causal determination. As Heidegger stresses, to understand causal determination to then use its relative indeterminacy to initiate an action depends on a prior use of freedom qua transcendence: “Only because freedom consists in transcendence can freedom make itself known as a distinctive kind of causality in existing Dasein” (Heidegger 1998: 217). This may seem odd, as it seems to imply that causal determination is itself a form of freedom. Heidegger indeed claims that freedom qua transcendence is necessary for causality to “make itself known” (sich bekunden) in experience. But what this means is simply that to recognize the causal order as defining a trait in our experience requires the minimal commitment of taking this experience as truly revealing a feature of reality. This commitment is already an exercise of freedom in the deeper sense of transcendence. Heidegger is thus advocating here what can be labeled a normative understanding of phenomenology, where each form of phenomenal presence, including in the recognition of causality (including 247

Tobias Keiling

in perception, Doyon and Breyer 2015), involves a response to some normative claim. Although Tengelyi does not refer to these authors, in relying on the account of freedom as transcendence, he situates his discussion of action within readings of Heidegger developed by Haugeland (2013), Crowell (2013), Golob (2014), and others that focus on normativity in this sense. I shall have more to say on what follows from this for Tengelyi’s view in the concluding section. A second implication of Heidegger’s account of freedom is that there are different notions of “reason” or “ground” irreducible to a single, coherent notion. This point has already been encountered in the last section with regard to the discussion of the different forms of determinacy when Tengelyi denied a hierarchy among them, referring to Oedipus’ as a case in point. As Tengelyi emphasizes now, that reason-giving is not univocal has already been recognized in Aristotle’s account of the four causes or Schopenhauer’s treatise on the fourfold root of the principle of sufficient reason, to both of which Heidegger refers. Tengelyi highlights that Heidegger was always “careful to keep in view a manifold of different relations of grounding” (Tengelyi 2014: 385). This pluralism of forms of grounding runs parallel to Heidegger’s discussion of ontology. Although Heidegger never clearly endorses ontological pluralism (Keiling 2018) and an according grounding-pluralism relative to different understandings of the world (Weltentwürfe), Tengelyi seems to do so (Keiling 2020). A third important element in Heidegger’s account of freedom qua transcendence is its reference to the manifestation of entities. As Tengelyi stresses, to transcend for Heidegger does not mean to access “some kind of ‘beyond’ or ‘backworld’ [Hinterwelt]” (Tengelyi 2014: 386). Rather, “that towards which the subject transcends is what we call world” (Heidegger 1984: 166). Specifically, freedom qua transcendence engages with individual entities and their respective forms of manifestation. It takes the form, Heidegger says, of “letting oneself be bound [Sichbindenlassen]” in such a way that one becomes beholden to the “measure: beings as they are [das Seiende, wie es ist]” (Heidegger 1995: 342). Haugeland (2013: 201) aptly calls this the free subject’s “beholdenness to entities.” Despite the plurality of different forms of grounding, the specific situation and the entities encountered there emerge as the each-time relevant normative correlate. As Tengelyi highlights in reference to another quote from Heidegger’s lectures, “being open to beings” (Heidegger 1995: 342) and an “openness for relations of grounding in the world as such” (Tengelyi 2014: 385) are co-constitutive. Note that this view concurs with an element in Tengelyi’s critique of Davidson. While Davidson aimed to show “how to construct a primary reason” (Davidson 2001: 4) for an action and that this reason is its cause, Tengelyi’s reference to Heidegger’s account of what it means to be a reason or ground implies that, although an action may have a primary reason, it is not identical to its cause. Where Davidson aimed to show that any rationalization of action depends on its rationalization in causal terms, Heidegger and Tengelyi hold that all rationalization of action depends on the exercise of freedom qua transcendence, of which a rationalization in causal terms is only one form. Unlike views “inspired by the later Wittgenstein” (Davidson 2001: 10), which Davidson criticizes, this position does not imply that reasons cannot be causes; they are. It only implies that the question “how reasons explain actions” (Davidson 2001: 10) admits to different answers, depending on prior assumptions about how to understand the action; pointing to a cause as its reason is only one possible rationalization of the action. But even if neither their description in intentional nor that in causal terms exhausts the meaning of what actions are, Tengelyi, like Davidson, holds that actions are entities (Tengelyi 2012a: 269). Given the third aspect of Heidegger’s view, this raises the question how the relation of these entities to the subject that is free in the sense of transcendence is to be understood. What could it mean to be “open” or “beholden” to 248

László Tengelyi

actions? In the setup of Tengelyi’s theory, this question directly leads to describing the role of narrative intelligibility in making sense of these entities. I will sketch Tengelyi’s discussion of narrative identity before returning to the truth of actions in the concluding section.

My practice, my story, and I: narrative identity Recall that Tengelyi’s paper on narrative identity predates the account in World and Infinity and Tengelyi’s discussion and endorsement of Heidegger’s notion of freedom as transcendence. While there is continuity between the two texts in Tengelyi’s treatment of the problem of unintended consequences, I will argue that it is not clear that the earlier treatment of narrativity can be seamlessly integrated with his later discussion of freedom. This section aims to spell out Tengelyi’s view before discussing it in the next section. Recall from Section ‘From action to its agent. Tengelyi on Davidson’ that Tengelyi took Davidson’s theory to be incapable of accounting for the first-person character of the experience of action. Rather than referring to a semantic theory alone, Tengelyi argued that it must build on a metaphysics and a phenomenology of agency defined by the actual course of first-person experience, its “experiential way through the life-world” (Tengelyi 2012a: 274). This idea was central to Tengelyi’s charge that Davidson’s theory was incomplete: if we take seriously the task of deciding, to put it in Davidson’s terms, ‘what makes a bit of biography an action’, we shall have to pass over from a Davidsonian ontology of impersonal events to a phenomenology of experience in the life-world. (Tengelyi 2012a: 272) Although Tengelyi does not put it that way, I take this to mean that the relations between the descriptions a semantic theory recognizes should be seen as dependent on or founded in the meaning encountered in the first-personal experience of action and of taking the risk of unintended consequences in particular. How is such dependence to be understood? Tengelyi’s answer to that question is given not in a more elaborate account of how we narrate action such as the one I sketched in Section ‘From action to its agent. Tengelyi on Davidson.’ Somewhat indirectly, Tengelyi conceives the link between first-personal and third-personal elements in the meaning of an action by means of an account of the self. Central to that account is the idea that “the constitution of self hood is . . . a multi-layered process” (Tengelyi 2012a: 280). Tengelyi identifies three such layers building on one another: passive self-constitution, reflective self-awareness, and narrative identity. Tengelyi adopts this model to counter a certain tendency he finds in the theory of narrative identity, the “danger of reducing life and self hood to fictional constructions . . . to construct a meaningful life-history and, so to speak, invent an appropriate self hood” (Tengelyi 2012a: 275). Although Tengelyi recognizes that authors such as Ricoeur (1991: 32) take the constitution of narrative identity to be a discovery rather than a construction, he finds lacking a more detailed account of how such discovery of a pre-narrative self occurs: The idea of a narrative identity clearly implies that our selfhood is somehow given before it comes to be specified and determined by narratives. . . . However, it is by no means clear how a narrative view of the self could account for the constitution of a pre-narrative self-identity. There is no exaggeration in saying that this is a fundamental difficulty with which the proponents of the theory of narrative identity find themselves confronted. (Tengelyi 2012a: 276) 249

Tobias Keiling

The three-layered model of the self Tengelyi presents aims to provide a solution to this difficulty. The idea is straightforward: that the narrative identity of the self is discovered simply means that it is founded on the explanatorily more basic layer of self-awareness, which is founded in turn in the layer of passive self-constitution, as Husserl describes it in the Analyses of Passive Synthesis (Husserl 1966: 203). What connects the passive self with the layer of reflective self-awareness is memory or recollection, while narration or articulation connects the self consciously aware of its past with its narrative identity: “the layer of reflective self-awareness is articulated by the stories about one’s life” (Tengelyi 2012a: 280). Tengelyi takes this model to limit the “validity” (Tengelyi 2012a: 279) of theories of narrative identity, providing a safeguard against the danger of their “constructivist interpretation” (Tengelyi 2012a: 276). In the case of recollection, even “an active search for the self ” does not construct or invent the self but discovers it, and “what is discovered” is precisely “a passively constituted self ” (Tengelyi 2012a: 278). As Tengelyi stresses in opposition to fantasy, recollection supposes the existence of the object; it is, consequently, a positional act, and as such it necessarily remains in touch with reality. There is no conscious past—and therefore no life-history—without this adherence to reality. However, the relationship with reality that is characteristic of recollection is by no means constituted by story-telling; on the contrary, it precedes narration . . . the adherence to reality that marks the rememoration of one’s own conscious past is constituted in the passive sphere. (Tengelyi 2012a: 279) The “adherence to reality” relevant for recollecting facts thus carries over into the higher layer of the narrative self: “telling stories about one’s own life,” Tengelyi writes, “can only contribute to the constitution of one’s own self, if it preserves the adherence of recollection to reality” (Tengelyi 2012a: 279). At first, this account seems to have lost sight of the central idea that narration and action are essentially intertwined: one can recollect one’s past perceptions just as well as one’s past deeds, both belong to the experience of the passively constituted self. If narration is only inter alia founded in recollection, it is unclear how Tengelyi can make good on the claim that structures of narrative intelligibility play a privileged role in the constitution of the meaning of action. So what does it mean when Tengelyi takes actions to be essentially narratable? To answer that question Tengelyi returns to the phenomenon of unintended consequences and the case of Oedipus, contrasting an understanding provided by an action as part of a practice with that provided by the notion of narratibility. Both can articulate the “relation between the different descriptions of one and the same action,” but only one of them “provides this action with a specifically narrative intelligibility” (Tengelyi 2012a: 273). What sets both kinds apart is the way in which they relate different descriptions of actions. Tengelyi’s example of a practice is what is involved in the job of a philosophy professor. Here, different descriptions make the same action intelligible as part of what it means to be a professor of philosophy: the professor (i) presents a university lecture on the rise of Greek metaphysics, (ii) he sets forth his views on the relationship between Plato and Aristotle, (iii) he gives a course on the history of philosophy, (iv) he prepares his students for an examination, and (v) he insists on the standards of his profession. (Tengelyi 2012a: 273) 250

László Tengelyi

These different descriptions make the action only intelligible as part of a practice because “the activity in question is intentional in all of the five descriptions”: “the internal coherence of a practice . . . assures the intentionality of action” (Tengelyi 2012a: 273). Because they grow out of the agent’s intentions, Tengelyi argues, practices constitute a structure of intelligibility that is “at least in principle . . . transparent” (Tengelyi 2012a: 273). For this reason, doing something as part of a practice typically does not “give rise to any surprise” and does not “deserve the attention of the story-teller” (Tengelyi 2012a: 273): “the network of practices, with which our life-world is permeated, protects us against the radical turns that our actions are apt to take” (Tengelyi 2012a: 274). This is different in the case of genuinely narrative intelligibility, which typically includes “the relation between an action and its unintended consequences” (Tengeyi 2012a: 273). It is precisely the feature revealed in the above discussion as defining cases like that of Oedipus, “the fact that every agent takes at least the risk of experiencing some unforeseeable and unexpected turns that result from his or her action” (Tengelyi 2012a: 274), that cannot be made intelligible by the intentional and principally transparent structure of practices but requires genuinely narrative intelligibility. Only the more encompassing structure of narrative intelligibility allows the self to see her deeds as part of her personal identity and life-history. A practice not only “rests on a systematic, and to some extent, impersonal articulation of certain intentions to act” (Tengelyi 2012a: 273), the “network of practices” defining the life-world even effects a certain “depersonalization” (Tengelyi 2012a: 274). As long as an action is considered merely as taking part in a practice, it does not contribute to the constitution of a personal self on the level of her narrative identity. This idea is certainly controversial, and I will discuss it in the concluding section. But it does tie in with several strands in the above discussion. One possible consequence, which Tengelyi does not draw explicitly, is that one should distinguish a broader sense of action, which includes all intentional involvement of a practice (all rule-following), from action in a more restricted sense. Action in this genuine sense happens outside the safeguards of an established practice, it requires acceptance of the risk of unintended consequences, and this is what makes it narratable. Arendt, for instance, whom Tengelyi credits with first pointing out “that action gives rise to a story to be recounted, that it is, moreover, in search of a narrative” (Tengelyi 2012a: 272), makes a similar distinction between a doing that, like production (poiesis), is merely an anonymous partaking in a practice, and a genuine sense of action (Handlung, prâxis), in which the agent is involved in such a way that she can emerge as a member of a community of agents (Arendt 1958, 136–247; Tengelyi 2004, 42–52; Loidolt 2018). Another implication, which Tengelyi does recognize, is that only narrative intelligibility can be understood as adequate articulation of the first-personal experience of the agent. This confirms a point discussed in Section ‘From action to its agent. Tengelyi on Davidson’: with regard to the self’s involvement in action, the best a semantic approach can do is identify the register specific for talking about certain practices and describe what it means for any agent to take part in these practices. But it cannot get a grasp of the first-personal experience of the indeterminacy typical for the constitution of the meaning of action and the indexed specificity of what it is like for me to act in this ways. The metaphysical possibility of unintended consequences then not only confirms the epistemological priority of first-personal knowledge; it is also a necessary requirement for the constitution of a genuinely narrative self. This is how Tengelyi puts the point: because actions can have unintended consequences, the “process of depersonaliziation” typical for practices is necessarily bound to come to an end. For even when he or she is entangled in a network of different practices, an agent necessarily remains the originator of his or 251

Tobias Keiling

her actions. And this fact connects the different actions in his or her life, even if they do not belong to the same practice. Such transversal relations preserve the traces of an experiential way through the life-world, which is characteristic of personal history. Such an experiential way can, however, never be entirely immune to the unintended consequences of actions. That is why action, as we may put it by converting the Davidsonian phrase, remains always a bit of biography. (Tengelyi 2012a: 274) The consequence Tengelyi draws from this is that “the narrative view of selfhood” must be refined “in order to account for the possibility of radical turns in life-history” (Tengelyi 2012a: 280). Given Tengelyi’s account of agency as presented above, it is clear that action “implies that we become the accomplices of a reality that ultimately cannot be fully mastered” (Tengelyi 2012a: 281). Therefore, the upshot of a narrative account of personal identity is the reverse of what the constructivist reading anticipated. It reveals that narration is bound up with the subject’s continuous “complicity” with reality. At the end of the narrativity paper, Tengelyi summarizes this point in reference to Ricoeur (1992: 308) by defining the subject in its double role as agent and narrator as a “decentered self.” For Tengelyi, this notion gathers the two original and “most important contributions” that a “narrative interpretation of action makes . . . to the characterization of the freedom of action and the selfhood of the agent”: first, that our actions must be always understood “in their intertwinement with the causal mechanisms of the world”; and second, that the agent is “not only the originator of certain initiatives, but also . . . the decentered co-originator of the unintended consequences that supervene on these initiatives” (Tengelyi 2012a: 282). The account of the narrative self as decentered thus takes up central ideas from the phenomenology of action and the metaphysics of freedom it calls for. Only these elements taken together give a full picture of action in Tengelyi’s system.

Conclusion: a case for truth Tengelyi’s account of human freedom in World and Infinity, outlined in Sections ‘Freedom and necessity: the meaning of unintended consequences’ and ‘Grounding grounds: on different ways to be free,’ addresses a lacuna in the discussion of narrative identity in the earlier paper “Action and Self hood.” Both accounts have the same central theme, the constitution of the meaning of action and the phenomenological traits and metaphysical implications of unintended consequences. But it is not obvious how both accounts can be integrated. In the following, I will try to show that this difficulty is due to some features of Tengelyi’s view of narrative identity presented in the last section. To get to a more coherent picture, a stronger emphasis on the way in which human freedom engages normativity is needed, which I aim to bring out by highlighting the role of truth and the ideal of authenticity in the constitution of the meaning of action. A discussion of truth can function as the missing link between the different elements of or stages in the development of Tengelyi’s view. There are two elements in Tengelyi’s account of narrative identity I find unsatisfactory. First, recall that Tengelyi relied on the view that the constitution of self hood is multilayered. This view provided Tengelyi with an answer to how a narrative identity can be understood as discovery: it discovers, through articulation and recollection, two deeper layers of the self, i.e. reflective self-awareness and passive self-constitution. The problem with this layered account Tengelyi takes over from Husserl is that it does not get a grasp of a feature of subjectivity that is presupposed by the layer-account: its normative relation to truth. Thus the notion of foundation (Fundierung) to which Tengelyi refers, is one of the two elements, 252

László Tengelyi

alongside with fulfillment (Erfüllung), defining Husserl’s account of truth as Evidenz (Husserl 1970: 226–268; Crowell 2013: 37–41). To conceive one layer of the constitution of the self as founded on another presupposes this notion of truth. Further, to conceive the difference between construction and discovery as a contrast defining success in the constitution of a narrative identity, which Tengelyi clearly does, presupposes a prior commitment of the self to understand herself in light of the truths encountered in the process of constituting her life-history. Telling this story involves a standard of truth and a commitment to behold oneself to that standard. A commitment to constituting, retrospectively, her narrative self as a true self, is also defining the two acts through which the narrative self bases itself in the deeper layers of the self: in the case of articulation, that the narration of one’s past experiences is to be founded on acts of conscious recollection rather than fantasy involves a normative contrast to which articulation is beholden. Similarly in the case of recollection: if the narrative self is to be constituted in the narration of one’s deeds, then the subject must recollect its past actions rather than its perceptions or other past experiences. Again, an analysis of the semantic structure inherent in the intentional content of an act required for narration to succeed defines a standard for success and failure for this act (Crowell 2013: 156). For narration to be based on recollection and narration therefore involves the recognition of different instances of truth: recollecting other experiences than actions fails to constitute the narrative self as the agent she in fact is; basing one’s personal history on fantasies may be an instance of either a mistake or a self-deception, but clearly also fails to be true. Already on the two layers the narrative is founded upon, the subject is thus constituted in view of a form of truth as evidence. The idea that a narrative self is discovered rather than construed constitutes a normative standard necessary to conceive the narrative self as one layer in the layer structure of the self. Reference to truth in this sense is not foreign to Tengelyi’s treatment of human freedom. Although Tengelyi discusses truth with regard to neither narrative identity nor agency, he emphasizes its importance in the presentation of “Heidegger’s metontological foundation of metaphysics” (Tengelyi 2014: 228–264) in an earlier chapter of World and Infinity. Here, Tengelyi takes it to be defining Heidegger’s account of freedom as transcendence that it discovers entities in normative terms, referring to a passage from The Fundamental Concepts of Metaphysics. In this passage, Heidegger (1995: 339) describes “being free [Freisein]” as the “pre-logical being open for beings as such and holding oneself towards the binding character of things,” thereby opening the “leeway [Spielraum] . . . of truth and falsity.” In his discussion, Tengelyi emphasizes that this alethic function of freedom allows one to recognize an “epistemic dynamic [epistemische Dynamik]” (Tengelyi 2014: 258) that is not dependent on a specific form of rationalization and its respective Weltentwurf. Despite the variance in different forms of rationalization, which Tengelyi recognizes as a central element in Heidegger’s account, this dynamic can be seen as the genuine alethic correlate of human freedom. Tengelyi associates this dynamic with Heidegger’s account of the world (Tengelyi 2014: 257–259) as well as with Heidegger’s notion of an active manifestation or “essence of truth” (Heidegger 1998: 136–154; Tengelyi 2014: 391). Given Tengelyi’s reliance on Heidegger’s metontological account, it is thus in continuity with his own explicitly formulated views to situate his account of the narrative self within the account of truth implied by the account of freedom he endorses. It is simply addressing an incompleteness in Tengelyi’s discussion to say that in recollection and articulation, the constitution of the narrative self is beholden to different moments of an epistemic dynamic. But to recognize the subject’s constitutive beholdenness to different manifestations of truth also calls for a more explicit treatment of how such beholdenness contributes to the constitution of the self. 253

Tobias Keiling

This leads to the second problem I see in Tengelyi’s account of narrative identity, which regards the idea that practices are necessarily anonymous and depersonalizing. It is a consequence of that view that an action, if it is part of a practice, contributes nothing or very little to the self ’s narrative identity. Although I won’t argue for this here, this likely overestimates the importance of unintended consequences for the ordinary (‘non-Oedipus’) self. More importantly, Tengelyi seems to misconstrue the way in which adhering to a practice involves the recognition of normativity. Recall from the last section that the only clearly normative element in his account of what is involved in being a university professor is that “he insists on the standards of his profession” (Tengelyi 2012a: 273), which I take to mean that he is evaluating the actions of others. But there is a deeper normative element regarding the first-person: to genuinely uphold the standards of my profession, I must not only be following the rules pertaining to a specific practice; I must be committed to them, take them as binding for myself. Using, curiously, the same example, Crowell (2013: 249) refers to this as the attempt at “trying to be” a philosophy professor. This “trying,” however, refers to the norms inherent in what it means to be a true (good) professor, and these can contribute to how I understand myself. Adopting a term from Korsgaard (1996: 101), Crowell highlights that recognizing the standards of certain practices as binding yourself provides you with a “a description under which you value yourself,” a “practical identity.” At first sight, this contrasts with Tengelyi’s insistence that practices do not contribute to the identity of the self, and even if they did, committing to such typical practical identities would not contribute to the specifically narrative identity of the self. There is, however, a way in which practical identities may well contribute to my narrative self. To see how both practical and narrative identities can be related, however, it is again necessary to turn to a discussion of different instances of truth. Without embracing a narrative account of the self, Heidegger’s account of the authentic self offers a model for how to understand both the continuity and the contrast between these two layers or aspects of the self. Importantly, this discussion of authenticity helps not only to better understand how practical identities can tie in with a life-history (in more mundane cases), but it can also be extended to account for unintended consequences and thus to cover Tengelyi’s example (the extreme of Oedipus). In both kinds of cases, the account of a free, true (authentic) self can help to bridge rather than oppose the forms of rationalization provided by practices with that provided by narrativity. I present the central element of that account before connecting it to the discussion of metontological freedom, which Tengelyi clearly endorses. Heidegger opens his discussion of authenticity with the idea that there must be an “attestation” (Bezeugung, 2006: 267, 2010: 257) of the possibility of an authentic existence. He finds such attestation in the possibility of what he calls “Nachholen einer Wahl” (2006: 268), which I submit might best be translated as the “retrieving of a choice.” Such retrieval constitutes the link between the self constituted in every practices, what Heidegger calls the they-self (Man-selbst), and the authentic self: “the they-self is modified in an existenziell manner so that it becomes the authentic Being-one’s-self ” (1962: 313, 2006: 268, 2010: 258). This idea has room for Tengelyi’s observation that practices may well be anonymous and depersonalizing: simply adhering to the rules of a practice obscures that there is a choice involved in adhering to the practice. In Heidegger’s account, however, the meaning first constituted by the ‘depersonalized’ ‘they-self ’ is also a prerequisite for the self; the self becomes authentic by (first-personally) committing to the normative standards in which it is already engaged, which may, but need not, involve deliberation and a conscious decision. Inauthentic (third-personal) adherence to the practice, by contrast, never involves true choice, which motivates both Macquarrie-Robinson and Stambaugh to translate “Nachholen einer Wahl” 254

László Tengelyi

as “making up for not choosing” (1962: 313, 2010: 258). But this translation is somewhat misleading: Heidegger stresses that the self has already been involved in choosing, and it is this choice (and its consequences) that an authentic self discovers as having been a possibility of her dasein rather than a requisite imposed on the self by others. What the authentic self must “make up” for is its having, in the guise of the they-self, already made some (albeit inauthentic) decision. This is why Heidegger stresses that the retrieval of choice is essentially a reiteration and glosses it as “choosing to make this choice [Wählen dieser Wahl]” (1962: 313, 2010: 258, emphasis modified). It is decisive for normativity readings of Being and Time such as Crowell’s that the retrieval of choices involves what Heidegger calls “[taking] over being a ground in existing [existierend das Grundsein übernehmen]” (2006: 284, 2010: 273), i.e. recognizing grounds/reasons with which dasein founds herself already confronted (‘thrown into,’ in Heidegger’s idiom) as not merely third-personal forms of rationalization but as dasein’s genuine possibilities. Within the framework of Being and Time, Heidegger locates the free subject in the authentic experience of the contrast between a free self and the they as “basic concept of unfreedom” (Figal 2005: 115). Crowell (2013: 189) puts it succinctly: “Freedom is not essentially the ability to choose between possibilities, but the difference between the third-person and the first-person as such.” While there is no explicit treatment of authenticity in Heidegger’s later discussion of freedom as “ground” for different forms of rationalization presented in Section ‘Grounding grounds: on different ways to be free,’ the discussion of authenticity might provide a bridge from Tengelyi’s endorsement of that view to a modified account of the constitution of the narrative self. Thus the contrast between discovering and constructing the self ’s narrative identity can be seen as being part of the self ’s commitment to being authentic, namely as the agent trying, retrospectively, to be true to the meaning of her past deeds, including those actions that form part of a practice. This idea can be spelled out in two steps. First, authentically retrieving one’s action as a they-self effects a modification of the meaning from third-personal to first-personal terms that, even if the descriptions of the actions do not involve unintended consequences, may provide ample material for a story. Thus even if its description in third-personal terms may be unaltered, how a self experiences her action retrospectively may very well yield itself to narration. In particular, a first-personal account may involve an attestation of authenticity worthy of narration: ‘first I did it because everyone else did it, now I do it for my own reasons.’ In this case, authenticity quite precisely marks the transition from a ‘depersonalizing’ involvement in practices to one that, because it is authentic, can also be a formative element in the self ’s life-history. The phenomenology of agency thus yields a good reason why a philosophical account of self-constitution should not exclude narratives of the ordinary, a point Cavell has argued in numerous contexts (Mulhall 1994). Authentic involvement in a practice also involves risking unintended consequences, satisfying Tengelyi’s explicit standard for what is worthy of story-telling. Second, and more importantly, the idea of retrieving one’s choices may be reformulated to apply to cases where unintended consequences do occur and may lead, such as in Tengelyi’s example, to a complete reversal of the first-personal experience of the action. Recall from Section ‘Grounding grounds: on different ways to be free’ that Tengelyi’s most complex analysis of the case highlighted Oedipus’ attempts to regain control of his initial actions, the (causally determined) emergence of unintended consequences motivating his (teleologically determined) decision to initiate further action in the attempt to compensate for the deviation the development of the meaning of his actions has taken from his initial intention. Although Heidegger’s discussion in Being and Time is geared toward the inauthentic, 255

Tobias Keiling

third-personal meaning generated by the they of every-day practice, his discussion of authenticity may be extended to what Oedipus is trying to do: he is trying to retrieve his initial choice. Recognizing that, despite its meaning now mainly determined by factors outside his intention, his deeds are his, Oedipus attempts to again use his freedom and modify the ultimate outcome of the process in which the meaning of his initial action is constituted. As Tengelyi highlighted, with regard to the metaphysics of agency, this shows that the freedom to initiate action is an always renewable capacity. Importantly, in contrast to views such as Schelling’s, which Tengelyi explicitly rejected, authentically retrieving one’s actions does not involve reinstating a fully transparent and autonomous subject. It only leads to assuming that the self is trying to be authentic with regard to the indeterminate process in which the meaning of her deeds is constituted. The idea of authenticity adds to Tengelyi’s view an account of how the renewable exercise of this freedom remains beholden to the unfolding of the actions’ (true) meaning, providing a coherent picture of what motivates the sequence of choices and defines the emerging narrative structure. As Wrathall (2015: 193) highlights, it is a defining feature of Heidegger’s approach that he gives methodological priority to authenticity over autonomy. That is, Heidegger argues that one cannot understand what the self is, let alone figure out the right way for the self to ground action, until one has grasped authenticity as an ideal of human existence. Although there is no explicit reference to authenticity in Heidegger’s writing from the 1930s, if authenticity is associated with freedom, the metontological account of human freedom, too, requires a modification of the ideal of authenticity to include the aspects discussed in Section ‘Grounding grounds: on different ways to be free’: that the rationalization of action is an inherently pluralistic capacity yet remains beholden to the truth of particular entities. In an earlier discussion of the notion of experience, Tengelyi makes a similar claim regarding the nature of experience as such. Even if the disappointment of expectations and the possibility of conflict are defining feature of experience, this does not negate the directedness of experiences toward entities. Reiterating a point raised in Heidegger’s (2002: 86–156) critique of Hegel’s notion of experience in the Phenomenology of Spirit, Tengelyi writes: If the crossing-out of an expectation results from a conflict in experience that, as Husserl wrote, presupposes ‘a certain basis of agreement’, then it is misleading to speak [with Hegel, T.K.] of changing the object of experience. Rather, one has to insist that it is one and the same object that is suddenly seen in a new light or takes on a new form. (2007b: 15–16) Given this also holds true in the experience of action, it means that the notion of an “open essence [offenes Wesen]” of things, which Tengelyi (2014: 429) takes over from Husserl (1989: 313), should be extended to actions. Despite its incompleteness with regard to first-person involvement, the semantic framework then correctly identifies the process of describing and redescribing the same action as central element in how its “open essence” evolves. Tengelyi’s central idea that it is the experience of action and the self ’s experiential way that link action and narration can then be specified in the following way: what the self discovers in the recollection and the articulation forming part of narrating itself is the passively constituted self in its interaction with the open essence of entities. Aiming to be authentic, the self 256

László Tengelyi

is committed to discovering the truth about entities, and this includes the dynamic constitution of the meaning of its actions. That it tries to adhere to all different instantiations of truth qua evidence defines both the prospective and the retrospective action of the self: she attempts to redirect her deeds and influence how the meaning of her actions develops further by making use, prospectively, of her renewable capacity to initiate further action. Retrospectively, she attempts to give an account of her true self, committing to discovering her past through recollection and articulation. According to the modification of Tengelyi’s view I propose, what links both the prospective and the retrospective is the self ’s attempt to authentically relate to her past actions, or in Heidegger’s language: to retrieve her past choices. Importantly, this not only requires switching between a retrospective and prospective perspective. It also requires making use of different forms of rationalizing the changing meaning of actions. Narrating one’s agential self is then not merely another form of rationalizing action besides its rationalization in causal, teleological, etc. terms. Understood as outcome of the self ’s capacity to ‘ground grounds,’ a narrative rapport to one’s past actions requires making use of different forms of rationalizing it in causal, teleological. . . terms. In all of this, however, the authentic attempt to narrate one’s life-history remains beholden to the evolving meaning of individual actions. Being free in the sense of transcending into the world, the changing truths about an individual action constitute the genuine albeit not uniform measure of the self ’s attempt at giving a true account of her past deeds, even if these truths include recognizing unintended consequence of the most extreme kind. Referring to the risk of experiencing unintended consequences, Tengelyi in an earlier discussion of freedom in passing calls this risk the “character of adventure” (Abenteuercharakter, 2009b: 251) that accompanies every intentional action. As Tengelyi argues, a precise sense can be given to the idea that the life of an agent can be called an adventure. But what is an adventure if not a quest for truth?1

Related topics See Chapters 8 (on Heidegger) and 15 (on Ricœur).

Note 1 For very helpful comment and encouragement, I am indebted to Inga Römer and Steven Crowell.

References Anscombe, E. (2000) Intention, second edition, Cambridge, MA: Harvard University Press. Arendt, H. (1958) The Human Condition, second edition, Chicago, IL: The University of Chicago Press. Crowell, S. (2001) Husserl, Heidegger, and the Space of Meaning: Paths towards Transcendental Phenomenology, Evanston, IL: Northwestern University Press. Crowell, S. (2013) Normativity and Phenomenology in Husserl and Heidegger, Cambridge, MA: Cambridge University Press. Davidson, D. (2001) Essays on Actions and Events, second edition, Oxford: Clarendon Press. Doyon, M. and Breyer, T. (eds.) (2015) Normativity in Perception, London: Palgrave Macmillan. Figal, G. (2005) “Being-with, Dasein-with, and the They as the Basic Concept of Unfreedom,” in R. Polt (ed.), Heidegger’s Being and Time. Critical Essays, Lanham, MD: Rowman & Littlefield, pp. 105–116. Golob, S. (2014) Heidegger on Concepts, Freedom, and Normativity, Cambridge, MA: Cambridge University Press.

257

Tobias Keiling Gondek, H.-D. and Tengelyi, L. (2011) Neue Phänomenologie in Frankreich, Berlin: Suhrkamp. Hartmann, N. (1932) Ethics. Volume III. Moral Freedom, London: Allen & Unwin. Hartmann, N. (1960) Die Philosophie des Deutschen Idealismus, second edition, Berlin: De Gruyter. Haugeland, J. (2013) Dasein Disclosed, Cambridge, MA: Harvard University Press. Hegel, G.W.F. (1975) Aesthetics: Lectures on Fine Arts, Oxford: Clarendon Press. Hegel, G.W.F. (2010) The Science of Logic, Cambridge, MA: Cambridge University Press. Heidegger, M. (1962) Being and Time, trans. by John Macquarrie and Edward Robinson, Oxford, Cambridge: Blackwell. Heidegger, M. (1984) The Metaphysical Foundations of Logic, Bloomington: Indiana University Press. Heidegger, M. (1995) The Fundamental Problem of Metaphysics, Bloomington: Indiana University Press. Heidegger, M. (1998) Pathmarks, Cambridge, MA: Cambridge University Press. Heidegger, M. (2002) Off the Beaten Track, Cambridge, MA: Cambridge University Press. Heidegger, M. (2006) Sein und Zeit, nineteenth edition, Tübingen: Max Niemeyer. Heidegger, M. (2010) Being and Time, trans. by Joan Stambaugh, revised edition, Albany: State University of New York Press. Husserl, E. (1966) Analysen zur passiven Synthesis, Husserliana vol. XI, The Hague: Martinus Nijhoff. Husserl, E. (1970) Logical Investigations. Volume II, London: Routledge and Kegan Paul. Husserl, E. (1973) Zur Phänomenologie der Intersubjektivität, Dritter Teil: 1929–1935, Husserliana XV, ed. Iso Kern, Den Haag: Martinus Nijhoff. Husserl, E. (1989) Ideas Pertaining to a Pure Phenomenology and Phenomenological Philosophy. Second Book, Dordrecht, Boston, MA and London: Kluwer. Husserl, E. (2014) Ideas for a Pure Phenomenology and Phenomenological Philosophy. First Book, Indianapolis, IN and Cambridge: Hackett. Keiling, T. (2018) “Phenomenology and Ontology in the Later Heidegger,” in D. Zahavi (ed.), The Oxford Handbook of the History of Phenomenology, Oxford: Oxford University Press, pp. 251–267. Keiling, T. (2020) “Freiheit und Determination bei Tengelyi,” in I. Römer and A. Schnell (eds.), Phänomenologie und Metaphysik, Hamburg: Meiner, pp. 213–229. Korsgaard, C. (1996) The Sources of Normativity, Cambridge, MA: Cambridge University Press. Kriegel, U. (2015) The Varieties of Consciousness, Oxford: Oxford University Press. Loidolt, S. (2018) Phenomenology of Plurality: Hannah Arendt on Political Intersubjecivity, New York and London: Routledge. Mulhall, S. (1994) Stanley Cavell. Philosophy’s Recounting of the Ordinary, Oxford: Clarendon. Ricœur, P. (1991) “Life in Quest of Narrative,” in D. Wood (ed.), On Paul Ricoeur. Narrative and Interpretation, London and New York: Routledge, pp. 20–33. Ricœur, P. (1992) Oneself as Another, Chicago, IL: Chicago University Press. Römer, I. (2017) “Was ist phänomenologische Metaphysik?,” in M. Gabriel, C. Olay and S. Ostritsch (eds.), Welt und Unendlichkeit. Ein deutsch-ungarischer Dialog in memoriam László Tengelyi, Freiburg and München: Alber, pp. 115–130. Römer, I. (2020) “Laszlo Tengelyi,” in T. Keiling (ed.), Phänomenologische Metaphysik. Konturen eines Problems nach Husserl, Tübingen: Mohr Siebeck, pp. 403–414. Schear, J. (ed.) (2003) Mind, Reason, and Being-in-the-World: The McDowell-Dreyfus Debate, London: Routledge. Schelling, F. W. J. (1989) The Philosophy of Art, Minneapolis: University of Minnesota Press. Tengelyi, L. (1998) Der Zwitterbegriff Lebensgeschichte, München: Fink. Tengelyi, L. (2004) The Wild-Region in Life-History, Evanston, IL: Northwestern University Press. Tengelyi, L. (2007a) “Narratives Handlungsverständnis”, in K. Joisten (ed.), Narrative Ethik. Das Gute und Böse erzählen, Berlin: Akademie Verlag, pp. 61–73. Tengelyi, L. (2007b) Erfahrung und Ausdruck: Phänomenologie im Umbruch bei Husserl und seinen Nachfolgern, Dordrecht: Springer. Tengelyi, L. (2009a) “Une interprétation narrative de l’action,” in P. Kerszberg, A. Mazzù and A. Schnell (eds.), L’oeuvre du phénomène. Mélanges offerts à Marc Richir, Bruxelles: Ousia, pp. 197–214. Tengelyi, L. (2009b) “Betrachtungen über die Handlungsfreiheit und die Selbstheit des Handelnden,” in M. Pfeifer and S. Rapic (eds.), Das Selbst und sein Anderes. Festschrift für Klaus Erich Kaehler, Freiburg and München: Alber, pp. 245–258. Tengelyi, L. (2012a) “Action and Self hood: A Narrative Interpretation,” in D. Zahavi (ed.), The Oxford Handbook of Contemporary Phenomenology, Oxford: Oxford University Press, pp. 265–286.

258

László Tengelyi Tengelyi, L. (2012b) “Nicolai Hartmanns Metaphysik der Freiheit,” in G. Hartung, M. Wunsch and C. Strube (eds.), Von der Systemphilosophie zur systematischen Philosophie. Nicolai Hartmann, Berlin and Boston, MA: De Gruyter, pp. 277–295. Tengelyi, L. (2014) Welt und Unendlichkeit. Zum Problem phänomenologischer Metaphysik, Freiburg and München: Alber. Tugendhat, E. (1979) Selbstbewußtsein und Selbstbestimmung, Frankfurt am Main: Suhrkamp. Waldenfels, B. (2016) Antwortregister, second edition, Frankfurt am Main: Suhrkamp. Wrathall, M. (2015) “Autonomy, Authenticity, and the Self,” in D. McManus (ed.), Heidegger, Authenticity, and the Self, London and New York: Routledge, pp. 193–214. Zahavi, D. (2017) Husserl’s Legacy: Phenomenology, Metaphysics, and Transcendental Philosophy, Oxford: Oxford University Press.

259

PART II

Systematic Perspectives

Phenomenology of Agency 1: General Issues

19 ON THE SATISFACTION CONDITIONS OF AGENTIVE PHENOMENOLOGY A dialogue Terry Horgan and Martine Nida-Rümelin What is it like to experience one’s own behavior as the exercising of one’s agency? What are the satisfaction conditions of agentive phenomenology (the conditions that must be met in order for experiences of agency to be veridical)? Is such phenomenology typically veridical, or is it instead usually (or even always) illusory? In this dialogue we pursue a debate about these issues that began as a series of oral exchanges between us at recent conferences in Fribourg and in Munich. Horgan has been seeking to articulate and defend a position that (i) fully acknowledges the phenomenal character of agentive experience,1 (ii) treats such experience as veridical, and (iii) renders agentive phenomenology compatible with familiar, broadly “materialist,” tenets in metaphysics and philosophy of mind. Nida-Rümelin has been challenging Horgan; she has maintained that he does not successfully offer a position incorporating each of features (i)–(iii), and she has expressed persistent skepticism about the very possibility of such a package-deal position. We share some important common ground. We agree that human action is suffused with rich phenomenal character of a specifically agentive kind, and we largely agree about the introspectable features of this phenomenology. We agree that agentive phenomenology is richly intentional, in the sense of having content with satisfaction conditions.2 And we agree that such phenomenology typically is veridical rather than illusory. We disagree sharply, however, about the nature of the satisfaction conditions that accrue to agentive experience. Horgan maintains that these conditions are satisfiable consistently with a broadly materialist metaphysics of mind (e.g., Horgan 2007a, 2007b, 2011, 2015), whereas Nida-Rümelin is doubtful (e.g., Nida-Rümelin 2007, 2016, 2018a). We will not resolve this disagreement in what follows. We do agree, however, that the ensuing debate nonetheless manifests philosophical progress: viz., Horgan refines and improves his position through the course of his successive installments, in ways that are directly prompted by Nida-Rümelin’s successive formulations and reformulations of her fundamental objection to Horgan’s project. We expect our debate to continue beyond what is said here. And we sincerely hope that others in philosophy will recognize the importance of the issues we address and will join the ongoing discussion. 264

Conditions of agentive phenomenology

Terry Horgan: proposal I In this section I will first describe what I take to be some salient aspects of ordinary first-person agentive phenomenology, and I will then set forth some of my key claims, in prior writings and conference presentations, about the satisfaction conditions of such phenomenology. What is it like, phenomenologically, to apparently perform an action? (Here and throughout, I deploy the modifiers ‘apparent’ and ‘apparently’ in order to maintain neutrality about whether or not such agentive phenomenology is veridical.) I will describe seven introspectively discernable aspects of such experiences. First is the aspect of self as source. One’s apparent actions are not experienced passively, e.g., as bodily or mental events that are merely happening to oneself, or as the outcome of passively experienced state-causal processes one undergoes (e.g., a process consisting of an occurrent desire-belief combination state-causing an intention-formation or a bodily motion).3 Rather, one experiences one’s apparent actions as emanating from oneself as their source—as opposed to experiencing them, passively, as the end products of an experienced state-causal process. (When a process is experienced as transpiring state-causally, however, the phenomenology is as-of the process evolving by itself, without any agentive intervention either by oneself or by others.) Second is the phenomenological aspect of optionality. One’s apparent actions are experienced as being under one’s voluntary control, and thus as being such that in the circumstances one can/could act otherwise. In advance of apparently performing an action that one is considering performing, one experiences the future as being “open” to oneself, with respect to whether or not one will perform it. And once an apparent act has occurred (either spontaneously or upon prior deliberation), one experiences both its potential performance and its potential non-performance as having been open to oneself. Third is the phenomenological aspect of purposiveness. Normally, one experiences one’s actions as being done on purpose, and as being done for specific purposes both coarse-grained and fine-grained. For example, the coarse-grained purpose for which one enters one’s office in the morning involves further, more fine-grained aspects of purposiveness, such as pulling out one’s keys as one approaches one’s office, grasping one’s office key, extending it just so into the lock and then twisting it just so, etc. Fourth is what Matjaž Potrč and I call chromatic illumination (cf. Horgan and Potrč 2010). Often, many of the purposes informing one’s actions affect the character of one’s agentive experience by being implicitly appreciated consciously without actually being represented in consciousness. These purposes are phenomenologically operative as reasons for one’s actions, coloring the specific character of one’s agentive experience without being overtly “before the mind.” Fifth, the practical reasons (i.e., purposes) for which one (apparently) acts, and which figure in the character of one’s agentive experience, normally exhibit a means-ends teleological structure. In the familiar simplest case, one has a desire D for a particular outcome, and one has a belief B that performing action A is an available best-means to secure that outcome. This belief-desire combination constitutes a practical reason for performing action A. One acts in light of, and in response to, such a practical reason toward achieving an end. (Some aspects of the belief-desire combination might be implicitly appreciated experientially via chromatic illumination, without being overtly present in consciousness.) Sixth, appreciation of a practical reason for performing an action A normally exerts a motivational pull toward A-ing. The appreciated reason is experienced as authoritatively favoring A-ing. In the case of ordinary practical reasons, this authoritativeness has a “hypothetical 265

Terry Horgan and Martine Nida-Rümelin

imperatival” character rather than a “categorical imperatival” character, because its authority is desire-dependent. Such desire-dependence is intimately connected to the experienced voluntariness of one’s actions—the aspect of optionality. One’s experience is as-of being free to act on one’s desire or to refrain from doing so. Finally, when one (apparently) acts on the basis of a practical reason, normally one experiences oneself as so acting because of that reason, qua reason. (Again, this can be implicit, via chromatic illumination.) This “rational becausal” aspect of agentive experience is not an aspect as-of passively experienced state causation—say, as-of one’s belief-state and desire-state jointly state-causing one’s body to move. On the contrary, the experience of performing an action for a reason has the phenomenological aspect as-of self-sourcehood—and, moreover, as-of purposive self-sourcehood, and, indeed, as-of self-sourcehood involving a specific purpose. That experienced purpose—the consideration that one experiences as the reason for which one now acts in the particular manner one does—is the “rational becausal” aspect. Phenomenologically, such experience resides squarely, indeed paradigmatically, in the “space of reasons,” rather than residing in the “space of causes,” i.e., the space of phenomena experienced passively as state-causal transactions. Agentive phenomenology, in virtue of the several aspects I have been describing, is rife with intentionality—in the sense of ‘intentionality’ involving representational content. (It is also rife with intentionality in the sense of purposiveness, which is a part of its intentionality in the former sense.) The following two important, deeply intertwined, philosophical questions therefore arise. First, what are the satisfaction conditions of ordinary agentive phenomenology? Second, are these satisfaction conditions normally (or ever) actually met when human beings undergo agentive experience? I will now briefly rehearse some of my views pertaining to these two questions, which elsewhere I have articulated at greater length and have defended (e.g., Horgan 2007b, 2015). First, it is introspectively obvious that agentive phenomenology is not as-of one’s (apparent) action being a state-causal process. It is not as-of one’s bodily motion or one’s intention-formation being state-causally generated by a mental state or a combination of mental states in oneself (e.g., a belief-desire combination). Nor is it as-of a physical statecausal process, qua physical (e.g., a state-causal process consisting of certain brain-states, qua brain states, state-causing one’s body to move in a specific manner). Second, it is crucially important to distinguish between (1) one’s agentive phenomenology being not being as-of a state-causal process, and (2) one’s agentive phenomenology being as-of not a state-causal process. Feature 1 is a way that one’s apparent action is not experienced to be, and nothing more. However, if feature 2 were indeed an aspect of one’s agentive phenomenology then this would be a way that one’s apparent action is experienced to be—albeit a negative way-of-being, viz., being not a state-causal process. Third, although there is presumably a determinate fact of the matter about whether or not ordinary agentive phenomenology has feature 2, introspection alone cannot reliably ascertain whether it does. Thus, those philosophers who think that they reliably detect the presence of feature 2 in their agentive phenomenology are mistaken; they are conflating feature 2 with feature 1. (This is not to say that introspection alone can reliably detect the absence of feature 2 in ordinary agentive phenomenology; on the contrary, introspection is not powerful enough by itself to yield a reliable verdict either way.) Fourth, there is excellent reason to believe, based on well-documented scientific evidence, both (i) that all human behaviors that are experienced as actions are in fact state-causally generated by neuro-physical causal processes in the brain and central nervous system, and (ii) that all neuro-physical states and events in the brain and central nervous system are 266

Conditions of agentive phenomenology

themselves physically state-causally generated. (This is the hypothesis of “physical causal completeness,” specifically with respect to humans.) Fifth, there is also strong reason to believe that the satisfaction conditions of agentive phenomenology include the following condition of psychological state causation: PSC: In order for an instance of human behavior that is experienced as one’s own action to really be an action, this behavior must be state-causally produced by a psychological state that “rationalizes” it, in the sense of rendering it rationally appropriate—e.g., a state comprising both a desire for a certain outcome and a belief that such behavior would bring about (or probably would bring about) that outcome. This is so even though, as I have already emphasized, experiences of agentively instigating one’s behavior are not experiences as-of such a psychological state-causal process. The PSC harks back, of course, to Donald Davidson’s hugely influential 1963 paper “Actions, Reasons, and Causes”— even though Davidson himself did not acknowledge agentive phenomenology at all, let alone frame his account of agency as being about the satisfaction conditions of such phenomenology. Sixth, concerning the PSC itself, the state-causal efficacy of psychological states is compatible with the hypothesis of physical causal completeness—even though, because of the multiple realizability of psychological state types by neuro-physical state types, numerous psychological state types are distinct from the neuro-physical state types that realize them on specific occasions. (In response to Jaegwon Kim’s infamous problem of “causal exclusion,” I advocate a version of contextualism about causation and causal explanation; cf. Horgan 1989 1998, 2001; Horgan et al. 2009.) Seventh, normally the satisfaction conditions of agentive phenomenology are indeed satisfied on occasions when people are undergoing such phenomenology. That is, normally this phenomenology is veridical, and hence normally people really do exercise genuine free agency of the kind they experience themselves to be exercising. Finally, eighth (and as a consequence of the above), the contention that people normally exercise free agency of the kind they experience themselves to be exercising is compatible with the various other claims I embrace; in particular, it is compatible with the claim that the exercise of free agency is a psychological state-causal process (despite not being experienced as one), and it is compatible with the hypothesis of physical causal completeness. In addition, it is also compatible with state-causal determinism. (Although I regard it a seriously open question whether or not state-causal determinism is true at the level of fundamental physics, I believe that there is strong scientific evidence that neurobiological processes in humans and other organisms are virtually deterministic—with any quantum-level indeterminacies (if such there be) getting “washed out,” rather than getting amplified in such a way as to matter at the neurobiological level of temporal evolution.)

Martine Nida-Rümelin: rejoinder I Experiencing oneself as active Horgan distinguishes six phenomenally manifest aspects present in all or many cases of agentive experience: the aspect of self as source, of optionality, of purposiveness, of chromatic illumination, of apparently acting for certain practical reasons, of motivational pull, and finally the “rational-becausal” aspect of agentive experience. I will focus on the first of these, on what Horgan calls the aspect of self as source. 267

Terry Horgan and Martine Nida-Rümelin

If one wishes to discuss whether an aspect of human agentive experience is veridical without talking past each other, one must first establish a situation of shared reference: one needs to make sure that one is talking about the very same aspect of human experience. Based on numerous conversations I am confident that Horgan and I are in fact in such a situation of shared reference concerning the first aspect he mentions. What Horgan calls ‘the aspect self as source’ is what I have in mind when I use the locution “experiencing oneself as active” (see Nida-Rümelin 2007, 2016, 2018a). We are both confident that this specific aspect is present in all humans when they are acting. We agree on a stronger thesis as well: no bodily movement should count as a case of acting if that specific experiential aspect is absent. I would like to make sure that we are in a situation of shared reference concerning that specific experiential aspect with the reader as well. In order to do so, let me first cite parts of a well-known passage occurring in several of Horgan’s previous writings about the topic. I would like to invite the reader to use that passage in order to individuate the specific aspect we are talking about in his or her own agentive experience: What is behaving like phenomenologically, in cases where you experience your own behavior as action? Suppose that you deliberately perform an action—say, holding up your right hand and closing your fingers into a fist. [. . .] In order to help bring into focus this specifically actional phenomenological dimension of the experience, it will be helpful to approach it a negative/contrastive way, via some observations about what the experience is not like. For example, it is certainly not like this: first experiencing an occurrent wish for your right hand to rise and your fingers to move into clenched position, and then passively experiencing your hand and fingers moving in just that way. [. . .] It would be very strange indeed, and very alien. Nor is the actional phenomenological character of the experience like this: first experiencing an occurrent wish for your right hand to rise and your fingers to move into clenched position, and then passively experiencing a causal process consisting of this wish’s causing your hand to rise and your fingers to move into clenched position. [. . .] it seems patently clear that one does not normally experience one’s own actions in that way—as passively noticed, or passively introspected, causal processes consisting in the causal generation of bodily motion by occurrent mental states. That too would be a strange and alienating sort of experience. How, then, should one characterize the actional phenomenal dimension of the act of raising one’s hand and clenching one’s fingers [. . .]? Well, it is the what-it’s-like of self as source of the motion. You experience your arm, hand, and fingers as being moved by you yourself—rather than experiencing their motion either as fortuitously moving just as you want them to move, or passively experiencing them as being caused by your own mental states. You experience the bodily motion as caused by yourself. (Horgan et al. 2003: 327–329) Here are various ways to describe the relevant aspect of agentive experience, two of them taken from the passage above: • • • • •

You experience the relevant part of your body as moved by yourself. You experience the bodily motion as caused by yourself. You experience your own movement as brought about by yourself. You experience yourself as active in what you do. You experience yourself as the causal origin of your movement. 268

Conditions of agentive phenomenology

These descriptions of what appears to be the case when one has an experience with that aspect should help attract the reader’s attention to that specific phenomenally manifest aspect. They are used to fix reference to a specific part of how things appear to be (of how one’s own behavior appears to be generated) while one is executing an action. One cannot refer to that specific phenomenal aspect without saying something about what appears to be the case in such experiences. Therefore, when one tries to establish a situation of shared reference, one seems to be saying something about the veridicality conditions of the kind of experience one has in mind. (Let us work, for the moment, with the following notion of veridicality conditions: the veridicality conditions of a given type of experience E are those conditions, which render, if satisfied, experiences of type E veridical where an experience is veridical if and only if it presents or represents the world as it is.) After all, on a natural understanding and in a first approximation (we will both say more on that issue below), the veridicality conditions of a given type of experience E are simply the conditions which appear to be fulfilled when one has an experience of type E. Therefore, each of the five ways chosen above (with the intention to fix reference to an aspect manifest in agentive experience) involves already an assumption about the veridicality conditions of experiences with that aspect. It is however important to note that one can nonetheless remain neutral (at this point) about whether any of these assumptions is correct. The various descriptions above can successfully serve their communicative purpose (to establish a situation of shared reference to a specific experiential aspect we are all familiar with) even if the corresponding claims about veridicality conditions were, strictly speaking, all false. In what follows, I will call experiences of the relevant kind experiences of oneself as being active in one’s behavior, or experiences of the type OBA for short and, analogously, I will call the phenomenal aspect which characterizes these experiences the OBA-aspect. For various reasons, the OBA-aspect is of particular interest in the context of our discussion. First, it is phenomenologically fundamental in the sense that any other aspect of agentive experience includes it. For instance, experiencing oneself as being able to act otherwise is to experience the capacity to actively bring about a different movement. A parallel claim applies to all the other aspects listed above. Second, it is of all six aspects listed by Horgan the one which is metaphysically most relevant. To have an experience with that aspect is to experience oneself as active in a given movement, as bringing that movement about rather than something one passively undergoes. The line between actions and non-actions does not coincide with the line between active changes (doings) and passive changes (non-doings) since all actions are doings in that sense but not vice versa (see, e.g., Nida-Rümelin 2007). For reasons I develop in some detail in other papers, the metaphysically more significant line is the one between doings and non-doings (see, e.g., Nida-Rümelin, 2018a). Third, if we wonder about whether agentive experience is broadly veridical or rather illusionary, we should focus on the aspect of self as source (on experiencing oneself as active). If Horgan and the vast majority of philosophers and scientists today are right in assuming that human beings are what we may call biological robots (that is, if their behavior is causally determined by preceding conditions plus microphysical laws), then this particular aspect of agentive experience poses a serious problem. It is plausible to think that your experience that you bring about a given movement cannot be veridical if your movement is, in fact, the result of microphysically determined processes caused by preceding conditions prevailing in your brain. Horgan disagrees with that claim and I will defend it against his arguments. However, we can easily agree on the following. If the standard scientific view of human beings forces us to accept that agentive experience is non-veridical, then the aspect of self as source is the most plausible suspect. If agentive experience is in large part non-veridical then this is so in virtue 269

Terry Horgan and Martine Nida-Rümelin

of that specific phenomenally fundamental aspect of agentive experience. Experiences with any of the other aspects are then non-veridical as well but only because they include the aspect of self as source. (This claim is however put into doubt by Baierlé 2016 who argues that the aspect of optionality is an additional ‘suspect’.) In the passage cited above Horgan points out that it would be alienating to experience one’s own movements as caused by our own occurrent mental states such as wishes and desires. Let us call that type of experience (or that aspect of potential experiences) experiences of type MC (or aspect MC—“MC” for “mental causation”). According to Horgan, we never have MC-experiences with respect to our own behavior involved in action. Human agentive experience does not have the aspect MC. Yet, human behavior involved in acting actually is generated—according to that view—precisely in the way in which it would appear to be generated in MC-experiences. In short, MC-experiences would be alienating and yet veridical if we had them with respect to our own behavior in acting. In what follows, I will argue that such a result should not satisfy a philosopher who wishes to defend agentive experience against the charge of being illusionary.

A thought experiment: Alexa’s transformation One morning shortly after waking up, Alexa realizes with surprise and with horror that something has radically changed. Until the evening before, Alexa was just like all of us. She experienced her own movements involved in her actions as brought about by herself. Her agentive experience had the OBA-aspect. Now that aspect is absent. Instead, Alexa is now under the impression that her own wishes and beliefs cause her movements. As every morning, she stands up, goes to the kitchen, and prepares a cappuccino. In the past, she always experienced herself as bringing all those movements about. Not so now. Her wish to have a cappuccino and her beliefs about how she can realize it appear to cause her legs to move in the way they do and her hands and fingers to move in the way required. Alexa is shocked. She experiences the absence of the familiar experience of being herself the source of what happens as deeply disturbing. We can summarize Alexa’s transformation saying that up to the crucial moment (a) Alexa’s agentive experience has the OBA-aspect and does not have the aspect MC and that (b) after that moment Alexa’s agentive experience has the aspect MC but does not have the OBA-aspect. According to the view Horgan presents in the first part of our paper, Alexa’s experience is veridical before and after that change. Her pre-change and post-change agentive experiences are both veridical in virtue of the fact that certain mental events cause her bodily behavior. Before I use the example to formulate my objections, let us see if Horgan can object to anything in the above description of Alexa’s case. It does not seem so. Note that I did not claim that Alexa has the experience of not being actively involved or of not causing her own movements. I only state that, contrary to her previous experience, Alexa does not have the impression of bringing her movements about. One may safely assume that the mere absence of the OBA-aspect would be deeply disturbing. It would be alienating to realize that one’s body is moving as if one were acting without thereby having any impression of being the one who actively originates the relevant movements. To see the problem Alexa poses for the view Horgan describes above, one will have to keep in mind that its main motivation is anti-illusionist. Horgan finds the idea unacceptable that agentive experience is generally and fundamentally illusionary. He intends to avoid the illusionist claim that agentive experience misrepresents the genesis of our own action and 270

Conditions of agentive phenomenology

yet to preserve both the phenomenological insight that we experience ourselves as actively bringing our movements about in every action and the theoretical claim that actions are caused by occurrent mental states. The argument I would like to develop based on Alexa’s case is supposed to show that these claims are in serious tension. Horgan’s proposal avoids the illusionist claim only in a superficial way, which cannot satisfy the anti-illusionist with respect to agentive experience.

First problem raised by Alexa’s case: the appearance of change Imagine how it is like for Alexa to become aware of the radical change in her agentive experience. Until yesterday, she had the impression that she herself brought her movements about in an active manner. Now she has no such impression anymore. Rather, her occurrent mental states appear to cause the way she moves. Vividly imagining Alexa’s situation, one is pressed to accept a further claim. Not only does Alexa realize a radical change in the way her behavior appears to be generated. Rather, the way her behavior is generated will appear to her to have undergone a radical change. With the change in appearance, things will appear to have changed. However, according to Horgan’s proposal, that impression is an illusion. The view implies that no change has taken place in the way Alexa’s movements in her actions come about. The very same occurrent mental states cause her movements before and after the radical change in her agentive phenomenology. This result tells us something not only about the imagined case involving Alexa. More importantly, it tells us something about ourselves. We all would suffer an illusion of change if we underwent an analogous modification in agentive experience. After the change, Alexa has no experience with the OBA-aspect. She does not experience herself as active in her behavior. One must not jump to the conclusion that Alexa then experiences her movements as not being brought about by herself. Horgan stresses in various places and rightly so that we must not conflate the absence of an experience that p is the case with an experience of p not being the case. However, one may now use Alexa’s illusion of change to argue for the stronger claim (that after the change Alexa experiences herself as not being actively involved in her movement). In general, if the change from experiencing something as having property F to experiencing that same thing as having property G necessarily involves an experience of change (the impression that the object itself has changed from having F to having G), then one may conclude that experiencing an object as having F includes experiencing it as not having G. One could defend that principle by application to concrete cases. One would have to show that it leads to intuitively correct results when applied to various examples. I am confident that this can be done in a convincing manner. However, for reasons of space, I will leave it to the reader to submit that principle to intuitive tests.

Second problem raised by Alexa’s case: no genuine experiential access to the genesis of one’s own behavior The second problem I would like to raise against the version of Horgan’s view presently under consideration is perhaps more serious; it also goes deeper and is a little more complicated to explain. To put it briefly, the view implies that human beings have no genuine experiential access to the way their own behavior actually comes about and that, even worse, having such genuine experiential access is impossible for humans since it is incompatible for them with acting. According to the relevant version of the view, the veridicality conditions of OBAexperiences and MC-experiences are the same. Both kinds of experiences are veridical 271

Terry Horgan and Martine Nida-Rümelin

(when they are) in virtue of certain occurrent mental states causing the relevant behavior. So far, these two kinds of experiences are on a par. However, there is an important disanalogy between these two kinds of experiences. Having an OBA-experience does not reveal to the subject how her own behavior comes about while MC-experiences do. Having an OBA-experience does not involve the impression that occurrent mental states cause one’s own behavior. In fact, when a person acts, causation by occurrent mental states is the real genesis of the relevant piece of behavior (according to the view here at issue). It follows that OBA-experiences, even if they are veridical according to the view, do not present to the experiencing subject how her own behavior comes about. Even though her experience of how it comes about is veridical, it does not give genuine experiential access to what is in fact going on. This is so since, in an intuitively clear sense, OBA-experiences are not transparent with respect to their own veridicality condition. In having an OBA-experience one is not aware of the fact that it is mental causation which renders these experiences veridical. Nor could any reflection on how things appear in that experience make it obvious that this is so. To be told that one’s experience of being active is veridical just in case certain occurrent mental states cause one’s own behavior would come as a surprise. Contrary to this, one is aware of what renders one’s experience veridical in the case of MC-experiences. Putting aside issues about the metaphysical nature of causation, one may say that having an MC-experience comes with a clear phenomenally manifest awareness of how one’s own behavior must be generated for the experience to be veridical. A person who has an MCexperience is thereby phenomenally aware of the fact that the experience presents the genesis of her own behavior correctly if and only if occurrent mental states cause her own behavior. One may perhaps object that to say so is to over-intellectualize phenomenal content. I do not think it is, but for the purposes at hand one may weaken the thesis: in the case of MCexperiences, one can easily discover by a little phenomenological reflection that the experience one is having presents the genesis of one’s behavior correctly if it is in fact caused by the relevant occurrent mental states. (To make the present point precise, one needs to develop an understanding of what it is for a veridicality condition to ‘show up’ at the phenomenal level. See for a proposed account Nida-Rümelin 2018b.) Let me summarize the result as follows. According to the view proposed by Horgan in the first part of the present paper, OBA-experiences do not reveal the genesis of our own behavior when we act while MC-experiences, if only we could have them, would reveal how our behavior in acting comes about. This result is highly problematic for a philosopher who wishes to avoid illusionism about agentive experience. To see this one must fully appreciate a few further elements of the view proposed. Horgan acknowledges that it would be alienating to have MC-experiences. It follows that it would be alienating to be experientially aware of the way our own behavior actually comes about. It would be alienating because having MC-experiences excludes that one experiences oneself as actively involved in what one does. I take it that a theory with such a consequence does not genuinely avoid illusionism. It avoids illusionism in its letter (agentive experience is veridical) but not in its spirit (only alienating agentive experience could reveal the truth about the genesis of our behavior in acting). The situation is even worse. Since the view implies that acting necessarily involves OBA-experiences and that MC-experiences cannot co-occur with OBA-experiences in human beings, it follows, furthermore, that experiential access to the way our behavior actually comes about when we act is even incompatible (at least for humans) with acting. I conclude that the first proposal presented by Horgan in the first part of this paper, contrary to his intention, does not do duty to the anti-illusionist intuition. 272

Conditions of agentive phenomenology

Terry Horgan: proposal II Nida-Rümelin characterizes my views about the phenomenology of agentive experience, and about the satisfaction conditions of such phenomenology, in a way that accurately reflects both my prior published work on this topic and my presentations about it at conference venues at which we were both present. In the course of thinking about her challenge, however, I have found myself making refinements to my position that seem to me both independently plausible and helpful in seeking to respond to her. Let me now set forth these refinements, as a prelude to my subsequent replies to her multi-faceted challenge. (I label these refinements “R1,” “R2,” etc. to facilitate subsequent cross-references.) Refinement R1: In previous presentations at conferences I often said that a key satisfaction condition of agentive phenomenology is the requirement that the bodily motion which one experiences as one’s action is state-causally generated by a mental state, or a combination of mental states, that “rationalizes” such an action—for instance, a desire for a certain outcome, together with a belief that the pertinent action will produce (or probably will produce) the desired outcome. I now maintain that a more fundamental satisfaction condition involving mental state causation is this: one’s experience of agentively initiating the action must itself state-cause the pertinent bodily motion. Hereinafter, I will refer to such an action-initiation experience as a willing. (When a specific belief-desire combination constitutes the reason for which one acts, it too is state-causally operative, but indirectly: it state-causally generates one’s willing, which itself is then the proximal psychological state-cause of the pertinent bodily motion.) Refinement R2: My previous discussions gave the impression that I think a creature could exhibit full-fledged agency even without undergoing agentive phenomenology—for instance, by undergoing a belief-desire combination which state-causes a bodily motion that it “rationalizes,” without also undergoing an experience as-of willing the action. In fact, however, I maintain that a creature cannot be a full-fledged agent, and cannot perform full-fledged actions, without undergoing the pertinent agentive phenomenology. Thus, the presence of agentive phenomenology is among the satisfaction conditions of such phenomenology: you can’t exercise full-fledged agency without experiencing yourself as doing so. (I take it that this second refinement is entailed by the first: willings must be state-causally operative—and hence present—in exercises of genuine agency; and willings are experiences as-of agentive self-sourcehood.) Refinement R3: I now distinguish between two kinds of satisfaction conditions. On one hand are content-based satisfaction conditions, which are constitutive (or partly constitutive) of the intentionality of a given, intentional, psychological state type. On the other hand are implementational satisfaction conditions; these are not constitutive of the psychological statetype’s intentionality, but nonetheless must be met in order for that state type to be realized or implemented within a given kind of creature (e.g., by a human being). In humans, I claim, mental state types must always be physically realized by certain brain states; being thus realized is therefore an implementational satisfaction condition of mental state types (for humans), although not a content-based satisfaction condition. Refinement R4: Now, as I write this installment of the dialogue, I maintain that the requirement that one’s willing-experience state-causes one’s pertinent bodily motion is only an implementational satisfaction condition—and is not a content-based satisfaction condition—of one’s agentive phenomenology. (However, I leave open the possibility that this implementation condition is applicable to all possible creatures with mentality, and that this contention perhaps can even be established a priori.) 273

Terry Horgan and Martine Nida-Rümelin

Refinement R5: I now believe that the kinds of experiences as-of state-causal processes that humans are capable of underdoing are always—and can only be—experiences as-of non-agentive state-causal processes. That is, such experiences are always—and can only be— as of state-causal processes that evolve by themselves, without any agentive intervention either by oneself or by other creatures who one perceives as exercising agency. Since this “by itself ” aspect is constitutive of such experience, the experiential content must always be as-of non-agentive state causation. It is a vexed question whether this so because state causation simpliciter is inherently non-agentive, or instead is so even though—as I myself maintain— exercises of agency are themselves a species of state causation even though they cannot be experienced as such. (I return to this question in my third installment below.) I turn now to the two challenges that MNR raises for my position. The first concerns the appearance of change that her thought-experimental experiencer, Alexa, would undergo. The heart of the objection is this: Until yesterday, she [Alexa] had the impression that she herself brought her movements about in an active manner. Now she has no such impression anymore. . . . Rather, the way her behavior is generated will appear to her to have undergone a radical change. With the change in appearance, things will appear to have changed. However, according to Horgan’s proposal, that impression is an illusion. The view implies that no change has taken place in the way Alexa’s movements in her actions come about. The very same occurrent mental states cause her movements before and after the radical change in her agentive phenomenology. This result tells us something not only about the imagined case involving Alexa. More importantly, it tells us something about ourselves. We all would suffer an illusion of change if we underwent an analogous modification in agentive experience. My response is to appeal to my refinements R1 and R2. In fact, I contend, the genesis of Alexa’s movements has changed, according to my position—and hence I am not committed to the counterintuitive claim that the appearance of change is an illusion. The state-causal etiology has changed because the kind of conscious mental state which, in a case of genuine agency, is the proximal mental state-cause of the pertinent bodily movements—viz., a conscious willing of a specific action—is totally absent in Alexa’s case. Now her bodily movements are proximately state-caused by wishes, or by her belief-desire combination states—whereas before her wishes and beliefs were etiologically distal, giving rise to the willings that were the proximal state-causes of her motions back when she was still undergoing normal agentive experience. (She no longer experiences agentive phenomenology, and indeed she no longer exercises agency.) The second challenge posed by Nida-Rümelin is one that she herself considers more serious, as do I. She formulates it this way: According to the view proposed by Horgan in the first part of the present paper, OBA-experiences do not reveal the genesis of our own behavior when we act while MC-experiences, if only we could have them, would reveal how our behavior in acting comes about. This result is highly problematic for a philosopher who wishes to avoid illusionism about agentive experience. To see this one must fully appreciate a few further elements of the view proposed. Horgan acknowledges that it would be alienating to have MC-experiences. It follows that it would be alienating to be experientially aware of the way our own behavior actually comes about. It would be alienating because having MC-experiences 274

Conditions of agentive phenomenology

excludes that one experiences oneself as actively involved in what one does. I take it that a theory with such a consequence does not genuinely avoid illusionism. It avoids illusionism in its letter (agentive experience is veridical) but not in its spirit (only alienating agentive experience could reveal the truth about the genesis of our behavior in acting). The situation is even worse. Since the view implies that acting necessarily involves OBA-experiences and the MC-experiences cannot co-occur with OBA-experiences in human beings, it follows, furthermore, that experiential access to the way our behavior actually comes about when we act is even incompatible (at least for humans) with acting. I conclude that the first proposal presented by Horgan in the first part of this paper, contrary to Horgan’s intention, does not do duty to the anti-illusionist intuition. I take it that the heart of this objection—the source of her contention that my position, as she says, “does not do duty to the anti-illusionist intuition”—is the following first putative consequence (for short, “PC.1”) of the position: PC.1: If we were to experience our own (apparent) actions as mentally state-caused, then (i)

we thereby would be aware of their real nature and (ii) we would find the experience deeply alienating. The transition from this to her claim that my position is too close to the illusionist position is via this key contention (“KC”): KC: If experience of our own (apparent) actions as mentally state-caused would be both ve-

ridical and yet deeply alienating, then my position avoids the illusionist position only in a superficial sense. KC, together with PC.1, entails the following second putative consequence of my position: PC.2: My position avoids the illusionist position only in a superficial sense.

One way I might try to push back against this objection would be to embrace PC.1, but repudiate PC.2 by denying KC. But that is not the tack I wish to take. Rather, I concede KC, at least for argument’s sake (although I also find KC quite plausible). That is, I concede that if my position really did entail PC.1, then the position thereby would be so close to illusionism that it would fail to honor satisfactorily the contention that agentive experience is veridical. Instead I propose to argue that the version of my position that I now espouse, incorporating the five refinements R1–R5 set out above, does not have the untoward consequence PC.1. I will consider three different construals of what it would be like to experience our own actions as caused by mental states of ourselves, and I will take issue with PC.1 under each of these construals. (I will explicitly invoke the various refinements R1–R5, when applicable.) First construal (which presumably is what Nida-Rümelin actually has in mind): One’s experiences are like Alexa’s experiences of the originations of her bodily motions, after her transformation. Second construal: One’s experiences are simultaneously (i) as-of willing one’s action in the manner that includes the phenomenology of self-sourcehood and the phenomenology of optionality, and (ii) as-of this very willing state-causally generating the pertinent bodily 275

Terry Horgan and Martine Nida-Rümelin

motion. Moreover, aspect (ii) is an experience as-of a purely passive process in which agency is not involved. Third construal: One’s experiences are simultaneously (i) as-of willing one’s action in the manner that includes the phenomenology of self-sourcehood and the phenomenology of optionality, and (ii) as-of this very willing state-causally generating the pertinent bodily motion. Moreover, aspect (ii) is itself neutral as to whether or not agency is involved in this state-causal process. Under the first construal, the envisioned experiences would not “reveal to us how our behavior in acting comes about”—for two important reasons. First, now one would entirely lack the phenomenologically agentive willing-states which, in cases of genuine action (given R1 and R2), are the proximal psychological state-causes of one’s bodily motions. Instead, the proximal psychological state-causes of one’s bodily motions would be d ifferent—viz., states like wishes, belief-desire combinations, and the like. (In cases of genuine action, however, such states are only distal psychological causes of one’s bodily motions, with willing-states being the psychologically proximal ones). Second, now one’s bodily motions would be experienced as non-agentive, since (given R5) humans can only experience processes as state-causal at all by experiencing them as non-agentively state-causal. So although Alexa-type experiences would be veridical vis-à-vis the bodily motions that they themselves state-causally generate in herself after her transformation, in the respects mentioned they would reveal ways that those bodily motions are now being brought about in her case that are different from how ordinary human actions are really brought about. Thus, PC.1 is false under this construal, because the experience in question is not veridical vis-à-vis the real nature of actual human actions. Under the second construal (and given R5), aspect (ii) of the pertinent kind of experience would be as-of a non-agentive state-causal process. Thus, one’s overall experience (if one could undergo such an experience at all) would be as-of an impossible situation, experienced as impossible—viz., an experience as-of a phenomenon that is both agentive by virtue of feature (i), and non-agentive by virtue of feature (ii). Now, it is fairly plausible that such a putative experience, as-of such an impossible state of affairs, is itself impossible—at least for humans. But perhaps it is actually possible for humans, at least in something like the manner in which it is possible to experience apparent “impossible objects” like a visually presented Penrose triangle—in that case, by an overall diachronic experience consisting largely of successive, focally presented and possible, object-parts that collectively cannot be parts of a single possible entire object. (Many famous Escher drawings feature Penrose triangles; see also the cover of my 2017 collection Essays on Paradoxes.) Either way, however, such a putative experience certainly would not “reveal to us how actions are really brought about,” since it would be an experience as-of a bodily motion being brought about in a manner in which it could not possibly be brought about—viz., both agentively and non-agentively. Thus, PC.1 is false under this second construal, because the experience in question—if possible at all (and it may not be)—is non-veridical not only vis-à-vis the real nature of actual human actions, but also vis-à-vis the real nature of one’s behaviors in the envisioned counterfactual scenario itself. Under the third construal, one’s experiences would, I admit, be more revealing concerning “how our behavior in acting comes about” than are actual human agentive experiences. However (and given R1, R2, and R5), these are not experiences that creatures like us humans are capable of undergoing. Given R5, this is because the only kinds of experiences as-of state causation that humans are capable of are experiences as-of non-agentive state 276

Conditions of agentive phenomenology

causation. Of course, it would be alienating for us humans to experience our bodily motions as proximally mentally generated state-causally, because that would be to experience them as non-agentively generated (or as generated, impossibly, both agentively and non -agentively). But the super-humans now being envisioned, unlike ourselves, are capable of experiences as-of state causation that are neutral as to whether or not the experienced process is also agentive. Indeed, they are capable of experiencing their actions simultaneously both (i) as emanating from willing-states with the inherent phenomenological aspects of self-sourcehood and optionality, and (ii) as bodily motions that are proximally state-causally generated, mentally, by these very willing-states. For them, I contend, there is no reason for such dual-aspect experiences to be alienating. But we humans are not capable of such experience. Thus, PC.1 is false under this third construal, because the experience in question—one that humans are incapable of undergoing, or of even imagining what it would be like to u ndergo—would not, I claim, be alienating for creatures capable of undergoing it. The upshot is that according to my position PC.1 is false under each way of construing a counterfactual scenario in which one’s bodily motions are experienced as being proximally state-causally generated by certain mental states. With this as the dialectical situation, I deny that my position entails PC.1. I also deny that I am committed to PC.2—the claim that my view is only superficially different from illusionism about agency—even though I do not contest Nida-Rümelin’s claim KC. This is my reply to the first part of her principal objection. Two residual issues still remain, however, each of which involves a potential challenge to my position that is still broadly in the spirit of Nida-Rümelin’s own worries. First, one might think that it is it somehow philosophically problematic that according to my view, there are satisfaction conditions of agentive experience that are not reliably ascertainable from within agentive experience itself—i.e., are not phenomenally manifest. I contend, however, that there are several reasons why this consequence of my position is not problematic. For one thing (and given R3 and R4), the requirement that one’s willing-the-action be the proximal psychological state-cause of one’s bodily motion is only an implementational satisfaction condition of one’s agentive phenomenology, rather than a content-based satisfaction condition; and there is no evident reason to expect agentive experience itself to reveal either the nature of its merely implementational satisfaction conditions or the fact that they are satisfied on a given occasion of experienced action. (After all, agentive experience certainly does not reveal its own neuro-physical implementational satisfaction conditions, or the fact that they are satisfied on a given occasion of experienced action.) Second, human powers of introspection evidently are somewhat limited anyway, even vis-àvis the specific content-based satisfaction conditions of agentive experience; in particular, they are limited with respect to the satisfaction conditions that accrue to the optionality aspect of such experience, the aspect as-of “could/can do otherwise, given my actual circumstances and myself as I actually am.” (This second point is one I have defended at some length elsewhere; cf. Horgan 2011, 2012, 2015; Horgan and Timmons 2011.) Third (and given R5), there is a principled reason why humans simply are not capable of experiencing their own behaviors both as actions and as proximately psychologically state-causally generated: viz., the only kind of experiences as-of state causation that humans can undergo are experiences as of non-agentive state causation. This last point leads to a second residual issue, however—one which seems to me considerably more serious. The challenge is to explain why the only kinds of experiences as-of state causation that humans can undergo are experiences as of non-agentive state causation. 277

Terry Horgan and Martine Nida-Rümelin

One natural-looking candidate-explanation, of course, is the following. State causation is non-agentive by its very nature; hence, experiences as-of non-agentive state causation reveal the nature of state causation simpliciter, and do so more fully than would experiences as-of state causation that do not include the aspect of being non-agentive. I myself am committed to repudiating this candidate explanation. And I am committed not only to the claim that state causation simpliciter sometimes is agentive, but also to the claim that despite this fact, state causation that is agentive cannot ever be experienced as such by humans. The challenge I face is to explain why not—and to do so in a credible way. (Such an explanation is likely to be somewhat speculative, and to involve encroaching upon matters scientific and empirical. So be it.) In terms of evolution and natural selection, presumably it was important for nature to instill in humans (and in other critters too) the capacity to distinguish very vividly and very saliently between matters susceptible to effective influence by one’s potential bodily motions and matters not thus susceptible. Consider, for example, a bear on a mountainside, generating bodily motions that transfer berries from an adjacent bush to the bear’s mouth and on to its stomach. (I am assuming that the bear undergoes agentive phenomenology.) The bear now notices a giant boulder rolling down the mountain, heading straight toward himself. For him, there should be a very salient distinction between (on one hand) the experience of that boulder’s motion as something not effectively controllable by bodily motions the bear might engage in, and (on the other hand) the experience of agentively transferring berries from the bush to its mouth and stomach. (And in light of this salient difference, the bear should rightly realize that he had better move his bear butt to another location pronto, before the boulder arrives!) Now, one way for evolution to guarantee this kind of distinction would be to do two things together: (i) instill agentive phenomenology—in humans, bears, and various other reasonably sophisticated creatures capable of locomotion—that tends to arise only with respect to bodily motions that are under a critter’s internal state-causal influence; and (ii) instill only a limited capacity for experiences as-of state causation, viz., a capacity for experiences as-of non-agentive state causation. My suggestion is that this is exactly what evolution did do—and that evolution did it because of the clear advantages, in terms of fitness and survival/reproductive potential, thereby imparted upon creatures with both agentive phenomenology and inherently limited, inherently non-agentive, state-causal phenomenology. Distinctions of the kind experienced by the bear on the mountainside would be very vivid experientially, as they should be from the perspective of the evolutionary-biological design of agents with mentality. No doubt there is more to say about this matter, by various parties to the dispute. Some will contend that state causation is inherently non-agentive and that genuine agency is inherently non-state-causal (although they owe arguments for both claims). Among those who embrace this dual contention, some—like Nida-Rümelin, I take it—will contend that humans often do exhibit genuine agency—and that in doing so, they operate outside of the state-causal nexus altogether. Others will instead contend that humans and other critters never exhibit genuine agency at all, and that experiences of agency are systematically illusory and non-veridical. Were I among those who embrace the dual contention, I would opt for the second variant of it. I would do so because I believe that there is excellent scientific evidence for the thesis of physical state-causal completeness vis-à-vis humans and other critters, and also because I contend that willing-experiences are proximal state-causes of the bodily motions one experiences as one’s actions. I would claim too that experiences of agency are a very useful 278

Conditions of agentive phenomenology

systematic illusion, from an evolutionary perspective—that nature instilled them in order to render very salient the differences between phenomena that are susceptible to effective influence by bodily motions that are state-caused internally within oneself, and phenomena that are not thus susceptible. (In Horgan 2014, I argue that color experience is an evolutionarily useful illusion, and that color “presentational content” is systematically non-veridical even though color “judgmental content” is not.) But I repudiate the dual contention lately mentioned, in favor of the position I have been defending here. I contend that genuine, full-blooded, agency is a state-causal phenomenon in which one’s willings, with their inherent aspects of self-sourcehood and optionality, are proximal mental state-causes of the bodily motions one experiences as one’s actions. (I  also contend that such psychological state causation is implemented by neuro-physical state causation, and indeed that physical causal completeness obtains for humans and for all other critters.) Although I think it is possible to believe that color experience is systematically non-veridical even though one cannot help undergoing it, and although indeed I do believe this, to me it seems literally impossible to believe—really and thoroughly believe—that I am not an agent of the kind I experience myself to be. I would be loathe to embrace a philosophical position entailing that I ought rationally to believe what is psychologically impossible for me to believe.

Martine Nida-Rümelin: rejoinder II Horgan’s response to the argument from alienating experience (1) Horgan has an excellent reply to the argument from alienating experience. After the relevant change of her agentive experience, Alexa fulfills three conditions: (a) she is capable of acting, (b) in acting she experiences her movement involved in her own acting as statecaused by her mental states, and (c) that experience is alienating. Each of these conditions is necessary for the example to serve its role in the argument from alienating experience. According to Horgan’s refined view, condition (c) is fulfilled since—as a consequence of fulfilling (b)—Alexa does not have the experience of being active. It follows—according to the refined version—that Alexa violates condition (a) since the experience of being active is essential to acting. It follows, furthermore, that it is impossible to construct a similar example fulfilling both (a) and (c). If the relevant experience of one’s behavior as state-caused is alienating because that experience excludes the experience of being active, then both (c) and (a) cannot be satisfied. My original argument thus has no bite against Horgan’s refined view. However, I remain convinced that his proposal does not do duty to the anti-illusionist motivation we share. I will argue that his theory, contrary to his own intention, implies that the genesis of our own behavior in acting is, in a sense I will explain, radically different from how it appears in agentive experience. This may be surprising since his theory also implies that agentive experience is largely veridical. Before I present my new argument for the claim that Horgan’s proposal is, as one may say, a hidden illusion theory, I would like to make a few preliminary remarks. They will help clarifying what follows.

Willings and appearances—a misleading terminological choice (2) Horgan introduces the term “willings” as a name for the experience of being active where that experience has two aspects, the one of ‘self as source’ (the OBA-aspect) and the one of 279

Terry Horgan and Martine Nida-Rümelin

‘optionality’. Given that terminology, Horgan’s claims about the experience of being active (EA1, EA2, and EA3 below) turn into claims about willings (W1, W2, and W3 respectively): EA1: The presence of an experience of being active is essential to action. EA2: The experience of being active causes the movements involved in acting. EA3: The relevant movement in a given action being state-caused by the experience of being active is a veridicality condition of the experience of being active. W1: The presence of a willing is essential to action. W2: Willings cause the movements involved in action. W3: A veridicality (or rather, satisfaction) condition of a willing (related to a given action) is that the willing causes the associated movement. “Willing” is an introduced term. We already have a pre-theoretical understanding of what willings are. Therefore, the terminological choice here at issue is not a harmless stipulation. Turning EA1, EA2, and EA3 into W1, W2, and W3 invites the reader to test her intuitions regarding EA1–EA3 using her implicit understanding of willings. To do so is however inadequate and leads to unfounded intuitions if, in reality, what we normally refer to using the term “willings” is distinct from experiences of being active. In fact, it is quite clear that this is so. Let me explain why. One may express an important difference between willings and experiences of being active using Searle’s famous notion of direction of fit. Willings have a world-to-mind direction of fit. Appearances have a mind-to-world direction of fit. Appearances, but not willings, present things to the subject as being the case. In experiences of being active, one’s movements appear to have a certain genesis; they appear to be brought about by oneself. Such appearances are not directed at something still to happen as willings are. These remarks should suffice to make evident that the agentive experience of being active is not a willing. Given the difference between experiences of being active on one side and willings on the other, one should not confuse the first three claims EA1–EA3 with the corresponding claims about willings W1–W3. Testing one’s intuitions with respect to the former by thinking about the latter is illegitimate. In fact, it will lead one astray. For instance, EA3 is a surprising claim with little intuitive appeal. Not so W3. It is plausible to think that a willing is only satisfied if the willing itself plays a causal role in bringing about what the subject wills. Therefore, testing EA3 by considering W3 will lend credibility to the former claim, which it does not merit. Similar remarks apply to EA2. One implicitly thinks of willings as something the subject does in an active manner. If one thinks of willings as active, then W1 and W2 may appear to be well motivated. One might tend to assent to W2 having in mind that movements in acting are active due to being caused by active willings. Implicitly presupposing that this is the only sense in which a subject can bring about her movements in an active manner, one then might tend to assent to W1 as well. Such intuitions may have their own merits but they surely cannot support the corresponding claims about experiences of being active. Intuitions about willings (in the already introduced sense) and intuitions about experiences of being active have distinct subject matter. I therefore will resist adopting Horgan’s proposed terminology.

Veridicality of agentive experience and the nature of being active (3) I interpret Terry Horgan’s proposal as implying a claim about what it is to be active, about what actively bringing one’s own movement about consists in. This will play a role in my 280

Conditions of agentive phenomenology

argument below. Horgan does not explicitly formulate any such ontological claim. Nonetheless, or so I claim, he is committed to the thesis that for a subject to be active in a given movement consists in that movement being caused by the subject’s experience of being active. My argument for this interpretation is simple but I need to clarify a small point before I go on. Horgan speaks of experiences of being active as having two aspects: the aspect of apparently bringing the movement about and the aspect of optionality. I would like to leave the second aspect apart for the present discussion. I will focus on the former. Furthermore, I reserve the term “experiences of being active”—as I did already in the discussion so far—to agentive experiences with the former aspect. I leave it open whether such experiences always include the aspect of optionality. If so, this does not change anything for what follows. Here is my argument for my ontological interpretation mentioned above. We all have an intuitive understanding of an important and fundamental distinction between changes of our own body or mind which happen to us (changes we passively undergo) on the one hand and changes we bring about ourselves (changes in which we are active) on the other. Not all philosophers agree with this first observation but I believe that Horgan has no objection. Furthermore, for a philosopher who concedes that we experience ourselves as active it seems unescapable to admit that to be active in one’s own movement is precisely what renders such experiences of being active veridical. One may add that such experiences are our primary access to what it is to be active but this further plausible assumption which, I suppose, is in accordance with Horgan’s ideas is not required for my argument. Let us simply suppose the following:

Thesis AV (relating being active to the veridicality of experiences of being active) S is active in her movement M is the veridicality condition of S’s experience of being active in M. Horgan does not talk about the nature of being active. He only expresses himself with respect to the veridicality condition of experiences of being active. However, if one accepts Thesis AV, then proposing a view about the condition which renders experiences of being active veridical and proposing a view about the nature of being active amount to the same. This is why I take it to be legitimate to presuppose in what follows that Horgan is committed to the following ontological claim.

Thesis NA (about the nature of being active) The condition expressed by the following sentence (S1) consists in the fulfillment of the condition expressed by the following sentence (S2): (S1) S is active in her own movement M. (S2) S’s experience of being active in M state-causes M.

Veridicality and the phenomenally manifest (4) I need to attract the reader’s attention to a further feature of Horgan’s view, which might go unnoticed but will be crucial to my main argument against his proposal. One might think 281

Terry Horgan and Martine Nida-Rümelin

that the veridicality conditions of a given experience are (or should be) most closely related to what appears to be the case to the subject undergoing the experience. If a thing appears to be spherical to me in a visual experience I am undergoing, then this justifies the claim that this thing being spherical is among the veridicality conditions of my present visual experience. This direction (the conclusion from what appears to be the case in a given experience to its veridicality conditions) is quite trivial. To deny it would be to miss the very notion of veridicality conditions. My own view is that the opposite direction should be accepted as well but I will not insist on this opinion nor presuppose it here (see for an argument in its favor Nida-Rümelin 2018b). Horgan’s view must be that the opposite direction does not hold. To block the conclusion from veridicality conditions to what appears to be the case is the only way to make sense of the following two central claims of his proposal which otherwise would be plainly in contradiction.

Thesis VCA (about the veridicality condition of human experiences of being active) The veridicality condition of the experience of a subject S of being active in her movement M is that M is state-caused by that very experience.

Thesis IE (about incompatible experience) Human subjects cannot experience themselves as active in a given movement M and, at the same time, experience that movement as state-caused by their own experience of being active. The relevant causal relation is thus the veridicality condition of experiences of being active and yet, in having these experiences it does not appear to the subject, on the phenomenally manifest level, as if her own movement were caused by her own experience. Horgan’s view thus presupposes a distinction between two kinds of veridicality conditions (related although different distinctions are developed in Chalmers 2006 and Nida-Rümelin 2006). Veridicality conditions in the first sense, I will call that sense “the strong sense,” can serve to describe the phenomenology of the experience at issue. They are manifest at the phenomenal level. In other words, if p is a veridicality condition in the strong sense, then p appears to the subject to be the case. Being active in one’s movement is a veridicality condition in the strong sense of experiences of being active. The perceived object’s being red is a veridicality condition in the strong sense of the corresponding kind of color experiences. Veridicality conditions in the weak sense are (or are supposed to be) relevant to assess whether the experience presents the world as it is (whether it is veridical). However, talking about how the experience presents the world in that context must not be misread as telling us what appears to be the case to the subject at the phenomenally manifest level. Arguably, a surface having a certain reflectance profile is a veridicality condition in the weak sense of visual experiences in which an object appears to be red. A movement being caused by the experience of being active is, according to my interpretation of Horgan’s view, a veridicality condition in the weak (and not in the strong sense) of experiences of being active. I suppose that the distinction between veridicality conditions in the weak and in the strong sense corresponds roughly to the distinction drawn by Horgan between content-based and implementational veridicality conditions. I take content-based veridicality conditions in Horgan’s sense to be determined by the phenomenology of the experience alone, while 282

Conditions of agentive phenomenology

implementational veridicality conditions are (at least in part) determined by facts about the experience’s causal role. If so, then the class of content-based veridicality conditions for a given experience should coincide with what I call veridicality conditions in the strong sense and the class of implementational veridicality conditions should either coincide or be part of what I call veridicality conditions in the weak sense. However, I do not presuppose this close relation between the two distinctions in what follows.

A regress problem (5) If my argument in section (3) of the present reply goes through, then Horgan is committed to the following claim about the nature of being active already stated before.

Claim NA (about the nature of being active) The condition expressed by the following sentence (S1) consists in the fulfillment of the condition expressed by the following sentence (S2): (S1) S is active in her own movement M. (S2) S’s experience of being active in M state-causes M. NA explains what it is to be active in terms of experiences of being active. This raises immediately a worry of a well-known type. (For a similar argument applied to dispositionalism about color, see e.g. Levine 2006.) If the term is used in the same sense on both sides, then we can plug the explanation given by NA of what it is to be active into (S2) and we get the following result.

Claim NA’ (about the nature of being active) The condition expressed by the following sentence (S1) consists in the fulfillment of the condition expressed by the following sentence (S2′): (S1) S is active in her own movement M. (S2′) S′s experience [of M being state-caused by her own experience of being active] state-causes M. We can apply the same procedure to the occurrence of “being active” in S2′ and we realize that we landed in an infinite regress. As far as I can see, there are only two ways to avoid the regress. The first is to deny that the occurrences of “active” in (S1) and (S2) are uses of the term “active” in the same sense. The second is to forbid the relevant replacement in the context of the attribution of an experience. The first way out is not available if one subscribes to thesis AV motivated in section 3. According to AV, for a movement to be actively brought about by the subject concerned just is for it to fulfill the veridicality condition of experiences of being active. It is hard to see how one can accept AV and yet deny that “active” is used in the same sense in (S1) and (S2) above. I therefore put the first reply aside. The second way out looks more promising. One may argue that attributions of experiences create intensional contexts analogously to belief attributions. One may take the example of color experience to illustrate the claim. Suppose that to be red is to have a surface with a certain reflectance profile R.4 Arguably it is not an argument against such an objectivist 283

Terry Horgan and Martine Nida-Rümelin

account of color that we cannot replace “appears red” by “appears to have reflectance profile R” in attributions like “The apple appears red to Maria.” One may argue that replacing the property referred to by “to be red” by the—supposedly—co-referring term “to have reflectance profile R” in such an attribution changes the truth value of the sentence because colors are not experientially given as reflectance profiles although they actually are such profiles (according to the ontological claim here presupposed). One may argue that an analogous point applies to experience attributions like the one formulated by (S3): (S3) Maria experiences herself as active in M. According to this view, Maria experiences herself as active but she does not experience her movement as state-cased by that very experience despite the fact that to be active in one’s movement is nothing but that movement being caused by that very experience. Such a replacement would change the truth value of the attribution because it is not experientially present to the subject undergoing an experience of being active what being active consists in. Even if (S3) is true of Maria, this does not imply that she is, in that experience, under the impression that her own experience of being active state-causes the movement. Attributions of appearances (that is attributions of experiences of what appears to be the case to the subject concerned in undergoing a given experience) thus create intensional contexts in a certain sense. One may explain that sense as follows. If the propositions p and q are rendered true by the very same state of affairs (by the very same ‘worldly’ conditions), then the replacement of q by p in an attribution of an appearance may nonetheless change the truth value of the appearance attribution. As far as I can see this is how Terry Horgan would have to respond to the above-formulated regress problem. To respond in this manner amounts to accepting the above distinction between veridicality conditions in the weak sense and in the strong sense. In a case where replacing p by q turns a true appearance attribution into a false appearance attribution, p is a veridicality condition of the experience in the strong sense and q is a veridicality condition of the experience only in the weak but not in the strong sense. If no other plausible way out of the regress problem can be provided within Horgan’s account, then this is a further reason to say that he should accept the proposed distinction between strong and weak veridicality conditions.

Incompatible experiences versus experiences of apparent incompatibles (6) I need just one further element to prepare my main objection. Horgan concedes in his response to my argument from alienating experience that the experience of being active is, at least for humans and perhaps more generally, incompatible with the experience of one’s own movement being state-caused. I agree that this is most likely the case. However, I would like to propose a different claim in the neighborhood, which I find phenomenologically convincing. Experiences of being active in a given movement on the one hand and of the relevant movement being state-caused on the other are experiences of apparent incompatibles. By experiences of apparent incompatibles, I mean this: a pair of experiences belongs to that kind if the subject undergoing these experiences is under the impression that the two experiences present conditions to obtain which cannot both be satisfied. In other words, on the phenomenally manifest level, the contents of the two experiences appear to be incompatible. To illustrate the phenomenological notion of experiences of apparent incompatibles, one may choose the example of visual shape perception. In certain phenomenal types of visual 284

Conditions of agentive phenomenology

experiences, the subject undergoing the experience is under the impression of being presented with a spherical object. In a different phenomenal type of visual experiences, the subject is visually under the impression of being presented with a cubical object. Here is my phenomenological proposal: a subject having an experience of an object as being spherical and of a different object as being cubical is thereby under the impression that no object could possibly have both shapes at once. The incompatibility of the two shape properties a subject is presented with in such a case is available at the phenomenal level. The incompatibility of the two shape properties one is presented with in such experiences is phenomenally manifest. Before I go on, I would like to note that not all incompatible experiences (for a given species) are experiences of incompatibles. For instance, humans are unable (under normal circumstances) to experience an object as reddish and greenish at once and in the same place. (See Nida-Rümelin and Suarez 2009 for a detailed discussion of supposed real counterexamples in unusual circumstances.) However, there is no phenomenological ground for supposing (as far as I can see) that experiences of being greenish and being reddish are experiences of apparent incompatibles. Furthermore, experiences of apparently incompatible conditions may nonetheless occur. One can be presented, in perception, with a situation that appears to satisfy conditions, which, at the same time, appear to be incompatible.5 Reflecting on the example concerning visual shape perception, a further observation comes to mind. It is not a contingent fact about visual experiences of something being spherical and of something being cubical that they are experiences of apparent incompatibles. If for some reason we were to suppose that a subject had an experience of something being spherical and of something being cubical and yet did not thereby have the impression that nothing could have both properties at once, then we surely would have to conclude that we were about to mis-describe the subject’s visual phenomenology. It is part of what it is to have experiences of these phenomenal types that the properties presented in those experiences appear to be incompatible. This observation motivates introducing a further phenomenological term: some pairs of experiences are essentially experiences of apparent incompatibles. There is a further phenomenally salient aspect to which I would like to attract the reader’s attention using the same example. When you see an object as spherical and another object as cubical, the incompatibility of the properties you are presented with appears to be visible. You simply seem to see that nothing can have both properties at once. One might put it in this way: you seem to have sufficient access to the nature of the properties just by seeing them in order for you to also have visual access to their incompatibility. Your experience of those properties as incompatible appears to be due to their very nature, where that nature is (or rather: appears to be) in part available at the phenomenal level. Let us say of such cases that the apparent incompatibility (of the relevant properties or of the relevant conditions) is experientially evident in having the experiences at issue.

More about what appears to be the case in agentive experience (7) I would now like to propose a few phenomenological observations about the case of agency. I will have to leave it to the reader to test their plausibility based on his or her own phenomenological reflection: PO1: Experiences of oneself as active in a given movement and experiences of that movement as being state-caused are (for human beings) experiences of incompatibles. PO2: Experiences of oneself as active in a given movement and experiences of that movement as being state-caused are essentially experiences of incompatibles. 285

Terry Horgan and Martine Nida-Rümelin

PO3: The incompatibility of oneself being active in a movement and one’s own movement being state-caused is experientially evident in any relevant potential agentive pair of experiences. A lot would have to be said to justify these phenomenological claims by careful reflection. An obvious difficulty resides in the fact that we actually never have experiences of our own movements in actions as state-caused. I presuppose that we can nonetheless imagine in a sufficiently clear manner how that would be like. I cannot go deeper into the phenomenology of agency in order to convince the reader of PO1–PO3. I hope that these claims have enough intuitive appeal for it to be of some interest to see what happens to Horgan’s proposal if one takes them to be true. Before I discuss these consequences, let me explain in greater detail what these claims state. One may wonder if experiences of oneself as active on the one hand and experiences of one’s own behavior as state-caused on the other are analogous to the case of experiences of reddish and greenish or rather analogous to the case of visual impressions of different shapes. My claim is that they are analogous to the latter and not to the former. They may be incompatible for humans (like the color experiences under consideration) but this is not all. They are, furthermore, like the pair of shape perceptions mentioned. It is part of their phenomenal character that—in having an experience of being active and in having an experience of your movement being state-caused (perhaps in the next moment like in the case of Alexa)—you are under the impression that the same movement could not possibly fulfill both conditions. This is the content of PO1. However, we can go further. Reflecting on what it is like to be active in one’s movement and on what it would be like to experience one’s own movement as state-caused, one will be led to a further and stronger conclusion. Suppose that a creature has experiences of type E1, which we are inclined to consider as experiences of being active in her movement and that it also has experiences of type E2, which we are inclined to consider as experiences of her own movement as being state-caused. Suppose we learn, in addition, that the creature does not experience what appears to be the case in E1 as incompatible with what appear to the case in E2. Then, or so I claim, we have reason to resist the former inclinations. If the appearance of incompatibility is lacking, then either experiences of type E1 are very different from our experiences of being active and thus do not deserve to be classified with them or experiences of type E2 are very different from experiences of being state-caused. This is what PO2 affirms. Suppose that PO1 is correct. Then PO3 is quite obvious, I would say. If the two conditions are given in your experience as incompatible, then you surely have the impression that these conditions are in fact incompatible by their nature. You surely are under the impression that you are aware, experientially, of their actual incompatibility. It seems to you that you have experiential access to what it is to be active in your movement and to what it is for a movement to be state-caused which allows you to experience their incompatibility. If you have doubts about that claim, I ask you to reconsider the shape case first in order to judge whether you can agree that this is a case of experientially evident incompatibility in the sense introduced. If you come to a positive result, you might be in a position to admit that the agentive case we are interested in is analogous to the shape case in that respect.6

Veridicality without genuinely veridical appearance (8) We are now in a position to see why and how Horgan’s proposal does not do duty to his own intention. We would like to be able to rationally endorse a theory about agency, which allows us to assume that our own behavior comes about in the way it appears to come about 286

Conditions of agentive phenomenology

in agentive experience. When a subject experiences her movements as actively brought about by herself then, in undergoing that experience, her movements appear to be generated in a particular way. We would like to develop a well-founded and scientifically acceptable theory about agency according to which the subject’s movement is actually generated in the way it appears to be generated in those experiences. This, I believe, is a central motivation behind Horgan’s work about agency, which I fully endorse myself. Horgan’s view saves the claim that the experience of being active is veridical. However, given the special situation in which we are if (a) his refined view is correct, if (b) the claim above about the ontology of being active (which I attribute to Horgan) is correct, and if (c) my additional phenomenological observations are adequate, then such veridicality is not enough. In that particular situation, there is a huge gap between the experience being veridical and the experience being such that things appear as they are to the subject undergoing the experience. Let us first leave my new phenomenological observations apart and let us focus on (a) and (b). We thus assume Horgan’s view and we presuppose that it commits him to the ontological claim that to be active in a movement M consists in the movement being state-caused by one’s experience of being active. In that case, our situation in agentive experience has the following features: F1 In agentive experience, we experience ourselves as actively bringing about the relevant movement M. F2 To actively bring about one’s own movement consists in the satisfaction of a certain condition C (which is that the subject has an experience of being active which state-causes the movement). F3 We are experientially blinded to condition C in the following sense: we cannot act and at the same time undergo an experience, which is such that condition C appears to be satisfied to the one who has that experience. F4 Condition C is experientially accessible in this sense: there are potential experiences E, such that in undergoing E the subject is under the impression that condition C is satisfied. F5 Experiences of being active in one’s own movement are veridical (in cases where they are) in virtue of the satisfaction of C. The last feature, F5, is crucial to what I find unsatisfying about Horgan’s view. It is a consequence of F2 (the ontological claim). Condition C is what really underlies the veridicality of veridical experiences of being active. However, that condition plays no role in what is phenomenally manifest to the subject who experiences herself as active. It is only a veridicality condition in the weak sense but it is no veridicality condition in the strong sense. Even worse, although one could in principle have an experience in which C appears to be satisfied (an experience such that C is a veridicality condition in the strong sense of that experience), human beings are unable to have such experiences. They are unable to be phenomenally aware of what makes their experience of being active veridical. This is already enough to see that veridicality is a very weak condition in the present context. The relevant experiences are veridical in virtue of the fulfillment of a condition, which does not appear (and could not appear in the case of humans) to be fulfilled to the subject concerned. Things become worse if we assume, in addition, that the above-motivated phenomenological observations are adequate. If they are, then our incapacity of having experiences with C as their veridicality condition in the strong sense prevents us from undergoing a very disturbing experience. Suppose a subject were given that capacity perhaps by removing a 287

Terry Horgan and Martine Nida-Rümelin

certain suppressing mechanism in the brain, which has evolved in members of the human species ‘in order’ to guarantee certain survival advantages. That person, call her Aliena, has the experience of being active when she moves in her actions and she also has the experience of that movement being state-caused by her experience of being active. If what I said before is correct, then Aliena cannot but have the impression that she is in an impossible situation. To have these two experiences necessarily involves that the conditions that appear to be satisfied appear to be incompatible. Like Alexa, Aliena would have an alienating experience but for different reasons. Her experience would be alienating, not in virtue of a lack of the experience of being active (as it is for Alexa); it would be alienating in virtue of the apparent satisfaction of a further condition which appears to be incompatible with how things appear in her experience of being active. We now can say this. We are experientially blinded to what makes our experiences of being active veridical. If we were freed from that blindness then we would be, necessarily, under the illusion that we are in an impossible situation, in a situation in which our movements are generated in two incompatible ways. Furthermore, according to PO3, we would then have the impression that the two conditions are, by their experientially accessible nature, incompatible. We would suffer the illusion that our experience gives us sufficient access to their nature for us to be aware, on a phenomenally manifest level, of their real incompatibility. One may even go further and argue for a stronger claim. When visually experiencing an object as spherical one is aware on a phenomenally manifest level of the object having a shape, which is incompatible with being cubical. So, not only are visual experiences of being spherical and of being cubical experiences of incompatibles in the above introduced sense, the relevant incompatibility is phenomenally present even when a subject has one of those experiences at a given moment and not the other. Arguably, a similar observation is in order with respect to experiences of being active. In experiences of being active, a condition appears to be satisfied which appears to be, due to its experientially accessible nature (accessible by that very experience), incompatible with being state-caused. If so, then experiences of being active are non-veridical after all. In having them we suffer the illusion that the way our behavior is generated is incompatible with a condition C which, in reality, renders the experience of bringing something actively about veridical. This is a radical illusion indeed. My above argument is supposed to show that Horgan’s proposal cannot satisfy its underlying motivation. The strength of the result obtained depends on how many premises one uses. Introducing only the additional ontological thesis about what being active consists in we reach the following result: the veridicality of veridical experiences of being active is due to the satisfaction of a condition C, which is experientially inaccessible to humans. Being active consists in C, but experiences of being active provide no access whatsoever to what being active consists in. C, and thus the condition which renders experiences of being active veridical, is totally absent at the level of experience. C is absent in human agentive experience despite the fact that it is, in principle, experientially accessible. Introducing the phenomenological claims PO1–PO3, we arrive at a more disturbing conclusion. A subject who is able to be phenomenally aware of being active and, at the same time, of the fulfillment of the condition C, which in fact constitutes being active, would be—in having both experiences— under the illusion that her behavior emerges in two incompatible ways. Adding a further phenomenologically plausible claim, we come to the most puzzling result: in experiences of being active, we suffer from a serious illusion about what it is to be active. In such experiences, our behavior appears to be generated in a way, which is incompatible with C while in reality C is even what constitutes being active and what therefore renders the experience of 288

Conditions of agentive phenomenology

being active veridical. Furthermore, we suffer from the illusion of being immediately aware of that incompatibility by having experiential access to what it is to be active.

An alternative approach (9) In order to avoid the problems highlighted in what precedes and in order to do duty to the motivation behind Horgan’s view, one may take a more radical decision and fully endorse what appears to be the case in agentive experiences with the specific aspect of being active in one’s movements. The following conditions appear to be fulfilled in experiences of being active: A1: The subject brings her movement about in an active manner. A2: The experience of being active provides partial access to what it is to be active. A3: That experiential access provides one with the capacity to be phenomenally aware of the real incompatibility between (a) a movement being brought about in an active manner by the subject and (b) the movement being state-caused. Such an account thus includes the claim that these conditions are actually fulfilled. It is thereby in contradiction with the received view that all physical events have only physical events as causes. Many philosophers, including Terry Horgan, are convinced that, in virtue of that feature, an account of this kind is seriously undermined by scientific wisdom. I doubt that this is true but will not start discussing that further issue here. (Compare for a discussion of that issue Nida-Rümelin 2016.) Horgan’s proposal is meant to fulfill the following two constraints: Constraint 1: The theory is compatible with the dogma that physical events have only physical events as causes. Constraint 2: The theory is compatible with the claim that our behavior appears to be generated in our experiences of being active in the way it actually is generated. Agentive experience is not radically and systematically illusionary. Based on the considerations presented in this part of our paper, I suspect that no account fulfilling both constraints is actually available. One must choose, I believe, between two options: respecting constraint 1 and abandoning constraint 2 or respecting constraint 2 and abandoning constraint 1. I would recommend the second option. I wonder what choice Terry Horgan would prefer if he agreed on my diagnosis that the choice is unavoidable.

Terry Horgan: proposal III Martine Nida-Rümelin’s latest installment contains a challenging new argument for her claim that my position “does not do duty to the anti-illusionist intuition.” The new argument is in much the same spirit as her earlier argument that was keyed to her thought experiment about Alexa, but does not deploy that thought experiment. Largely because of Nida-Rümelin’s ongoing objections, I now believe that my position needs to be refined yet further, above and beyond my prior refinements earlier in this dialogue. I also believe that the needed additional refinements are independently plausible and therefore should be embraced anyway, quite apart from the role they will play in my reply to her new argument. 289

Terry Horgan and Martine Nida-Rümelin

My earlier articulations of my position put heavy emphasis on the contention that although agentive phenomenology is not as-of one’s action being a psychological state-causal process, agentive phenomenology also is not as-of one’s action not being a psychological state-causal process. I stand by this contention, but I now realize that it can be refined in either of two incompatible ways. (Both ways are compatible with the contention itself, despite being incompatible with each other.) One way is to embrace the claim that agentive phenomenology has no negative content at all involving psychological state causation. But a second way is to acknowledge that agentive phenomenology is indeed as-of one’s action not being a certain kind of psychological state-causal process, while also claiming nonetheless that agentive phenomenology is not as-of one’s action not at all being a psychological state-causal process, I believe that my position should be refined the second way, not the first. Doing so is independently plausible, because this alternative accords better with the introspectable character of agentive phenomenology itself. Doing so also will enable me to resist Nida-Rümelin’s new argument against me. I begin with some preliminary remarks concerning satisfaction conditions. I now think that in principle, there can be three kinds: content-based, noncontingently implementational, and contingently implementational. Content-based satisfaction conditions are determined directly by the intentional content of the phenomenology: either they are literally part of that content, or else they are entailed by it as a matter of logic and/or analytic truths. Noncontingently implementational satisfaction conditions are not content-based, but are entailed by the content of the phenomenology plus synthetic necessary truths. Contingently implemental satisfaction conditions are neither content-based nor noncontingently implementational, and instead are dependent in part on contingent truths about the nature of the given kind of cognitive agent. I also embrace a distinction that is somewhat orthogonal to this three-way distinction: viz., between satisfaction conditions that are phenomenally manifest, and ones that are phenomenally non-manifest. This is Nida-Rümelin’s terminology, which she also aligns with ‘weak sense’ and ‘strong sense’. Another preliminary matter is that she objects to my use of the expression ‘willing’ as an allegedly misleading terminological choice. I think this concern can be set aside for present purposes, by just changing my terminology. So I will instead use the expression ‘agentive-instigation experience’. There is one more preliminary matter, before I turn to the main business at hand. She wonders whether I face a regress problem. She suggests that I do not, as long as I say that attributions of appearances create intensional contents—but that this commits me to the distinction between the weak and strong senses of ‘satisfaction condition’. I’m fully on board with all this. I turn now to further elaboration of the position I wish to espouse. Doing so involves attending carefully to refinement R5 from my second installment above, and at the same time acknowledging and embracing certain claims about negative satisfaction conditions of agentive phenomenology concerning matters of psychological state causation of one’s bodily motions—and all the while keeping clear about which claims are, and which claims, are not, phenomenally manifest. Let a process be experiential-acquaintance-wise state-causal (for short, EA-wise state-causal), just in case it is state-causal in a way with which humans are experientially acquainted. This feature will prove quite important in what follows. In particular, the following consequence of R5 will be important: all EA-wise state-causal processes are non-agentive state-causal processes. I hereby incorporate into my position the following additional claims, which I list under three categories: phenomenally manifest facts, facts about the limits of reliable introspectability, and phenomenally non-manifest facts. 290

Conditions of agentive phenomenology

Phenomenally manifest facts PM1. Any experience one can have as-of a process’s being state-causal is as-of that process’s being non-agentively state-causal. PM2. Any experience one can have as-of a process’s being not state-causal is as-of that experience’s being not non-agentively state-causal. PM3. A satisfaction condition of an experience as-of agentively moving one’s body a certain way is that the phenomenon thus experienced includes an agentive-instigation experience vis-à-vis the bodily motion. PM4. In experiences of being active, a condition appears to be satisfied which appears to be incompatible with one’s bodily motion being EA-wise state-caused by one’s agentive-instigation experience. PM5. A (negative) satisfaction condition of an experience as-of agentively moving one’s body a certain way is that the motion is not EA-wise state-caused by one’s agentive-instigation experience.

Facts about the limits of reliable introspectability L1. One cannot reliably ascertain introspectively whether or not an experience as-of a process’s being not EA-wise state-causal is an experience as-of that process’s being not state-causal at all. L2. One cannot reliably ascertain introspectively whether or not an experience of agentively moving one’s body a certain way is an experience as-of the process’s being not statecaused at all by the agentive-instigation experience.

Phenomenally non-manifest facts PNM1. Not all state-causal processes are non-agentively state-causal; some are agentively state-causal. PNM2. A satisfaction condition of an experience of moving one’s body a certain way is that the agentive-instigation aspect of the experience state-causes the appropriate bodily motion. PNM3. It is not the case that in experiences of being active, a condition appears to be satisfied which appears to be incompatible with one’s bodily motion being state-caused by one’s agentive-instigation experience. PNM4. It is not the case that a (negative) satisfaction condition of an experience as-of agentively moving one’s body a certain way is that the motion is not state-caused at all by one’s agentive-instigation experience. I have several comments about these various claims, by way of further elaboration. Concerning PM3: Although an experience as-of moving one’s body a certain way is trivially an occurrence condition of itself (because such an experience cannot occur without the presence of an instigation-experience as a component), it is not trivial to say that such an experience is a satisfaction condition of itself. (By contrast, consider for example a conscious belief such as the belief that Aristotle was a teacher of Alexander the Great; this state’s satisfaction condition—viz., that Aristotle was a teacher of Alexander the Great—would be met even if the state itself did not exist.) The claim that genuine agency requires agentive phenomenology is substantive and non-trivial. 291

Terry Horgan and Martine Nida-Rümelin

Concerning PM4 and PM5: These claims explicitly acknowledge that agentive experience includes negative content regarding state causation: agentive experience is phenomenologically as-of one’s action being not state-causal in a certain way (viz., being not EA-wise state-causal). Nida-Rümelin’s objections to my earlier articulations of my position have helped me appreciate that agentive experience does indeed have this kind of negative content, and that its having such content is introspectively manifest. I should have acknowledged such negative content all along, rather than emphasizing so heavily the fact that “not as-of Φ” is logically weaker than “as-of not Φ.” Concerning PNM3: Although I said in an earlier installment that I think this is an implementational satisfaction condition, I am now not sure about this. For the present I can be neutral between the following three alternatives: (a) it is actually a content-based satisfaction condition, albeit one that is not phenomenally manifest; (b) it is a non-contingent implementational condition (a view one might embrace, for instance, because one holds that it is a synthetic necessary truth that the only fully intelligible account of agency will incorporate this satisfaction condition); or (c) it is a contingent implementional satisfaction condition that perhaps is specific to certain kinds of creatures (including humans). Finally, concerning all of the theses PM1–PM5, L1–L2, PNM1–PNM4 and the question “Why should it be that these theses are true?”: I would again reiterate my remarks at the end of my second installment above, about the bear and the berries. There are plausible evolutionary-biological reasons why the only kinds of state-causal processes that humans can experience as state-causal are non-agentive state-causal processes. Maybe there are further reasons too why all of these theses are true—a matter of much philosophical interest, in my view. Maybe no realistically possible kind of intelligent creature can be capable of appreciating practical reasons qua reasons, and of acting upon reasons qua reasons, unless such a creature resides phenomenologically within an experiential “space of reasons” that is (and must be) experientially outside the “space of causes”—even though the creature does not (and cannot) reside metaphysically outside the “space of causes.” Sellars (1962) looms large here, and it is worth juxtaposing Sellars’ own talk of the “manifest image” of human beings with Nida-Rümelin’s talk of what is “phenomenally manifest” in experience. I am now ready to reply directly to Nida-Rümelin’s principal argument, in her most recent installment. As nicely summarized at the end of section 8 of that installment, the argument comes in three stages. Each stage involves an alleged bad consequence of my position. Also, each successive stage appeals to additional premises beyond those at the previous stage(s), and involves an alleged consequence even worse than the alleged bad consequences at the previous stage(s). I will address the three stages in turn. The first stage focuses primarily on what she calls condition C, which I now have formulated as PNM2. She says this about what I am allegedly committed to by embracing this satisfaction condition: Being active consists in C, but experiences of being active provide no access whatsoever to what being active consists in. C, and thus the condition which renders experiences of being active veridical, is totally absent at the level of experience. C is absent in human agentive experience despite the fact that it is, in principle, experientially accessible. I have three points in response. First, I deny that on my account, experiences of being active provide no access whatsoever to what being active consists in. An important part of what being active consists in

292

Conditions of agentive phenomenology

is the presence of agentive-instigation experience itself; and this part is phenomenally manifest. Second, I deny that in the pertinent sense, C’s obtaining “is, in principle, experientially accessible.” The only kinds of state-causal experiences that are experientially accessible to humans, I contend, are non-agentive state-causal processes. Third, although I do contend that C itself is an experientially non-manifest satisfaction condition of agentive experience, I deny that this makes my view objectionably close to an illusionist position. Agentive-instigation experience is itself phenomenomenally manifest, after all. So is the experiential aspect of one’s body’s moving the way it does because of one’s having instigated that motion. Admittedly, this “rational becausal” aspect is not experienced as state-causal, even though it is state-causal. But that is because the only kinds of statecausal processes that humans can experience as state-causal are non-agentive state-causal processes. Stage 2 invokes Nida-Rümelin’s phenomenological claims PO1–PO3, involving the phenomenally manifest incompatibility between experiences of oneself as active in a given bodily movement and distinct experiences of such a bodily movement as being state-caused. She says: Introducing the phenomenological claims PO1 to PO3, we arrive at a more disturbing conclusion. A subject who is able to be phenomenally aware of being active and, at the same time of the fulfillment of the condition C, which in fact constitutes being active, would be—in having both experiences—under the illusion that her behavior emerges in two incompatible ways. She sets forth a new thought experiment involving a character named Aliena, who Nida-Rümelin describes as suffering just such an illusion. She writes: Like Alexa, Aliena would have an alienating experience but for different reasons. Her experience would be alienating, not in virtue of a lack of the experience of being active (as it was for Alexa); it would be alienating in virtue of the apparent satisfaction of a further condition which appears to be incompatible with how things appear in her experience of being active. I have the following response. Although I grant Nida-Rümelin’s phenomenological claims PO1–PO3, I also claim that for humans, experiences of a phenomenon as state-causal are always as-of the phenomenon’s being EA-wise state-causal (thesis PM1). Likewise, experiences of a phenomenon as not state-causal are always as-of the phenomenon’s being not EA-wise state-causal (thesis PM2). So although it is true that experiences of oneself as active in a given movement, and experiences of that movement as being EA-wise state-caused, are essentially experiences of incompatibles, and although being active in a given movement is indeed incompatible with that movement’s being EA-wise state-caused, it is not true—cf. thesis PNM4—that experiences of oneself as active in a given movement are experiences as-of that movement’s being not at all state-caused. Thus, poor Aliena is not undergoing an experience which, according to me, accurately represents the psychological state-causal nature of human agency. On the contrary, she is having a (deeply alienating) experience that is simultaneously both (i) as-of being active in a given movement, and (ii) as-of that movement’s being EA-wise state-caused. No movement can be both.

293

Terry Horgan and Martine Nida-Rümelin

Concerning a hypothetical subject who is able simultaneously to be phenomenally aware of being active in a given movement and also to be phenomenally aware of that movement’s being state-caused by the subject’s agentive-instigation experience, I have four things to say. First, it would not be like Aliena’s experience. Rather, it would be both (i) as-of being active in a given movement, and (ii) as-of that movement’s being agentively state-caused. Second, we humans are not capable of such an experience, and cannot envision what it would be like. We are only capable of experiencing a process as state-causal by experiencing it as nonagentively state-causal. Third, I do not see why a (super-human) creature who could have such an experience—if a creature like this is possible, in some sense of ‘possible’—need necessarily find it alienating. Fourth, it is worth pondering, somewhat in the spirit of Kant-style reflection about transcendental ‘conditions of possibility’, whether any possible creature capable of appreciating and acting upon reasons, qua reasons, would necessarily need to be experientially blind to whatever processes within itself causally implement its rational agency. (Perhaps a condition of possibility for occupying “the space of reasons” is to be experientially outside of the “the space of causes.”) The upshot is the following, regarding stage 2 of Nida-Rümelin’s argument. Although I am indeed committed to claiming that we humans are experientially blinded to part of what makes our experiences of being active veridical, I am not committed to claiming that if we were freed from that blindness then we would be, necessarily, under the illusion that we are in an impossible situation. (This latter putative commitment is the alleged bad consequence of my position that supposedly arises in stage 2 of Nida-Rümelin’s argument.) Instead, if, per impossibile, we were freed from that blindness, then we would experience our bodily motions both as the products of our agency and as agentively statecaused by our agentive-instigation experiences—rather than experiencing them, in the manner of Aliena, both as the products of our agency and as non-agentively state-caused by our agentive-instigation experiences. Stage 3 invokes the claim that the incompatibility between being agentively induced and being psychologically state-caused is phenomenologically present “even when a subject has one of those experiences at a given moment and not the other.” Given this claim, and citing my contention that she calls “condition C” (viz., that the subject has an experience of being active which state-causes the bodily movement), Nida-Rümelin says, we come to the most puzzling result [to which Horgan is allegedly committed]: in experiences of being active, we suffer a serious illusion about what it is to be active. In such experiences, our behavior appears to be generated in a way which is incompatible with C while in reality C is even what constitutes being active and what therefore renders the experience of being active veridical. Furthermore, we suffer from the illusion of being immediately aware of that incompatibility by having experiential access to what it is to be active. The key contention here, I take it, is that a phenomenally accessible part of the content of agentive experience is that one’s motion is not state-caused by one’s agentive-instigation experience. This being so, the argument goes, my own position commits me—in spite of my intentions—to the claim that agentive experience is illusory. Such experience presents one’s behavior as generated in a way that is incompatible with C, despite C’s constituting the real nature of agency.

294

Conditions of agentive phenomenology

My response is essentially the same as my response to Stage 2. Although the experience of being active does have significant negative content regarding matters of psychological state causation, it does not actually have content that is incompatible with C. Rather, the experience of being active vis-à-vis a given bodily motion is, inter alia, an experience as-of that motion’s not being EA-wise state-caused by one’s agentive-instigation experience. There is no illusion here. On the contrary, being active really is incompatible with such EA-wise state causation, because EA-wise state causation is non-agentive state causation. I turn finally to some concluding remarks. By my lights, Nida-Rümelin thinks she can reliably introspect certain phenomenally manifest aspects of agentive experience which I claim are not present at all in agentive experience (and hence are not really phenomenally manifest). I have come to think, though, that the underlying mistake here is more subtle than I once thought it was. (And I have always thought that it’s quite subtle!) The real mistake, I now think, is not the putative mistake of conflating (i) the phenomenally manifest fact that agentive phenomenology is not as-of one’s action being a psychological state-causal process with (ii) the putatively phenomenally manifest fact that agentive phenomenology is as-of one’s action being not a psychological state-causal process. Rather, it is the mistake of conflating (iii) the phenomenally manifest fact that agentive phenomenology is as-of one’s action being not an EA-wise psychological state-causal process with (iv) the putatively phenomenally manifest fact that agentive phenomenology is as-of one’s action being not at all a psychological state-causal process. And this mistake is all the easier to make because one cannot reliably ascertain introspectively whether or not an experience as-of one’s action being not an EA-wise psychological state-causal process is also an experience as-of one’s action being not at all a psychological state-causal process. Suppose, though, that I am wrong about all this, and that she is right in claiming that we must choose between (on one hand) a broadly materialist metaphysics of mind and (on the other hand) the claim that our behavior really is generated in the way it appears in our experiences to be generated. She wonders, at the end of Section 9 of her recent installment, which of these options I would choose if I found myself accepting that this choice is unavoidable. My answer is this: I would choose the first option, embrace an error theory about agentive phenomenology, and then seek philosophical psychotherapy regarding my residual inability to believe—really believe—that I am not an agent of the kind I experience myself to be.7

Outlook: agreements, disagreements, and conflicting background assumptions We will conclude by summarizing both the principal contentions about which we agree and the principal points of disagreement between us. We agree on each of the following claims (labeled with ‘A’ for ‘Agreement’): A1 Human action typically has a distinctive phenomenology, including the aspect of action-instigation (self as source/of being active), that is richly intentional. A2 Because human agentive experience is intentional, it has satisfaction conditions. A3 In order for human agentive experience to be veridical on a given occasion of human behavior, its satisfaction conditions must be met on that occasion. A4 Human agentive experience normally is veridical; i.e., normally when such experience occurs, its satisfaction conditions are indeed satisfied.

295

Terry Horgan and Martine Nida-Rümelin

A5 At least some of the satisfaction conditions of agentive experience are phenomenally manifest—i.e., are manifest in the specific way that human behavior appears in human experience to be generated. A6 Humans are agents of the kind they experience themselves to be. A7 Any essential aspect of how human behavior appears in human experience to be generated is a phenomenally manifest satisfaction condition of human experience. A8 Condition C—the putative requirement that one’s experience of act-instigation statecauses one’s pertinent bodily motion(s)—is not a phenomenally manifest condition of human experience. A9 There is a prima facie incompatibility between (i) the way that human behavior appears in human experience to be generated and (ii) the contention that condition C is among the satisfaction conditions of human agentive experience. The principal matters of disagreement between us involve the following claims (labeled ‘D’ for ‘disagreement’): D.1 Condition C is among the satisfaction conditions of human agentive phenomenology (although neither C itself nor its being a satisfaction condition is phenomenally manifest). D.2 There is no genuine incompatibility between the way that human behavior appears in human experience to be generated and condition C. Terry Horgan affirms D.1 and D.2, whereas Martine Nida-Rümelin denies each of them. Horgan has been making the following background assumption throughout this dialogue (labeled ‘B’ for ‘Background’): B1 If condition C does not conflict logically with the manifest satisfaction conditions of human agentive experience, then there is no genuine incompatibility between (i) the way that human behavior appears in human experience to be generated and (ii) the contention that condition C is among the satisfaction conditions of human agentive experience. In light of B1 and A9, he therefore has been regarding the dialectical situation as imposing on him a two-part burden of proof: viz., the burden of arguing (1) that condition C does not conflict logically with the phenomenally manifest satisfaction conditions of agentive experience, and (2) that the prima facie incompatibility cited in A9—which Horgan has been regarding as a prima facie logical incompatibility—can plausibly be explained away as subtly mistaken.8 He has attempted to shoulder this burden above. One key question, of course, is whether Horgan has successfully discharged this two-part dialectical burden. Nida-Rümelin is not persuaded that he has done so. Others too might not be persuaded, and might even contend that among the phenomenally manifest satisfaction conditions of agentive experience is the requirement that one’s pertinent bodily motion is not state-caused by one’s act-instigation experience (or by any other psychological state of oneself ). But here at the end-stage of this dialogue, as we two have sought to reach agreement about the nature of our residual disagreement(s), it has become clear to us that perhaps our most fundamental dispute all along has really been about B1 itself. Nida-Rümelin has been making the following two background assumptions throughout this dialogue:

296

Conditions of agentive phenomenology

B2 In order for human behavior really to be generated in the way it appears in agentive experience to be generated, a further requirement must be met: the non-manifest satisfaction conditions which, if satisfied, make it the case that the manifest satisfaction conditions are fulfilled must provide a phenomenologically adequate theoretical account of the way that human behavior appears in agentive experience to be generated. B3 C does not provide a phenomenologically adequate theoretical account of what appears to be the case in agentive experience. Nida-Rümelin is acutely aware of the fact that B2 and B3 remain quite obscure as long as no explication is provided of what it is for a condition to be ‘a phenomenologically adequate theoretical account of what appears to the case’ in a given type of experience. At present, she has no such explication to offer.9 But the guiding idea behind B2 is this: fulfilling A6 requires that C must be nothing other than a more precise and theoretically informed reformulation of the way our own behavior appears to be generated in experiences of being active. Nida-Rümelin contends that condition C fails to meet this further requirement. Horgan agrees. Thus, if B2 is true then Horgan’s assumption B.1 is mistaken: Horgan’s claim D1 (together with the assumption that C is in general fulfilled when the relevant kind of agentive experience occurs) does not imply A6 even if he is right that C does not conflict logically with the phenomenally manifest satisfaction conditions of agentive experience. So our most basic disagreement turns out to be about conflicting respective background assumptions that we have been making, respectively, throughout: Horgan’s B1 vs. Nida-Rümelin’s B.2. Our efforts in trying to agree about how we disagree have brought this fact into the open. The debate will continue, and we invite others to join in.

Notes 1 Fully acknowledging the phenomenal character of agentive experience, in my view, requires acknowledging that such experience has distinctively agentive phenomenal aspects—over and above whatever other aspects (e.g., aspects as-of one’s body moving a particular way) it might have in any specific agentive experience. 2 This use of ‘intentional’ and ‘intentionality’ is fairly standard in analytic philosophy of mind, with the term ‘intentionality’ often being glossed as ‘aboutness’. It is an interesting issue just how this usage comports with the notion of intentionality inherited from the Phenomenological tradition descended from figures like Brentano and Husserl, in which intentionality is understood as “directedness to,” or awareness of, “intentional objects.” 3 Here and throughout, I employ the expression ‘state causation’ rather than the expression ‘event causation’. I am using the term ‘state’ in a broad sense that is intended to include not only persisting “non-changes” but also (nomically governed) events and processes. My principal reason for eschewing ‘event causation’ is that there is something phenomenologically episodic—something temporally located, and thus “event-ish”—about experiences of self-as-source. Thus, the expression ‘state causation’ works better than ‘event causation’ as a way of expressing how one’s behavior is not presented to oneself in agentive experience: one’s behavior is not presented as occurring within a causal nexus that is evolving, by itself, apart from any agentive intervention. 4 I do not presuppose the truth of that objectivist theory about the nature of color. In fact, I believe that objectivism about color must be rejected (see Nida-Rümelin 2018: section 7). 5 Examples for this may be found in the artistic work of Escher. 6 Perhaps I should underline that I am not proposing a claim, which Horgan already discussed at some length in his writings. I am not just claiming that it is part of the veridicality conditions of experiences of being active that the behavior at issue is not state-caused. PO1, PO2, and PO3 do not imply that different claim without further premises and these claims are, as far as I can see, immune against Horgan’s arguments against the former (See Horgan 2007a: section 2).

297

Terry Horgan and Martine Nida-Rümelin

298

Conditions of agentive phenomenology Nida-Rümelin, M. (2016) “Active Animals and Human Freedom,” in W. Freitag, H. Rott, H. Sturm and A. Zinke (eds.), Von Rang und Namen. Philosophical Essays in Honour of Wolfgang Spohn, Paderborn: Mentis Verlag, pp. 339–378. Nida-Rümelin, M. (2018a) “Freedom and the Phenomenology of Agency,” Erkenntnis 83/1, 61–87. Nida-Rümelin, M. (2018b) “Colors and Shapes,” in F. Dorsch and F. MacPherson (eds.), Phenomenal Presence, Oxford: Oxford University Press, pp. 77–101. Nida-Rümelin, M. and Suarez, J. (2009) “Reddish Green. A Challenge for Modal Claims about Phenomenal Structure,” Philosophy and Phenomenological Research 78/2, 346–391. Sellars, W. (1962) “Philosophy and the Scientific Image of Man,” in R. Colodny (ed.), Science, Perception, and Reality, New York: Humanities Press and Ridgeview, pp. 35–78.

299

20 AMBULO! Structures of phenomenology and ontology in action David Woodruff Smith

The complex structure of action The phenomenology of action was foreshadowed long ago by René Descartes in his reply to Pierre Gassendi’s objection that the cogito inference (cogito ergo sum: “I am thinking, therefore I exist”) should apply to “any of your other actions [actiones]”. Descartes wrote: I may not, for example, make the inference ‘I am walking, therefore I exist’ [ego ambulo, ergo sum], except in so far as the awareness [conscientia] of walking is a thought. The inference is certain only if applied to this awareness, and not to the movement of the body which sometimes—in the case of dreams—is not occurring at all, despite the fact that I seem to myself to be walking. (Descartes 1641/2013, Fifth Replies (to Gassendi): 137) Certainty, however, is not at stake here. Our concern is with the structure of a bodily action such as walking with awareness of walking. The distinction between action and awareness-of-action was precisely put by Descartes and would be amplified significantly as phenomenology took shape in our time (cf. Smith 2004b on distinguishing different aspects of the cogito). On the analysis to follow here, an action—a conscious, intentional, volitional, embodied action—is a highly structured whole. One part of my action of walking is a certain bodily movement (a point of ontology). Another part of my action is my experience of so moving and doing so by will (a point of phenomenology). Central to my experience of acting is my volition to so act, my willing to walk. Where successful, my volition effects my movement as I walk along. In a different way, though, my volition “intends” or, where successful, is intentionally related to my movement: in the Husserlian sense of intentionality, i.e., not purposefulness, but content-directedness. We shall assess the content of my volition in my experience of walking, treating the content as a complex form of meaning that Husserl called “noematic” content. We may gloss that content as something like , or , or a bit more fully (Angle quotes will be used to denote ideal contents of experience in a broadly Husserlian style.) 300

Ambulo!

Within my action as a whole, then, we find two distinct relations between my willing and my movement. My willing causes my actual movement in the world. And my volition—my “act” of consciousness in so willing—is intentionally related to my movement. In successful action, that is, my movement satisfies the intentional content of my “act” of volition, which aims at said movement with a certain awareness of effecting that movement. Thus, we distinguish my movement itself from the intentionality of my volition in acting, my volition being both “directed” toward and effecting that movement. The content informs my volition as “aiming” toward my movement, but it is neural impulses near the frontal region of my brain that causally effect my bodily movement. To conflate the intentional relation with the physical causal relation within the action is simply a category mistake. Yet my action itself includes both of these relationships as dependent parts, each requiring the other within the action as a whole. Crudely, the “act” of volition is the “action” shorn of the de facto bodily movement—that is, “bracketing” the question of the volition’s successfully effecting the movement. Alternatively, the “action” is the embodiment of the successful “act” of volition. Accordingly, we distinguish the phenomenological structure of my experience of walking from the ontological structure of the action. The phenomenological structure is itself surprisingly complex. My sense of my body as I experience my ambulatory movement involves my kinesthetic awareness of my movement in walking. And my experience of willingly walking with that kinesthetic sense is interactive with my visual and tactile and auditory perception of my environs as I walk along the road. Furthermore, my “intended” action in walking is part of my immediate surrounding world, or Umwelt, wherein my action itself both depends on and is intentionally guided by my perception of my surroundings as I walk—lest I walk into a parked car or a spiny cactus, or lest my dog on leash bark at our approaching neighbor. However, my kinesthetic, visual, tactile, and auditory forms of perception are not purely sensory “impressions”, but rather forms of perception both sensory and meaningful. As Husserl insisted: these perceptual acts involve a fusion of “sensory” and “noetic”, or meaningful, elements. Still, the point is not that I am attentively doing all these things at once: volitionally walking, while perceptually tracking my movements through kinesthetic, visual, tactile, and auditory elements of experience. Rather, as I walk along, my attention may move around among my bodily movements and parts of my environment as I perceive the pavement beneath me, the trees alongside the road, the pelicans I note skimming the waves in the ocean. To be precise, my volition in walking “intends” my action as a whole, within my surrounding world. My bodily movement per se is an essential part of my action, but what I will in walking is my whole action in its significant context. I may focus for a moment on how I move my foot as I step over a slippery rock, or I may be focused on an abstract idea even as I walk along. Yet, in any case, my volition in walking is a foundational part of my consciously walking—I am neither robot nor zombie. Now suppose I am seated by the fire as I “think” or day-dream that I am walking along my familiar road. My mental action in so thinking is itself dependent on maintaining my bodily posture, in my “sitting” meditation, and so on my volition to sit in a way amenable to my meditating. Even this mental action is thus grounded in a form of embodied action. The possibility of disembodied consciousness or “action” we shall not consider here, for this possibility is not a “motivated” or “real” possibility—as Husserlian phenomenology would find, with apologies to science fiction enthusiasts, not to mention Descartes and Gassendi.

301

David Woodruff Smith

Suffice it to say, a simple action like walking is a complex phenomenon, within which we distinguish phenomenological and ontological structures of the action. (The present analysis extends that in Smith 1992/2004, “Consciousness in Action”, reprinted in Smith 2004. Relevant aspects of intentional causation are addressed, in a different context, in Searle 1983, 2015.) Note, I shall use the term “phenomenology”—in the traditional sense following Husserl et al.—to mean the discipline of studying the character or structure of conscious experience (including action) from the first-person perspective. A cognate use of the term has recently taken root where an experience is said to have a “phenomenology”, that is, a phenomenal character. This character is commonly glossed as the character of “what it is like” for a subject to have such-and-such an experience. (This idiom has been canonized since Thomas Nagel’s classic 1974 essay “What Is It Like to be a Bat?”. Cf. Siewert 2013 for a nuanced treatment of phenomenal character.)

My lived body versus my physical body Within classical phenomenology Husserl focused often on perception as a paradigm of intentional experience. What is less appreciated is the rich characterization Husserl gave of the “lived” body, which led into Merleau-Ponty’s characterization of the “phenomenal field” of experience centered on the body as experienced. With a view to the phenomenology of body and embodied experience, in Husserl and Merleau-Ponty, we can develop a detailed analysis of the structure of conscious intentional action. Ultimately, the model will include an appraisal of the ideal “noematic” content of volition in action, which itself involves a concrete sense of the subject’s “lived” body in action. That is the aim of the present study. (Cf. Husserl 1913/2014 [Ideas I]: §§28, 116; Husserl 1989/1912 [Ideas II]; Merleau-Ponty 1945/2012. Cf. Smith 2013 for the reconstruction of Husserl’s theory of intentionality on which I rely.) Husserl did not lay out a detailed theory of action per se along the lines pursued below, nor did Merleau-Ponty. However, we may draw fruitfully on details in Husserl’s account of the lived body in Ideas II along with remarks in Ideas I about the structure of volition or will. Husserl frequently spoke of “willing” (Wollen) alongside seeing, thinking, feeling, and other forms of intentional experience, taking these as different types of “acts” of consciousness. And to “acts” of willing, as well as “acts” of seeing and thinking and desiring, he applied his mature theory of intentionality featuring “noetic” and “noematic” contents. Bear in mind that, in Husserl’s scheme, my “act” (Akt) of willing-to-walk is a proper and dependent part of my “action” (Handeln) of consciously walking along the road. A theme in recent philosophy of mind is the issue of “cognitive phenomenology”. What (if any) is the phenomenal character or “phenomenology” of various types of consciousness: in pure sensation, in perception as cognitively informed sensory experience, in thought including thinking more abstractly than about what one sees or hears? Further, what (if any) is the phenomenal character of “conative” (effortful) or “volitional” (willful) forms of experience, that is, in the lived experience of action? There we find pointed issues of “action theory”: in the relevant structures of phenomenology and ontology in action, beginning with simple everyday actions like walking or throwing a ball or digging in the garden. At stake is the phenomenal character in one’s sense of one’s lived body in action. So here I am, walking along my neighborhood road. I pause, gazing at the silvery gloss on the Pacific, while my dog sniffs her way along. I resume walking, in my normal gait. What could be more familiar?

302

Ambulo!

What’s new is my hip of Theseus. I now have a hip structure formed from titanium and polyurethane, replacing my original hip structure of bone and cartilage. In this respect we are naturally speaking of my physical body. Nonetheless, as I walk along, I experience no unusual sensations in my movement, no novel kinesthetic feedback, and no variations in my “willing” to walk along. I am walking: experiencing a familiar form of conscious, phenomenal, intentional, volitional, embodied action. And my hip replacement notwithstanding, it is still I walking: myself-in-action, moving volitionally, as this body, my lived body. Accordingly, my experience of walking is a case of everyday conscious volitional action, something I do, an activity incorporating both my embodied movement and my consciousness in and of and effecting that movement. We cannot make sense of the structure of action, however, without a careful distinction between my “physical body” and my “lived body”.

The phenomenology versus ontology of my action in walking The ontology of my body in action as I walk is one thing: my body now incorporates a structure of titanium. ’Tis still me, my walking, my body doing its normal biological thing, with a twist of titanium, the physiology and biology and physics flowing as they will. The phenomenology of my body in action as I walk is however quite another thing: my lived experience of walking along, habitually moving one leg and then the other, in a thoroughly familiar form of activity. My experience in walking is conscious, phenomenal, intentional, volitional, and experienced as embodied: in short, an experience of volitional bodily activity, or action. My experience of walking has a distinctive phenomenal intentional content: in brief, the content , or better , or (if you will) , executing my self-executive order to walk. The Latin phrase “ambulo” is suggestive of an executive form of phenomenal intentional content: . We’ll return to this suggestion, but first we need to home in on the rich phenomenology of action launched by Husserl in his Ideas II (and expanded by Merleau-Ponty in his Phenomenology of Perception). To forestall a potential misconception, let us be clear that my experience of walking is not a form of thinking, say, where I think “Here I am walking” in some form of higherorder monitoring that “represents” my action while my body does its perambulatory thing. The phenomenology of action is very different from that of consciously thinking “I am walking”—thus our title “”. Nor is my experience of walking a purely habitual phenomenon with no consciousness, say, like sleep-walking with no memory thereof. Hubert Dreyfus has long stressed the skillful practices involved in action, and rightly so, but a Dreyfus-style account of “skillful coping” in walking should be compatible with the account unfolding here (cf. Dreyfus 2014). To be sure, Husserl held that the form of an “act” of consciousness is that of “the cogito” (1913/2014: §28). However, the Cartesian term belies the rich detail in Husserl’s phenomenology of the “lived” body in Ideas II. Accordingly, we turn to phenomenological reflection on my experience of walking. Fundamentally, my body appears in my experience as me walking: “I am walking!”, consciously, intentionally, volitionally. Note the first-person form of my experience. By contrast, from a third-person scientific point of view, I may think about my walking by virtue of my new titanium hip, part of the structure of my physical body. Here we emphasize what should be an obvious distinction between the phenomenological structure of my experience of walking and the physiological structure of the worldly movement

303

David Woodruff Smith

that defines my walking per se. Titanium is now part of that physical structure, but the sense does not appear in the phenomenological content of my experience of walking. The case of my hip of Theseus serves to underscore the difference between my biological body per se and my body “as lived”. To bring out this contrast, we turn, in further detail now, to Husserl’s distinction between the “lived body”, or Leib, and the “physical body”, or Körper. Importantly, that distinction we should see as a specialization of Husserl’s distinction between the object of an experience and its mode of presentation in consciousness: that is, between the object itself and an appropriate sense of that object appearing in a particular act of consciousness. Thus, we distinguish between my actual physical body in motion—featuring neuro-muscular contractions supported by skeletal configurations—and my sense of my body as it appears in my conscious experience of walking. And that distinction we shall set within Husserl’s phenomenological theory of intentionality. Edmund Husserl launched the discipline of phenomenology a century ago. The foundations he laid in his Logical Investigations (1900–1901). His later, “transcendental” version of phenomenology he detailed in Ideas I (1913). There he used his famous method of epoché or “bracketing”: as I bracket the thesis of the existence of the world around me, I thereby focus on my consciousness of the world from the first-person perspective. Now, a common brief of Husserl’s conception of transcendental phenomenology sets consciousness out of connection with one’s surrounding world—in a supposedly “Cartesian” model of disembodied conscious experience. However, this common view of Husserlian phenomenology is strangely misbegotten. For in the summer of 1912 Husserl drafted Ideas II along with Ideas I, though he withheld Ideas II from publication. This authorial strategy led to endless misunderstandings. For in Ideas II Husserl practiced phenomenological analysis in close detail, while in Ideas I Husserl pitched his phenomenology in a style that emphasized high-altitude metaphilosophical themes largely in abstraction from the concrete analyses detailed in Ideas II. Yet the experience of embodiment is central to the forays of Ideas II. In short, consciousness is not experienced as disembodied—certainly not in our everyday life where an experience of walking is typical of conscious life, as I move around in my daily adventures, even as I practice phenomenological analysis here, typing away but going to the kitchen for a glass of wine. In Ideas II, amplifying the story of phenomenology detailed in Ideas I, Husserl drew a sharp distinction between—in the first person—my “lived body”, or Leib, and my “physical body”, or Körper. As glossed above, my physical body is the biological system including my brain and limbs and organs and bones in complex interaction involving neural processes and all sorts of physiological activity—governed by the physics of relativity and quantum phenomena well beyond the reach of my consciousness. By contrast, my “lived body” is my body only as I experience it, for example, in my quotidian experience of walking. Husserl called my lived body the “zero point” or center-of-orientation of my everyday experience in my surrounding world, my Umwelt, and indeed the “organ of my will”. Perception and action are thus intertwined, on Husserl’s analysis. Merleau-Ponty followed suit in saying “my body” is the center, the organizing principle, of my “phenomenal field”. At issue, for the present essay, is how we should characterize “embodiment”, for the ontology and the phenomenology need to be carefully parsed and carefully related. In the present analysis we respect a broadly Cartesian aspect of the experience of walking, yet we reject the metaphysical dualism of mind and body that is often held up as a sort of bogeyman, not least since Gilbert Ryle cast Descartes’ vision as that of “the ghost in the machine”. On the story told here, I am no “ghost” and my body is no “machine”. We need a careful understanding of where the ontology and phenomenology of “my body” interact. 304

Ambulo!

Interestingly, Husserl explicitly included volition or “willing” (wollen) under his broad sense of “cogito” (see Husserl 1913/2014: §28). And accordingly we may see in the “act” of willing the locus of interaction between the phenomenology of acting and the ontology of acting, namely, in the lived body. Thus, when I do something through my own volition, do something “bodily”, as in my action of walking, my embodied movement is part of the action. Within my very movement, however, we distinguish my “lived” body and my “physical” body. There is, in the world, one entity that is my body, but that entity has different aspects: properties given variously as “lived” and as “physical”. My volition in walking carries a content that prescribes my body by appeal to its “lived” properties, as in moving my legs by volition; but that content makes no appeal whatsoever to the physical composition of my hip, now formed partly from titanium. We characterize the content of my “act” of willing as the special type of content , or . The “hereby” means “by the force of my so willing”. This element of content indicates the causal force of my willing in action. However, the sense of causation in the content of willing does not “intend” a physiological process in my “physical” body, where a neural process in the frontal region of my brain initiates a transmission of signals ultimately stimulating muscle activities in my lower limbs. Rather, the “hereby” element in means a power I experience in my “lived” body, a causal process as experienced. The “!” articulates that sense of executing my bodily movement in walking: by fiat, as it were. Bear in mind that the phenomenological structure in my action takes its place within the wider ontological structure in my action within my surrounding world. My “act” of volition is a dependent part of my action of walking along the road (assuming I am not the victim of a Cartesian evil demon or, in the scifi variant, a brain-in-a-vat). And the content in my volition is satisfied only under a complex condition wherein my actual “physical” bodily movement is in accord with how my “lived” body is moving just as “intended” in my volition. In other words, normally there is a successful intentional relation between my “act” of volition and my bodily action of walking. And, furthermore, this relation occurs in a still more complex situation featuring my action in my surrounding world including road and trees and distant ocean. The intentional relation thus links my “act” of volition with my “action” of walking, thereby structuring my action within my Umwelt. In Husserl’s idiom, the volition is a “moment”, or dependent part, of the action as a whole. Where successful, the action being in accord with the volition, the “act” of will is thus a part of the action that could not occur apart from the action. So the volition is dependent on its effect in bodily movement. In the normal course of events, then, the willing is embodied in that it is intentionally related to the action of which it is a dependent part. (Husserl’s notion of Leib arises in various places in Husserl’s philosophical system: compare Smith 2013. Recent studies of the phenomenology of embodiment are the essays by Dermot Moran, Dorothée Legrand, and Komarine Romdenh-Romlunc, and Shaun Gallagher in Dahlstrom et al. 2015.)

From cognitive to conative to volitional phenomenology Classical phenomenology—in writings of Husserl, Merleau-Ponty, and others—seems to take it for granted that all conscious experiences are “phenomena” and so have a “phenomenal” or “phenomenological” character. I take it that a conscious experience normally and indeed constitutively involves a certain phenomenal character: the character of an experience as lived or experienced from the subjective first-person perspective. I see phenomenality itself as one factor among several in the structure of “awareness” characteristic of our typical 305

David Woodruff Smith

conscious experiences. Our concern here is how an action appears phenomenally in the experience of action. (My view on phenomenality is developed in Smith 1986, 2004, 2005, 2013, 2016a. Cf. Siewert 2005 on sensorimotor intentionality, Siewert 2013 on phenomenality, Dreyfus 2014 on skillful action, and Zahavi and Kriegel 2015 on phenomenality as for-me-ness.) A recent line of debate in philosophy of mind has focused on the question of which—if any or if all—forms of mental activity have a bona fide phenomenal character, or “phenomenology” (in the current vernacular): the character of “what it is like” to experience that type of mental activity. Some philosophers have held, in a neo-Humean vein, that only sensory experiences like seeing a red round patch have a phenomenal character. Other philosophers hold that cognitive experiences also have a phenomenal character, where “cognitive” experiences range from seeing a red, round, ripening tomato on the window sill (where the content outruns sensory feel) to thinking about Kant’s Categorical Imperative (where the content is purely conceptual). Well, now, what about actions? Do we find a phenomenal character in a typical conscious, intentional action such as walking down the road, or hitting a tennis ball cross-court, or simply picking up a fork at dinner? If “cognitive” experiences have a phenomenal character, what about embodied conscious actions? (The spread of recent views on phenomenality and cognitive phenomenology is articulated in Bayne and Montague 2011. For my take on cognitive phenomenology, see Smith 2011, 2016a.) “Conative” or “agentive” experiences, typically enacted in intentional bodily actions, include trying to do something, in using one’s body, guided by beliefs and desires and other background mental activities. What is the phenomenal character, if any, of conative or agentive experiences, viz., in experiencing an action? (See Horgan 2011 on “agentive” phenomenology, and Siewert 2005 on sensorimotor experience.) The term “conation” or “conatus” literally means trying (from the Latin “conatus”), as in trying to do something (out of desire, belief, will). So the notion of “conative” experience connotes that aspect of action, of doing something, which in effect factors away the success of the “doing”. If we apply Husserl’s technique of epoché to an action, we “bracket” the actual bodily movement and thereby focus on the “pure” experience of acting: if you will, a “conative” act of consciousness, which consists in trying to do something. Uriah Kriegel has outlined a phenomenological analysis that opens the door nicely to conative experiences: in his 2015 book The Varieties of Consciousness (ch. 2 on conative phenomenology; cf. Horgan 2011), Kriegel begins by grouping various types of “conative phenomena” including desiring to, wanting to, intending to, deciding to, being willing to, etc. All are aspects of the experience of effort, or “conatus”, as in everyday intentional actions. And all are offered specifically as having a distinctive phenomenal character, or “phenomenology”. Kriegel then develops a particular account of the structure of phenomenal conative intentionality (ibid.: 83–96). First, he argues for a bona fide “primitive conatus”, a form of effortful experience with its own phenomenal character, a distinctive character not reducible to that of some other type of experience, such as a kinesthetic awareness of one’s arm moving, as opposed to moving one’s own arm conatively. Kriegel then proposes an analysis of the structure of this form of experience: a conative experience, he proposes, is a complex form that combines decision and effort. So a conative experience, for Kriegel, is a complex comprising two proper parts: a phenomenal deciding to do such-and-such, and a phenomenal effort to do such. I would add, as a friendly refinement, that these two components in the conative experience of an action are normally fused, i.e., interdependent. Then there is but one form of 306

Ambulo!

phenomenal character, whereby the phenomenal experience of conatus is a “decisive-effort”. And if the intentional content of this conatus is satisfied by the prescribed movement, then I take it the action consists in a structure comprising the conatus joined with the resulting successful movement, the embodied action. This “conatus”, or “conation”, model strikes me as on the right track. However, I find the traditional term “volition” resonates more fully, and arguably for an interesting reason. In the theory of conscious intentional action, I would like to draw in a broadly Husserlian theory of ideal content, in a certain rendition of what Husserl called the “noema” of an act of consciousness. Specifically, the core content of an action, I want to say, is the ideal form of willing to act in a certain way, and that form of content deserves its own logical force: the force of “Do this!”—an executive order, as it were, rather than simply a feeling of effort. The aspect of “deciding” what to do (per Kriegel) I see as resolving instead into the “sense” (Sinn) of which form of action is willed. For example, as “I will to walk”, we shall parse the structure of this volition into the character of willing and the characterization of what is willed. In Husserl’s terms, the noema of a volition parses into the “thetic” act-character of willing and the noematic “sense” (Sinn) of what is willed, here, my lived bodily movement in walking along the road. My volition in walking is thus informed by the content , which in the context of my action invokes a “horizon” of desire, belief, value, etc. The central player in the action, however, is the constituent volition bearing a properly volitional content. For the record, Kriegel sees his account of “primitive conatus” as a development of Paul Ricœur’s phenomenological theory of the will. And Kriegel aptly notes the rich contributions of French phenomenologists to the characterization of various forms of conative experience, not least the existential model of choice in Jean-Paul Sartre and Simon de Beauvoir, among their contemporaries. Broadly speaking, existential analyses carried phenomenology from concrete embodied action into the wider social realm, where meaning is formed through intersubjective “constitution” of values, of self, or society. (Consider Crowell 2001, 2013, drawing on Husserl and Heidegger; Fricke and Føllesdal 2012 on intersubjectivity and values.)

The “fiat!” character of volition in action Husserl often mentions willing (wollen) as an “act” of consciousness. Yet we turn to William James for a specific model of the form of consciousness in willing in action. In his Principles of Psychology (1891/1983: 1098), William James addressed the will in terms appropriate to our discussion: Desire, wish, will, are states of mind which everyone knows . . . If with the desire there goes a sense that attainment is not possible, we simply wish; but if we believe that the end is in our power, we will that the desired feeling, having, or doing shall be real; and real it presently becomes . . . The only ends which follow immediately upon our wiling seem to be movements of our own bodies. The proto-phenomenologist James soon declared: . . . whether or no there be anything else in the mind at the moment when we consciously will a certain act [i.e. action], a mental conception made up of memory-images of these sensations, defining which special act it is, must be there. 307

David Woodruff Smith

. . . there need be nothing else, and . . . in perfectly simple voluntary acts there is nothing else, in the mind but the kinaesthetic idea, thus defined, of what the act is to be. ( James 1891/1983: 1104) Thus, for James, willing aims specifically at “the act that is to be”, that is, the intended or willed action as defined by that “kinaesthetic idea”. In Husserlian terms, I will my bodily action just as defined by the noematic sense in my volitional act of consciousness, that is, I will my “lived” bodily action, which I experience kinesthetically. With a fine nose for the phenomenology of volition, James notes: “One has only to play ten-pins or billiards, or throw a ball, to catch his will in the act . . ., when it says ‘Now go!’ ” (1891/1983: 1105). In our running example of walking: my will is exercised by fiat as I will, “Now I walk along this road!”—nicely put as, “Now go!” “Effort of attention”, James holds, “is . . . the essential phenomenon of will” (1891/1983: 1167). That is, where will is applied to the type of action to be executed, James holds that attention is essential to holding in mind that action type. Husserl would translate the point into “intention”: framing the type of action willed through an appropriate meaning—such as “walking”, “throwing a ball”, or “shooting billiards”. Broadly, then, James affirms a basic structure of action whereby a mental “act” of willing is applied to an “attended” type of bodily “act” defined kinesthetically. Looking to the phenomenological structure of a voluntary action, we draw from the Jamesian account of will a distinction between: (a) my willing per se, in the form , and (b) my sense of the action willed, e.g., , where the form of action is defined kinesthetically— as a “lived” bodily movement of walking. So the executive fiat “Now go!” applies to the kinesthetic form of movement “I walk”: the form of volitional action applying the fiat to the form of movement, thus < Go! (I walk) >. These two aspects of volition form a Jamesian account of the structure of action. But we need to draw in details of Husserl’s mature theory of intentionality, featuring the structure of “noematic” content in an intentional action.

The noematic structure in volition: “Ambulo!”, “I walk!” Husserl distinguished the “real” or temporally flowing content of an “act” of consciousness from the “ideal” meaning content entertained in the act. These two types of content Husserl called “noesis” and “noema” respectively. Interpolating among details in Husserl’s theory of intentionality in Ideas I, I propose, we may mark out a distinctly Husserlian theory of the formal structure of volition: the structure of the noematic content or noema of volition in an action such as walking. On Husserl’s analysis, the noema of an intentional act of consciousness has two formal parts: (i) a sense (Sinn) that prescribes the object toward which the act is directed, carrying the phenomenological force of “the object as intended”; and (ii) a thetic or positing (setzen) component that prescribes the type of act executed, carrying the phenomenological force of how the object as intended is “posited”, say, in seeing, judging, imagining, etc. Thus, when I think that Descartes was born in La Haye, the sense of my experience is the propositional sense , and the thetic character of my experience is the modifying or “modalizing” content , or rather , or . Similarly, when I see a black dog sauntering along the road, the sense of my experience is the content , and the thetic character of my experience is the modifying visual content . The full noema of an experience 308

Ambulo!

is thus a structured meaning content combining Sinn and thetic content, for example, in these two cases: < I think that Descartes was born in La Haye >, < I see that black dog sauntering along the road >. (Cf. Husserl 1913/2014: §§28, 88–94 ff., and §§128–133, on the structure of an act’s noema featuring Sinn plus thetic character, and §95 on valuing and willing. For details on Husserl’s model of the noema, see Smith 2013: ch. 6, especially 267ff., and ch. 9. Where Husserl used German-style quotation marks to denote noematic content, we are here using angle brackets for content “quotation”.) Husserl introduced the theory of noema, famously, with an example of seeing a tree, characterizing the Sinn in the noema of the visual experience as “the perceived tree as perceived”. What would the noema of an experience of action look like—viz., in an “act” of willing in action? Intriguingly, Husserl explicitly addressed the form of the noema of an act of willing: in Ideas I, §95 on valuing and willing. In an act of valuing (Werten), Husserl holds, the noesis and the corresponding noema has a structure of “intentional layering” (Schichtungen). Within the act’s noesis, there is a dependent layer (Moment) of “valuing” (Werten) that is “built-upon” (“aufgeschichtet”) a layer of presentation that is not dependent on the valuing layer. Accordingly, within the Sinn component of the noema, there is a higher layer (Schicht) of meaning, presenting an intended value, that is founded on a lower layer of meaning, presenting an object or state of affairs which ostensibly carries the intended value. Thus, the higher level of meaning (presenting a value) is dependent on the lower level of meaning (presenting an object), but the lower is not dependent on the higher. The noematic Sinn thus forms a “foundational whole” (Fundierungsganzes). Similarly, Husserl holds, in an act of willing (Wollen), the noesis and corresponding noema has a layered structure. Thus, within the act’s noema: a higher layer of Sinn, presenting the action-as-willed, is founded on a lower layer of Sinn, presenting the action (Handeln) itself (the effected action “simpliciter”). The higher level of Sinn within the act’s noema—in willing as in valuing—is thus a layer “built-upon” (“aufgeschichtet”) the lower level of Sinn. For Husserl, the “willed as such” (das Gewolltes als solches) is thus founded upon the “presented as such” (the “action as presented”), the former meaning dependent on the latter. In his discussion in §95 Husserl pointedly uses quotation marks to denote the Sinn (or Meinung) within the noema of the act of willing. We should think of the noema in an act of willing as the ideal meaning content abstracted from the concrete act of willing. And accordingly Husserl explicitly analyzes the “logical” structure of the noematic content of willing. (My translations here from the German. Cf. Peucker 2012 for a discussion of Husserl’s evolving views of volitional consciousness.) Let us see how Husserl’s model would work for everyday cases. Suppose I see a “good dog” sitting patiently by the door of a café. Within the noematic sense , the noematic meaning is founded on the meaning . Husserl uses quotation marks, as in characterizing my seeing “this good dog . . .”, to signify noematic meanings, so what is at stake is the structure of the noema, rather than the actual properties “meant” or “intended” of the animal. In this case, then, the meaning is founded on the meaning , and this “layering” structures the noematic Sinn in my value-laden perception of the sitting dog. The act’s full noema we then articulate as: < I see this good dog sitting by the café >, 309

David Woodruff Smith

where the sense “is supervenient on” the sense . Interestingly, Husserl’s term “aufgeschichtet” is translated as “supervenes on” in the 2014 translation of the passage in §95—a term resonant with recent interests in supervenience as a form of dependence. Now suppose that I am walking along the road. A key part of my action of walking is my willing to so walk: I will that “I hereby walk!” On a Husserlian analysis, then, the noema of my act of willing includes the Sinn . We add the “!” to articulate the executive “fiat” aspect of my perambulatory movement as willed. Within the Sinn in the act’s noema, then, we find a “layered” structure where the executive force , expressed in , is founded on the sense of my movement . And so the full noema of my act of willing we may specify as: < I will that I hereby walk! >. The logic of this form of noema is remarkable. For Husserl, the act’s thetic character is manifest within the act’s full noema in two ways: first, in the thetic content ; and second, in the Sinn content , where “hereby!” means “so willed by fiat” in the intended action. Husserl’s claim, by interpolation, is that the Sinn is “layered” so that the higher level is founded on the lower level . Then is treated like . Moreover, Husserl pointedly treats willing as itself founded on valuing and “deciding” what action is willed (cf. §95). We should not take the phenomenology of willing to hold that a temporal sequence of independent “acts” ensues within action: first a presentation of a putative action, then a deciding-valuing of that action, then a willing of that valued action. The whole point of foundation (Fundierung) is that some parts of the action are ontologically dependent on others, and founding relations can link various parts without a temporal ordering or even a “layering” order. We should see the action as a whole with various parts, some dependent on others. We thus experience an act of willing, say, in the case of my walking, as a unified whole whose noematic Sinn is structured so that the fiat content is founded on the presentational content . The intentional force of the Sinn, whence of the noetic elements in my willing, is dependent on the way these meaning elements interact. The center of action in willing, per Husserl, is thus this structure in its volitional meaning content in the act of willing. However, we should not treat the character of being-willed as solely a feature of the Sinn in the act’s noema, as in . For the Sinn is itself modified by the thetic content, as in . Accordingly, a stronger analysis treats willing, like thinking or seeing, as a form of “intentional modality”—following what I have called the “modal model” of (self-) consciousness. Let us see how that approach treats the experience of action.

The modal model of volition in action A particular account of phenomenological structure follows the “modal model” of consciousness. Where Husserl spoke of “modalities” of belief (such as expected probability), the modal model articulates elements in the form following a conception of intentional modalities including perception, thought, and now volition. Thetic or positing character begins with an intentional modality such as perceiving or judging or thinking or, here, willing. And this “modal” character would modify the Sinn content in an intentional act of consciousness. 310

Ambulo!

(See Ideas I: §§129 and 132–133 on the thetic component in an act’s noema, and §§103–105 on “modalities” of belief including a sense of probability, falling under thetic or positing character. See Hintikka 1962, 1969 on intentional modalities, and see Smith and McIntyre 1982 on applying the notion of intentional modalities to Husserl’s conception of the “horizon” of possibilities “predelineated” by the content of an experience. On the modal model of (self-) consciousness, see Smith 1986, 2004a, “Return to Consciousness”, and Smith 2013, 2016a.) Following the modal model, we may articulate the structure of the noematic content of my act of volition in walking as follows: < Phenomenally in this very experience I now here will that I hereby now here walk down this road! >. Specifically, the form of my lived bodily action as intended is articulated by the Sinn content: < . . . I hereby now here walk down this road! >. And the form of my consciously executing my action through willing is articulated by the “modal” content: < Phenomenally in this very experience I now here will that . . . >. This complex thetic content distinguishes several factors in the way I enact the action by willing: phenomenally, with awareness, by the subject “I”, from a viewpoint “here now”. Notice how the Sinn content is tied back—reflexively or anaphorically—into the modal thetic content. First, the content in (Sinn) is bound back into the content in (modal content). Second, the content in < I hereby walk! > is bound back into the content in the modal content. Thus, in the action of walking, the object of my volition is not simply my physical body nor even my lived body moving in a walking style. Rather, the object of my volition is my action in walking. What I will is the action as a whole within which my willing is an essential, dependent, and executive part.

Toward a formal phenomenological ontology of action The types of structure we have charted in action—featuring parts, wholes, and dependencies in an action and its phenomenological content—fall under formal ontology. In opening Ideas I, Husserl outlined an ontology of formal and material categories. Material categories include Nature, Consciousness, and Geist (social “spirit”). These categories are constrained by formal categories including Individual, Property, Relation, Part/Whole, Dependence, Number, Set, Manifold, and so on. Meaning seems to be a “logical” category that either falls under formal categories of object (such as those just listed), or merits its own place in formal ontology (including distinctive meaning categories articulated in mathematical logic and model theory). (Husserl’s ontology is reconstructed in Smith 2013.) Over the long course of Ideas I, Husserl developed in effect a feedback loop as his analysis of the phenomenology of perception, judgment, and action both draws upon his formal ontology and grounds or justifies the ontology. The point should not be that the ontology is reduced to phenomenology, where being is reduced to appearance. Rather, the ontology 311

David Woodruff Smith

is interdependent with the phenomenology. Indeed, Husserlian epoché itself follows a “zigzag” back-and-forth between the world and consciousness thereof (cf. Smith 2016b). (I see a kindred meta-ontology in the neo-Carnapian approach to ontology defended by Amie L. Thomasson in Thomasson 2015.) In the spirit of formal ontology, action provides an especially forcing and revealing case study in the interplay between phenomenology and ontology. Indeed, action is the crux of our relation to the world!

Related topics See Chapters 2 (on Pfänder and Husserl), 21 (Hanna), 24 (De Monticelli), and 25 (Drummond).

References Bayne, T. and Montague, M. (eds.) (2011) Cognitive Phenomenology, Oxford and New York: Oxford University Press. Crowell, St. G. (2001) Husserl, Heidegger, and the Space of Meaning: Paths toward Transcendental Phenomenology, Evanston, IL: Northwestern University Press. Crowell, St. G. (2013) Normativity and Phenomenology in Husserl and Heidegger, Cambridge, MA and New York: Cambridge University Press. Dahlstrom, D.O., Elpidorou, A. and Hopp, W. (eds.) (2015) Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, New York and London: Routledge. Descartes, R. (1641/2013) Meditations on First Philosophy. With Selections from the Objections and Replies. A Latin-English Edition. Edited and translated with textual and philosophical introductions by John Cottingham, Cambridge, MA: Cambridge University Press, 2013. (Latin original, including objects and replies, 1641). Dreyfus, H.L. (2014) Skillful Coping: Essays on the Phenomenology of Everyday Perception and Action, ed. by M.A. Wrathall, Oxford and New York: Oxford University Press. Fricke, C. and Føllesdal, D. (eds.) (2012) Intersubjectivity and Objectivity in Adam Smith and Edmund Husserl, Frankfurt and Paris: Ontos. Hintikka, J. (1962) Knowledge and Belief, Ithaca, NY: Cornell University Press. Hintikka, J. (1969) Models for Modalities, Dordrecht and Boston, MA: D. Reidel Publishing Company. Horgan, T. (2011) “From Agentive Phenomenology to Cognitive Phenomenology,” in T. Bayne and M. Montague (eds.) (2011), Cognitive Phenomenology. Husserl, E. (1913/2014) Ideas for a Pure Phenomenology and Phenomenological Philosophy, trans. by D.O. Dahlstrom, Indianapolis, IN and Cambridge, MA: Hackett Publishing Company. (=Ideas I) Husserl, E. (1912/1989) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Second Book: Studies in the Phenomenology of Constitution, trans. by Richard Rojcewicz and André Schuwer, Dordrecht and Boston: Kluwer Academic Publishers. (Original manuscript dating from 1912, posthumously published in German in 1952.) James, W. (1891/1983) The Principles of Psychology. Introduction by George A. Miller, Cambridge, MA: Harvard University Press. Kriegel, U. (ed.) (2013) Phenomenal Intentionality, Oxford and New York: Oxford University Press. Kriegel, U. (2015) The Varieties of Consciousness, Oxford and New York: Oxford University Press. Merleau-Ponty, M. (1945/2012) Phenomenology of Perception, trans. by Donald A. Landes, London and New York: Routledge. Peucker, H. (2012) “Husserl’s Approaches to Volitional Consciousness,” in C. Fricke and D. Føllesdal (eds.) (2012), Intersubjectivity and Objectivity in Adam Smith and Edmund Husserl, pp. 45–60. Searle, J.R. (1983) Intentionality, Cambridge and New York: Cambridge University Press. Searle, J.R. (2015) Seeing Things as They Are: A Theory of Perception, Oxford and New York: Oxford University Press. Siewert, Ch. (2005) “Attention and Sensorimotor Intentionality,” in D.W. Smith and A.L. Thomasson (eds.) (2005), Phenomenology and Philosophy of Mind, pp. 270–294.

312

Ambulo! Siewert, Ch. (2013) “Phenomenality and Self-Consciousness,” in U. Kriegel (ed.) (2013), Phenomenal Intentionality, pp. 235–259. Smith, D.W. (1986) “The Structure of (Self-) Consciousness,” Topoi V, 149–156. Smith, D.W. (1992/2004) “Consciousness in Action,” Synthese 90, 119–143. Reprinted in D.W. Smith (2004), Mind World: Essays in Phenomenology and Ontology, pp. 122–146. Smith, D.W. (2004) Mind World: Essays in Phenomenology and Ontology, Cambridge and New York: Cambridge University Press. (See in particular ch. 2, “The Cogito circa A.D. 2000”, ch. 3, “Return to Consciousness”, and ch. 4, “Consciousness in Action”, reprinted from Synthese 90, 119–143). Smith, D.W. (2004a) “Return to Consciousness,” in D.W. Smith (2004), Mind World: Essays in Phenomenology and Ontology, pp. 76–121. Smith, D.W. (2004b) “The Cogito Circa AD 2000,” in D.W. Smith (2004), Mind World: Essays in Phenomenology and Ontology, pp. 42–75. Smith, D.W. (2005) “Consciousness with Reflexive Content,” in D.W. Smith and A.L. Thomasson (eds.) (2005), Phenomenology and Philosophy of Mind, pp. 93–114. Smith, D.W. (2011) “The Phenomenology of Consciously Thinking,” in T. Bayne and M. Montague (eds.) (2011), Cognitive Phenomenology, pp. 345–372. Smith, D.W. (2013) Husserl, second edition, London and New York: Routledge. Smith, D.W. (2016a) “Cognitive Phenomenology,” in D.O. Dahlstrom, A. Elpidorou and W. Hopp (eds.) (2015), Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, pp. 15–35. Smith, D.W. (2016b) “Truth and Epoché: The Semantic Conception of Truth in Phenomenology,” in P. Livingston, J. Bell and A. Cutrofello (eds.), Beyond the Divide: Philosophy in the Twenty-First Century, London and New York: Routledge, pp. 111–128. Smith, D.W. and McIntyre, R. (1982) Husserl and Intentionality, Dordrecht and Boston: D. Reidel Publishing Company. Smith, D.W. and Thomasson, A.L. (eds.) (2005) Phenomenology and Philosophy of Mind, Oxford and New York: Oxford University Press. Thomasson, A.L. (2015) Ontology Made Easy, Oxford and New York: Oxford University Press. Zahavi, D. and Kriegel, U. (2015) “For-Me-Ness: What It Is and What It Is Not,” in D. Dahlstrom, A. Elpidorou and W. Hopp (eds.) (2015), Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, pp. 36–53.

313

21 WILL-POWER Essentially embodied agentive phenomenology, by way of O’Shaughnessy Robert Hanna

In the beginning was the Act. (Goethe 1990: 153)

Introduction Goethe got it right. While I was a graduate student at Yale in the mid-to-late 1980s, during that heady, wonderful period between finishing coursework and starting actual PhDdissertation-writing labors, I read three big, brilliant books that significantly influenced my later work on cognition, mental content, knowledge, logic, rationality, the mind-body relation, mental causation, intentional action, free will, practical agency, and persons: Gareth Evans’s Varieties of Reference (Evans 1982), Brian O’Shaughnessy’s The Will (O’Shaughnessy 1980), and Richard Wollheim’s The Thread of Life (Wollheim 1984). At the same time, I was also steeped in Kant’s Critical philosophy; in existential phenomenology (especially the later Husserl, Scheler, Heidegger, early Sartre, and Merleau-Ponty); in Wittgenstein’s Investigations (Wittgenstein 1953); in Harry Frankfurt’s essays, many of them collected in The Importance of What We Care About (Frankfurt 1988); and in Richard Rorty’s radical metaphilosophy in Philosophy and the Mirror of Nature (Rorty 1979) and Consequences of Pragmatism (Rorty 1982). Those were the days of philosophical wine and roses. A Recipe for Long-Term Philosophical Agency about Rational Human Minded Animal Agency: Throw all the above-listed philosophical ingredients, including the wine and roses, into a big pot, add water till it reaches the brim, and heat slowly to an intellectual boil, stirring occasionally, for twenty years. And then, as the novelist John Cheever remarked in a closely-related context, “you apply butt to chair.” In the beginning, and also at the end, is the Act.

314

Will-power

Following that recipe step by step, I eventually wrote Rationality and Logic (Hanna 2006), Embodied Minds in Action (co-authored with Michelle Maiese, Hanna and Maiese 2009), Cognition, Content and the A Priori (Hanna 2015), and Deep Freedom and Real Persons (Hanna 2018a). The present chapter/essay is a rational reconstruction of some of the results of all that long-term philosophical agency about rational human minded animal agency, concentrating especially on three issues. First, I briefly describe the causally efficacious power of the minded animal will (or what Kant calls Willkür, aka “the power of choice”) in basic intentional actions, that is, intentional body movements, against the metaphysical backdrops of my essential embodiment theory of the mind-body relation and my metaphysics of free agency, Natural Libertarianism. Second, I describe some relevant comparisons and contrasts between my views about essentially embodied intentional action and O’Shaughnessy’s dual aspect theory of intentional action, about which he very provocatively says: Schopenhauer said: “I cannot imagine this will without this body.” If this wonderful remark is right – and . . . it is – then the psychological phenomena that occur when one engages in intentional physical action . . . depend in general upon the actual existence of the body. . . . Then it follows that [my position] cannot be an interiorist position, even though I believe that trying is a psychological event that causally explains the purely physical willed bodily event in the case of all bodily actions. For what here has come to light is, the essential mutual interdependence of bodily strivings and of the animal body upon each other. (O’Shaughnessy 2008: vol. 2, 600) And third and finally, I argue for the existence of a self-evident, veridical phenomenology of essentially embodied free agency, aka rational human minded animal agency.

Essentially Embodied Trying and Natural Libertarianism Here is a thesis about the nature of basic intentional action: A is a basic intentional act of an essentially embodied rational human minded animal, aka a real human person P, if and only if A is an intentional body movement of P that is structurally caused and actively guided and controlled by P’s simultaneously trying to perform A, which in turn is a physically irreducible conscious effective first-order desire to perform A, which in turn is P’s will. For convenience, I will call this thesis Essentially Embodied Trying. One crucial thing to notice about the Essentially Embodied Trying thesis is that what I am calling “intentional bodily movements” are essentially embodied events in a minded animal’s life, and not anything exclusively or fundamentally physical. This in turn directly implies what, in Embodied Minds in Action (Hanna and Maiese 2009), Michelle Maiese and I call The Essential Embodiment Theory of the mind-body relation and mental causation. Here are the six central theses of that theory: 1

The Essential Embodiment Thesis: Creatures with conscious, intentional minds are necessarily and completely neurobiologically embodied.

315

Robert Hanna

The Essentially Embodied Agency Thesis: Basic acts (e.g., raising one’s arm) are intentional body movements caused by an essentially embodied mind’s synchronous trying to make those very movements and its active guidance of them. The Emotive Causation Thesis: Trying and its active guidance, as the cause of basic intentional actions, is primarily a pre-reflective, desire-based emotive mental activity and only derivatively a self-conscious or self-reflective, deliberative intellectual mental activity. The Mind-Body Animalism Thesis: The fundamental mental properties of conscious, intentional minds are (a) non-logically or strongly metaphysically (namely, synthetically) a priori necessarily reciprocally intrinsically connected to corresponding fundamental physical properties in a living animal’s body (mental-physical property fusion), and (b) irreducible truly global or inherently dominating intrinsic structures of motile, suitably neurobiologically complex, egocentrically centered and spatially oriented, thermodynamically irreversible living organisms (neo-Aristotelian hylomorphism). The Dynamic Emergence Thesis: The natural world itself is neither fundamentally physical nor fundamentally mental, but is instead essentially a causal-dynamic totality of forces, processes, and patterned movements and changes in real space and real time, all of which exemplify fundamental physical properties (e.g., molecular, atomic, and quantum properties). Some but not all of those physical events also exemplify irreducible biological properties (e.g., being a living organism), and some but not all of those biological events also exemplify irreducible fundamental mental properties (e.g., consciousness or intentionality). And both biological properties and fundamental mental properties are dynamically emergent properties of those events. The Intentional Causation Thesis: A mental cause is an event or process involving both consciousness and intentionality, such that it is a necessary proper part of a nomologically jointly sufficient essentially mental-and-physical cause of intentional body movements. In so doing, it is a dynamically emergent structuring cause of those movements. Then, under the appropriate endogenous and exogenous conditions, by virtue of synchronous trying and its active guidance, conscious, intentional essentially embodied minds are mental causes of basic acts from their inception in neurobiological processes to their completion in overt intentional body movements.

2

3

4

5

6

Notice especially that according to this conception, a structuring cause i ii

is not a prior event-cause, because it is simultaneous and synchronous with the effect, as a complex dynamic pattern spread out over time, is equally formal and final in the Aristotelian senses and not (merely) efficient or material in the Aristotelian senses, in that iia it actualizes potential energy, and it is inherently goal-directed or at the very least forward-directed in time, iib it thereby activates, controls, and guides causally efficacious energy flows, and iic it is itself genuinely causally efficacious (as opposed to merely causally relevant) in an independently necessary, independently insufficient, but also jointly sufficient way, via those causally efficacious energy flows, but iid it is not reductively identical with – in the “downward-identity” sense of being “nothing but,” or “nothing over and above” – causally efficacious energy flows themselves, and

316

Will-power

i ii

is a non-reducible proper part of a whole causally efficacious complex dynamic organismic process, that is, a causally efficacious non-reversible, non-equilibrium self-organizing organismic energy flow.

In this way, The Essential Embodiment Theory says that our dynamically emergent, irreducible, sentient, and sapient minds are also necessarily interdependent with our own living organismic animal bodies and not essentially distinct from them; that we are farfrom- equilibrium, asymmetric, complex, self-organizing thermodynamic systems; that we act by intentionally moving our bodies by means of our desire-based emotions and trying; and that our conscious, intentional, caring, and rational necessarily and completely neurobiologically embodied minds are basically causally efficacious precisely because they are metaphysically continuous with our biological lives, and life is basically causally efficacious in physical nature. The simple upshots of The Essential Embodiment Theory, then, are: i

ii

in thinking about the mind-body problem we should decisively replace the early modern Cartesian and Newtonian ghost-in-the-machine metaphysics with a postCartesian and post-Newtonian but also at the same time neo-Aristotelian hylomorphic immanent-structure-in-the-non-equilibrium-thermodynamics metaphysics, and the irreducible conscious, intentional, caring minds of cognizers and agents grow naturally in suitably complex living organisms, as irreducible, non-dualistic, nonsupervenient, asymmetric thermodynamic structures of those organisms. Correspondingly, the three crucial points in the Essentially Embodied Trying thesis are:

that the trying is itself essentially embodied, that the trying is simultaneous and synchronous with the action, and not a prior event-cause of it, and iii that the causation characteristic of trying is inherently structural-dynamic and goal-directed, that is, it is genuinely efficacious neo-Aristotelian hylomorphic structural causation, fusing together “formal” causation and “final” causation in the classical Aristotelian senses. i ii

To put the Essentially Embodied Trying thesis in a name-dropping context, however, it is a constructive extension of Harry Frankfurt’s devastating critique of classical causal theories of action, his hierarchical desire conception of the will and personhood, and his guidance-control conception of intentional action (Frankfurt 1988a, 1988b, 1988c, 1988d, 1988e, 1988f ), together with O’Shaughnessy’s action-theoretic notion of trying (O’Shaughnessy 1973), when framed against the dual backdrop of my theory of logically driven cognitive rationality in Rationality and Logic (Hanna 2006), and The Essential Embodiment Theory of the mind-body relation and mental causation in Embodied Minds in Action. But in order to make all of this more clear and distinct, I will also have to provide a brief description of a background metaphysical theory I defend: Natural Libertarianism. What is free will? What is practical agency? What is human personhood? And how are human free will, practical agency, and human personhood really possible in the natural world as it is correctly characterized by the modern natural sciences, especially physics, chemistry, biology, and cognitive neuroscience? Or more compactly put: given the truth of modern science, how is human free agency really possible? Let us call this the freedom question.

317

Robert Hanna

In Deep Freedom and Real Persons (Hanna 2018a), I provide what I think is a rationally decisive and true answer to the freedom question. But in order to formulate this answer properly, I will need some precisely defined terminology. I will say that free will is a conscious subject’s power to choose or do what she wants to, or to refrain from so choosing or so doing, without preventative constraints and without internal or external compulsion, with at least causal responsibility; that practical agency is a conscious subject’s power to choose or do things freely in the light of principles or reasons, including moral principles or reasons, on the basis of self-conscious processes of deliberation and decision; and that a real person is an animal that is capable of free will, practical agency, and (moral or non-moral) responsibility. So the freedom question is asking how human free will, human practical agency, and real human personhood really exist, and how we truly have them, in the natural world as correctly described by modern science. Or, on the contrary, given the truth of modern science, are we really nothing but “biochemical puppets” (Harris 2012) or “moist robots” (Daniel Dennett, as quoted in Schuessler 2013), that is, nothing but natural automata, or natural machines, whose evolutionary and neurobiological mechanisms continually generate the cognitive illusion that we are free agents? If so, then we would be in an even worse cognitive place than Pinocchio, a wooden puppet who longed to be a real boy. We would be nothing but “meat puppets” (Meat Puppets 1994) dreaming that we are real human persons. The issue being raised here, then, is how we should understand the implications of the modern natural sciences for our classical conception of ourselves as rational and moral animals, in the face of the possibility that we are under a serious cognitive illusion about this. Indeed, some contemporary philosophers even think that once we are liberated from this serious cognitive illusion, we will finally clearly see that we are nothing but highly complex “biochemical puppets” and that “physics makes us free” in a deterministic, block universe (Ismael 2016). By sharp contrast to these philosophers, like Kant, I hold that we directly experience practical agency, or what he calls “practical freedom,” and also that practical agency requires metaphysically robust human free will or deep freedom – what he calls “transcendental freedom” – which really exists in the natural world: Practical freedom can be proved through experience. . . . We thus cognize practical freedom through experience as one of the natural causes, namely a causality of reason in the determination of the will, whereas transcendental freedom requires an independence of this reason itself (with regard to its causality in initiating a series of appearances) from all determining causes of the world, and to this extent seems to be contrary to the law of nature, thus to all possible experience, and so remains a problem. (Kant 1997: A 802–803/B 830–831, 675–676) I also fully agree with Kant that this kind of freedom “seems to be contrary to the law of nature, thus to all possible experience, and so remains a problem.” In Kantian terminology then, in Deep Freedom and Real Persons (Hanna 2018a) I am trying to solve Kant’s problem by developing a metaphysics according to which “transcendental freedom” and “practical freedom” or “autonomy” are themselves fully natural facts. But in my own terminology, I am trying to solve Kant’s problem by developing a metaphysics according to which deep freedom and practical agency are themselves fully natural facts. On the view I propose, our double capacity for free will and practical agency – which, for convenience, I will call free agency – is an irreducible fact. At the same time, our irreducible capacity for free agency does not exist over and above the rest of the physical world – it 318

Will-power

is categorically not a mysterious dualistic, extra-physical fact. On the contrary, it is a fully natural, biological, and neurobiological fact – a natural fact of life. So the key to our free agency is not that we possess mysterious, non-natural, atemporal causal powers to choose or act in violation of the causal laws of nature. On the contrary, it is simply that, insofar as we are minded living organisms, all of our intentional activities – by which I mean the things that we ourselves do, and do not merely happen to us – are inherently vital and non-mechanical, ii some of our intentional activities are naturally creative and authentically original, in just the way that a work created by a “human, all too human” artistic genius is authentically original, but not in any god-like or magical way, and iii none of the general causal laws of nature, especially including the Conservation Laws, is ever violated by us. i

These three natural facts, namely, i* vital, non-mechanical sourcehood, ii* natural creativity, and iii* living in bounded natural open space (having non-equilibrium-thermodynamic “degrees of freedom”), consistently with all of the general causal laws of nature, especially including the Conservation Laws, are all strongly supported by the self-evident, veridical phenomenology of essentially embodied free agency that I will spell out and defend in the final section of this chapter/essay.

O’Shaughnessy and Me I am more than happy to disclose the elective affinities and relations of influence between O’Shaughnessy’s work and mine, because I do think that his work is both exceptionally brilliant and also, unjustly, much-neglected. Moreover, it is clear from the following passage (and many others) in the second, 2008 edition of The Will that our theories of willing, trying, and intentional action are very much in philosophical harmony with each other: [According to the dual aspect theory of physical action] the event of acting and/or willing is one and the same as the full activation of the motor system. It should be noted that this is not that the event of willing activates the motor system. Rather, it is that the immediate act-progenitors, act-desire and act-intention, activate that system, and that this activation-syndrome falls simultaneously under “try,” “strive,” “do,” “will,” and finally by the end of the motor-process also under “arm-raising” or “walking” or “talking” (as the case may be). (O’Shaughnessy 2008: vol. 2, 511) Nevertheless, without actually having painstakingly compared the first, 1980 edition of The Will with the second, 2008 edition, line by line, I am also strongly of the impression that many passages in the first edition, as well as certain passages in his 1973 paper “Trying (as the Mental ‘Pineal Gland’),” suggest that between 1970 and 1980, O’Shaughnessy at times 319

Robert Hanna

slipped back into a classical causal theory of action, of the kind (to my mind) so devastatingly criticized in by Frankfurt (1988c). As Frankfurt crisply puts it: The problem of action is to explicate the contrast between what an agent does and what merely happens to him, or between the bodily movements that he makes and those that occur without his making them. According to causal theories of action, which currently represent the most widely followed approach to the understanding of this contrast, the essential difference between the two types is to be found in their prior causal histories: a bodily movement is an action if and only if results from antecedents of a certain kind. ... Despite its popularity, I believe that the causal approach is inherently implausible and that it cannot provide a satisfactory analysis of the nature of action. I do not mean to suggest that actions have no causes; they are as likely to have causes, I suppose, as other events are. My claim is rather that it is no part of the nature of an action to have a prior causal history of any particular kind. . . . It is integral to the causal approach to regard actions and mere happenings as being differentiated by nothing that exists or that is going on at the time those events occur, but by something quite extrinsic to them – a difference at an earlier time among another set of events entirely. . . . It is no wonder that [causal] theories characteristically run up against counterexamples of a well-known type. . . .No matter what kinds of causal antecedents are designated as necessary and sufficient for the occurrence of an action, it is easy to show that causal antecedents of that kind may have as their effect an event that is manifestly not an action but a mere bodily movement. . . . In my judgment causal theories are unavoidably vulnerable to such counterexamples, because they locate the distinctively essential features of an action exclusively in states of affairs which may be past by the time the action occurs. This makes it impossible from them to give any account whatever of the most salient differentiating characteristic of action: during the time a person is performing an action, he is necessarily in touch with the movements of his body in a certain way, whereas he is necessarily not in touch with them in that way when movements of his body are occurring without his making them. (Frankfurt 1988c: 69–71) Nevertheless, by 2008 or thereabouts, the later O’Shaughnessy and I are completely in agreement that i ii

the spontaneous, desire-driven, telic act of the will that consists in trying, and the event that consists in causally efficaciously bringing about intentional body movements are not only synchronous events, but also – and here is where O’Shaughnessy and I both go one big metaphysical step beyond Frankfurt – ontologically fused events that are, necessarily, complementary proper parts of the rational human minded animal agent’s life over that stretch of time.

It is trying that efficaciously causes the intentional body movement, but not as such or per se, rather only as ontologically fused with a larger causally efficacious dynamic process, because the conscious event of trying constitutes only a proper structural part of the whole dynamic process. So the metaphysical picture is one of individually necessary and individually insufficient but jointly sufficient synchronous and ontologically fused event-causation. Similarly a driver

320

Will-power

causally efficaciously drives a car, but nor as such or per se, rather only as a synchronous proper part of the whole dynamic process involving the causally efficacious car and the causally efficacious driver. The driver is causally efficacious in a different way from the car, insofar as he synchronously autonomously controls the car by means of his own intentional body movements, but he is not literally moving the car along the road by means of those movements – unless, of course, he’s Fred Flintstone. This is not to say, however, that O’Shaughnessy ever officially held something equivalent to The Essential Embodiment Theory of the mind-body relation: [The dual aspect theory advanced in The Will] is a dual aspect theory of bodily action, and nothing else. On the larger issues of materialism, dualism, and dual aspectism generally, I am intentionally silent. . . .To be sure, the specific conclusions reached concerning bodily action accord better with a general dual aspectism than with Cartesian dualism. But that is about all I have to say on these more general issues. (O’Shaughnessy 2008: vol. 1, 44) But at the same time, he also comes really, really close to The Essential Embodiment Theory, even lower down on the same page and onto the next: I believe we shall find in [the question of the constitution of bodily action] a key to the mind-body problem: a key through the agency of bodily action. . . . In my opinion, it enables us to put to philosophical use the fact that mind and body come closest together, and in a wholly overt mode, in the phenomenon of physical action. . . . These two ontological domains are incomparably wedded in the phenomenon of physical action. Through this phenomenon we encounter a partial resolution of a major philosophical problem: a resolution through the agency of the very special character of physical action. (O’Shaughnessy 2008: vol. 1, 44–45) One important domestic difference between O’Shaughnessy’s theory of bodily action and mine, however, concerns what he calls “mere bodily events” like breathing, laughing, and blinking – as opposed to fidgeting, humming a tune to oneself, or moving one’s tongue in one’s mouth while driving one’s car (see, e.g., O’Shaughnessy 2008: vol. 1, 37, 40, 42; vol. 2, 356–360, 463–469). Contrary to most other action-theorists, O’Shaughnessy and I do agree that fidgeting, humming, or moving one’s tongue in one’s mouth while driving are all intentional actions. But at the same time O’Shaughnessy holds that breathing, laughing, and blinking are all passive and non-intentional; whereas I do not. Why not? One reason is that I think that everything that is genuinely mental, no matter how “obscure,” and no matter how supposedly “sub-personal” or “unconscious,” is at least minimally phenomenally and pre-reflectively occurrently conscious. This is what I call The Deep Consciousness Thesis, aka The DCT (Hanna and Maiese 2009: chs. 1–2). In Embodied Minds in Action (Hanna and Maiese 2009), and again more recently in Cognition, Content, and the A Priori (Hanna 2015: 97–105), I have argued for The DCT in three ways: i

via a critique of the very idea of the “unconscious,” as inherently confused between, on the one hand, different degrees and modes of pre-reflective consciousness, e.g., “absorbed” or “absent-minded”/spaced-out consciousness while driving one’s car, and, on the other hand, underlying neurobiological processes (this rejection of the very idea of

321

Robert Hanna

the “unconscious” as confused and bogus is of course similar to Sartre’s critique of the Freudian unconscious [Sartre 1966: 90–96]), ii via the phenomenology of pre-reflective consciousness and other “obscure” modes of consciousness, and iii by pointing out, by way of a reductio argument, that if there are “sub-personal” or “unconscious” mental acts, states, or processes that are truly not conscious, then much or even most of our mental life and mental activity is merely mechanical and robotic, hence nothing that has basic normative value, (moral or non-moral) responsibility, any sort of real guidance or control, etc. The DCT, in turn, radically strengthens my action theory to comprehend anything that flows from the agent herself, as an ultimate source (not however in a classical agent-causal sense, but instead only in the natural-libertarian sense of what I call deep freedom – see the final section below), including absorbed or absent-minded/spaced-out actions like driving one’s car, but also sleeping (whether in dream or dreamlessly), forgetting, breathing, laughing, blinking, fidgeting, humming a tune to oneself, and moving one’s tongue in one’s mouth while driving one’s car. Of course, events that are mentally caused by the minded animal can also fail to flow from the agent herself as an ultimate source, as in mad-scientist-style manipulation cases, pathological mental causation while suffering from a mental illness or a pathological condition like Obsessive-Compulsive Disorder, aka OCD, and so on (again, what is missing here is not classical agent-causation, but instead only deep freedom in the natural-libertarian sense – again, see the final section below, and for its specific application to the case of OCD, see also the final section below). But I do also think that all of these too are at least minimally phenomenally and pre-reflectively occurrently conscious. Moreover, I think of all mental states, acts, or processes, especially including desires, not only as at least minimally phenomenally and pre-reflectively occurrently conscious but also in conative terms: they are also all inherently active, intentionally directed, telic, and at least poised for causing actions. Where mentality is concerned, then, it is through-and-through active or poised for action; hence it is all spontaneous to some extent (Hanna and Thompson 2005). At the same time, there are also different modes of spontaneity, in that some mental acts, states, or processes are inherently somewhat less self-initiated and somewhat more “primed” or “triggered” than others, in the sense that they are the result of vital responsiveness and sensible tracking – e.g., the conscious experience of bodily pain, or what I call nociperception (see Hanna 2018b: ch. 4). Nevertheless, none of the so-called “passive” mental acts, states, or processes, not only including absorbed or absent-minded/spaced-out actions like driving one’s car, but also including nociperception, sleeping (whether in dream or dreamlessly), forgetting, breathing, laughing, blinking, fidgeting, humming a tune to oneself, and moving one’s tongue in one’s mouth while driving one’s car, as well as intellect-infused activities like cognitive insight or discovery, aesthetic experiences, respect, and reverence, are in fact “passive” at all, but on the contrary, instead, they all manifest one or another kind of responsive, sensible spontaneity (cf. Schopenhauer 1907: §42, 169). But whether a first-order desire – say, wanting to drink a beer now – becomes effective and thus becomes the agent’s will at that time, or over time, thus moving the agent all the way to action, is a function of the desire’s position in the overall desire-hierarchy and the overall dynamic desire-structure constituting the agent’s will: as, e.g., when the agent also has the second-order desire of wanting to want to drink a beer now, and is thereby moved all the way to beer-drinking. The dynamic desire-structure, in turn, is the basic source of agency; and in this way, mental causation is agentive will-structuring. So my action theory is neo-hylomorphic, 322

Will-power

deploying an updated Aristotelian metaphysics of form-and-matter, or structure-and-stuffing, as indeed is my theory of the mind-body relation: the mind is the activating form of the living animal body and its far-from-equilibrium, complex, self-organizing thermodynamic processes; and, in turn, agency is a certain configuration (really: a designated set of dynamic configurations/patterns, in orientable space and over elapsed time, making up a real person’s life) of the minded animal’s desires. Thus, for me, the will is an immanent structuring cause of the complex dynamic process constituting the minded animal’s causally efficacious bodily life, via its organization of essentially embodied desires, and not, as in classical causal theories of action, a mental event temporally prior to and ontologically distinct from bodily movement.1 Crucial to this overall account is what Maiese and I, in Embodied Minds in Action (Hanna and Maiese 2009: chs. 3–5 and section 8.3), riffing on the fundamental O’Shaughnessyinspired notion of trying, call “effortless trying.” Effortless trying is pre-reflective and conscious, but non-self-consciously conscious, effective desiring. Hence, effortless trying is a mode of trying that synchronously guides and controls our causally efficaciously bodily life – say, various skillful body movements during dancing, playing sports, walking, or driving one’s car, but also breathing, laughing, and so on – but which does not require self-conscious desires, or reflective self-initiation, rather only pre-reflectively conscious desires and desire-structuring. In turn, moreover, effortless trying is presupposed by all self-conscious or reflectively self-initiated effortful trying. So the (everyday, normal) essentially embodied basic act of, e.g., willing to type a word on my laptop computer is an episode of effortless trying, because it is a pre-reflectively conscious synchronous mental immanent structuring cause of the specific complex dynamic process constituting the minded animal’s causally efficacious hand and arm movements, via its organization of essentially embodied desires; and the basic act itself in all its ontological glory is the complete dual aspect effortless-trying-event that synchronously structurally brings about the non-basic self-conscious, reflectively selfinitiated act of creating a meaningful sequence of words on the laptop screen. As I have indicated, I accept Frankfurt’s basic picture of the hierarchical desire-structure of the will – but not his somewhat intellectualist picture of agentive phenomenology: Now what conceptualization of [agentive phenomenology] fits its contours in the most authentic and perspicuous way? My own preference has been for a model that involves levels of reflexivity or self-consciousness. (Frankfurt 1988f: 164, underlining added) This seems somewhat intellectualist to me in its appeal to self-consciousness, which I construe in a broadly Kantian way as “apperception”: that is, a capacity for making higher-order psychological reports about lower-order conscious states. Now all such reports are judgments, and judgments are essentially conceptual activities. But for me, willing is fundamentally essentially non-conceptual, pre-reflective structural control of how my desires are integrated into my essentially embodied conscious life. Otherwise put, willing is not top-down mental impact from a separate highest-level executive self, but instead simply getting all of my minded bodily resources sufficiently structurally organized and together at any given time. Mental causation is structuring causation carried out by the whole minded animal. Or more idiomatically put: willing is simply getting your essentially embodied, desiderative sh** together, and thereby actually choosing or doing something. So a given second-order desire that fixes precisely which first-order desire is the effective one, at a certain time, is just a certain identifiable desiderative position (I want to want X, I want to want Y, etc.) in the overall will-structure 323

Robert Hanna

of the whole minded animal, indexed to a certain act-context, not some sort of little homuncular atomic willing self that initiates everything. In the final section, I will present a detailed picture of the full set of conditions that need to be satisfied in order for an agent to “get . . . all of [her] minded bodily resources sufficiently structurally organized and together at any given time.” So much for the important domestic differences between O’Shaughnessy’s views and mine. Another way in which O’Shaughnessy’s view and mine importantly converge is with respect to what he calls “the long-term body image” (O’Shaughnessy 2008: vol. 1, 273–298). O’Shaughnessy’s long-term body image, in turn, is equivalent to what Maiese and I, following Shaun Gallagher, call the body-schema (Hanna and Maiese 2009: 69–73; see also Gallagher 2005: 37–38). According to us, the body-schema is the self-evidently veridical, essentially non-conceptual content of primitive bodily awareness; and according to me, it also captures the core of the veridical phenomenology of essentially embodied free agency, to which I now finally turn.

The Veridical Phenomenology of Essentially Embodied Free Agency Many theories of free will – including Natural Libertarianism – hold that necessarily, I can freely choose or do X only if my choosing or doing X, i ii

lacks an antecedent nomologically sufficient cause, and I myself am the causally sufficient ground, origin, or source of choosing or doing X.

The conjunction of these two necessary conditions is what Kant rather misleadingly calls “transcendental freedom,” and what Robert Kane more aptly calls “ultimate responsibility” (Kane 1996: 35). But as I see it, these two necessary conditions provide only part of a single necessary but individually insufficient condition – namely, what I call the real causal spontaneity condition – for deep freedom, ultimate sourcehood, or up-to-me-ness. More precisely, according to Natural Libertarianism, the real causal spontaneity condition says this: A necessary but individually insufficient condition of deep freedom and (moral or nonmoral) responsibility is that all my choosings and doings are causally efficacious in the physical world in a way that is fully consistent with, but also not entailed or otherwise necessitated by, the total set of deterministic or indeterministic causal natural laws, especially including the Conservation Laws, together with all the settled quantity-ofmatter-and/or-energy facts about the past, especially The Big Bang. According to Kant, transcendental freedom is how a human person can, “from itself ” (von selbst) (Kant 1997: A533/B561, 533), be the spontaneous mental cause of certain natural events or processes. If I am that human person, then insofar as I am transcendentally free, it follows that I am an ultimate source of my choices and intentional actions precisely because certain events or processes in physical nature are up to me – or to use Kant’s own phrase, in meiner Gewalt (literally: “in my control” or “in my power”; Kant 1996: Ak 5, 94–95, 215–216). So otherwise put, Kant’s misleadingly labeled transcendental freedom is in effect deep freedom of the will (aka ultimate sourcehood, up-to-me-ness), which in turn necessarily includes but also significantly exceeds the real causal spontaneity condition alone. More specifically, it significantly exceeds the real causal spontaneity condition by adding

324

Will-power







325

Robert Hanna

Classical Libertarianism: Free will and Universal Natural Determinism are mutually inconsistent, free will exists, and Universal Natural Determinism is impossible, either (i) because free will exists as an essential property of a special agent-substance, existing outside the natural causal order or inside that order (classical agentcausationism), or (ii) because some indeterministic processes exist in nature and free will is among them (event-causal indeterminism), or (iii) because some indeterministic processes exist in nature and free will is not among them because it exists over and above natural processes (non-causal indeterminism). Correspondingly, the thesis that free will and Universal Natural Determinism are mutually logically or metaphysically consistent is Classical Compatibilism. So Soft Determinism is a form of Classical Compatibilism. And the thesis that, on the contrary, free will and Universal Natural Determinism are mutually inconsistent is classical Incompatibilism. So Hard Determinism and Classical Libertarianism are both forms of classical Incompatibilism. Here is a brief follow-up comment about classical agent-causal theories. If an (otherwise) classical agent-causal theorist were to reject the metaphysics of substance in favor of a metaphysics of dynamic processes in minded animals of a certain level of cognitive complexity, then her view would be very close to mine. In fact, this is the case with Helen Steward’s view in A Metaphysics for Freedom. For example, she writes: I am convinced that to solve the problems surrounding animal agency simply it is not enough to cast off mistaken theories of causation. . . . It needs also to be shown that real, biological processes might enable us to sustain the idea that an animal may be truly in charge of what it does, so that its actions are more than merely the byproduct of its innards and parts. The task requires some reflection on the organizational principles of living creatures, for it is only through such reflection . . . that we can start to understand where the difference really lies between, on the one hand those things that are true agents, and, on the other, mere machines, entities that nothing will ever be up to, however impressive they may be. . . . I am exceedingly hopeful that the next few years will see the beginnings of a revolution in our conception of the human person, as philosophical and everyday conceptions of the scientific picture of the world are freed from outdated Newtonian ideas and begin to take more note, both of the complexities of science as it really is and of the undeniable fact of our animal nature. (Steward 2012: 198–199) So whether, on the one hand, like Steward, one starts out by officially calling oneself a classical agent causal theorist and then adopts a process metaphysics of agency and persons that in fact undermines and radically transforms the classical metaphysics of agent-substances, or whether, on the other hand, like me, one starts out by officially rejecting the background substance-dualist metaphysics of the classical agent-causal theory and developing a process metaphysics of free will, practical agency, and persons, against the backdrop of The Essential Embodiment Theory of the mind-body relation and mental causation, and then incorporates the core ultimate sourcehood/up-to-me-ness/source-incompatibilist idea of agent causal theories within that non-dualist, process-metaphysics framework, is probably just a nominal or stylistic difference at the end of the day. Thus, a neo-classical or post-classical agent-causal theory like Steward’s, and a source-incompatibilist, Essentially Embodied-Agency theory like mine, can come out to pretty much the same thing. 326

Will-power

In any case, the thrilling upshot of all the argumentation in Deep Freedom and Real Persons (Hanna 2018a: chs. 1–5) is that Natural Libertarianism is neither one of the Three Standard Options, nor a version of Classical Compatibilism, nor a version of classical Incompatibilism. Triply on the contrary, it is a non-classical, non-standard, original and unorthodox, slightly mind-bending, anti-mechanist Libertarian view that constitutes an Incompatibilistic Compatibilism and fuses Wilfrid Sellars’s dialectically opposed conceptions of the Manifest Image and the Scientific Image (Sellars 1963) by means of a metaphysical synthesis I call liberal naturalism. So for the purposes of the rest of this chapter/essay, I will now concentrate on my sub-account of the capacity for veridical psychological freedom, because it captures the nature of essentially embodied agentive phenomenology. Strictly speaking, psychological freedom, when it is unspecified as to whether it is veridical or non-veridical and formulated as a weak disjunction – “veridical-or-non-veridical psychological freedom” – is also a necessary but not sufficient condition of deep freedom; and deep freedom, ultimate sourcehood, or up-to-me-ness, is a necessary but not sufficient condition of principled authenticity. Psychological freedom, per se, without regard to its veridicality or non-veridicality, is my first-order consciousness of being both negatively and positively free. Or otherwise put, psychological freedom, per se, is my subjective experience of having an unfettered and really causally spontaneous will. This consciousness, in turn, can be either (i) a correct or true consciousness of being both negatively and positively free, such that it is an actual fact that I am both negatively and positively free, in which case, it is what I call veridical psychological freedom, so that at any time, the intentional subject is either in one kind of state or else in the other, never both, or (ii) an incorrect or false consciousness of being both negatively and positively free, that is, a mere seeming to be both negatively and positively free, such that I in fact am neither negatively nor positively free, in which case, it is what I call non-veridical psychological freedom. Correspondingly, I am also committed to the thesis of strong metaphysical disjunctivism about the difference between veridical psychological freedom and non-veridical psychological freedom,2 which says that iii

iv

veridical psychological freedom and non-veridical psychological freedom essentially share no intentional or phenomenological content whatsoever, even if they accidentally share some (or even many) other psychological or non-psychological properties, and the difference between veridical psychological freedom and non-veridical psychological freedom is, in principle, inherently discriminable for rational human agents, even if, in context, or in a specific range of contexts, it is actually undiscriminated because the agent’s capacity for discrimination is adversely affected or suppressed in that context or those contexts.

So obviously psychological freedom, per se, when it is non-veridical psychological freedom, is consistent with my not really being negatively or positively free. Furthermore, non-veridical psychological freedom is also inherently conceptually determined, that is, inherently open to “cognitive penetration.” By sharp contrast, veridical psychological freedom is essentially non-conceptual; hence, it is also pre-reflectively conscious, or non-self-conscious and non-reflective, and therefore impervious to “cognitive penetration.” The currently popular 327

Robert Hanna

“cognitive penetration thesis” just means that for any given kind of human cognition, its representational content is necessarily and sufficiently determined by propositional content and thereby also necessarily and sufficiently determined by conceptual content, on the standard assumption that all propositional content is, ultimately, reducible to conceptual content. So my central point here is that since veridical psychological freedom is essentially nonconceptual, then it is impervious to cognitive penetration, and thereby immune to propositional or conceptual errors. Nevertheless, since rational human agents like us are also self-conscious and reflective, then the possession of psychological freedom for rational human agents like us, whether it is veridical psychological freedom or non-veridical psychological freedom, also entails a further capacity for having self-directed beliefs to the effect that we are negatively and/or positively free. In the case of veridical psychological freedom, obviously, those self-directed beliefs to the effect that we are negatively and/or positively free, by virtue of the essentially non-conceptual content of veridical psychological freedom, are true beliefs; whereas in the case of non-veridical psychological freedom, by virtue of its erroneous propositional and conceptual content, those self-directed beliefs are false beliefs. Thus, we can be deceived about our freedom and have what I call sheer illusions of deep freedom. These sheer illusions can occur in drug-induced or pathological dreams,3 hallucinations, or, most poignantly, in causally overdetermined pathological waking psychological states, such as cases of sociopathic paranoid schizophrenics who believe they are choosing and acting with free will, but are actually delusional and insane, and acting under an irresistible compulsion when they commit crimes, hence are correctly legally judged to be “not guilty by reason of insanity.” One such case – or it seems to me from the available evidence – is the real-world waking nightmare of a former Yale Law School student, Ketema Ross: Early one morning in 2007, Ross heard President George W. Bush [Yale] ‘68 telling him that his next-door neighbors were traitors who needed to be gotten rid of. Ross broke into the elderly couple’s apartment and beat them with a broom handle. (They both survived the attack.) Charged with assault, he pleaded not guilty by reason of insanity. Now Ross says he has recovered his sanity, and a court order says he is no longer “a substantial danger.” And, after seven years of confinement in a psychiatric hospital, he has regained his freedom, mostly: by court order, he was conditionally discharged on January 11 [2015]. (Bass 2015: 49) Now psychological freedom, per se, is not a sufficient condition of deep freedom, precisely because, as some philosophers (including Spinoza, Leibniz, Hume, and more recently Frankfurt in 1988b, and many other Frankfurt-inspired philosophers of agency) have correctly pointed out, both the pre-reflective consciousness, and also the self-conscious or reflective belief, of having an unfettered and spontaneously really causal will, and of being both negatively and positively free, are perfectly consistent with Universal Natural Determinism. Formulated in my terminology, if either Universal Natural Determinism, Universal Natural Indeterminism, or the disjunctive combination of these that I have called Natural Mechanism were true, then all of these conscious or self-conscious states would be cases of non-veridical psychological freedom and/or false self-directed beliefs about my deeply free will. Nevertheless, both veridical psychological freedom and also veridical psychological freedom’s discriminability from non-veridical psychological freedom collectively yield

328

Will-power

a conjunctive necessary epistemic condition of deep freedom, by way of the capacity for self-commitment to a live option, or Kierkegaardian Either/Or. This condition says that no one could have a capacity for self-commitment to a live option X and at the same time either (i) be in a state of non-veridical psychological freedom, or (ii) truly believe herself to be prevented from choosing X or doing X, or (iii) truly believe herself to be inwardly or outwardly compelled to choose X or do X, or (iv) truly believe herself to be otherwise unable to choose or do what she wants. More precisely, by a self-consciously or reflectively fettered, epiphenomenal, will I will mean one’s self-conscious or reflective awareness to the effect that, and correspondingly one’s belief about oneself to the effect that, one is helplessly violated by inner or outer forces. But more briefly put, this is when someone vividly feels like a natural automaton (biochemical puppet, moist robot, “meat puppet,” etc.), or like a tool in the hands of some other powerful manipulative agent or agency. This would be vividly true of, e.g., the Ketema Ross case, and also of pathological conditions like OCD. More specifically, mentally caused body-movements under OCD fall short of free agency in at least four ways: i

they fail the condition of veridical psychological freedom, since in mentally causing body-movements brought about by her OCD, the subject self-consciously experiences herself as being helplessly overwhelmed by inner forces, ii they fail the Kierkegaardian Either/Or condition, since in mentally causing bodymovements brought about by her OCD, the subject does not commit to these as a live option for her – on the contrary they are something ultimately forced on her by neurobiological processes internal to her body, iii they fail the ownership condition, since in mentally causing body-movements brought about by her OCD, the subject is not performing intentional body-movements that flow from that person herself, as an ultimate source – on the contrary, they are something that is external to the self, even despite their being mentally caused by the subject, and iv as a consequence of (i) to (iii), the bodily movements brought about by the subject’s OCD are not something for which she is deeply (morally or non-morally) responsible. So I am saying that in order to have deeply free will, then, we must not have a self-consciously or reflectively fettered, epiphenomenal, will: on the contrary, we must not only have veridical psychological freedom, but also be at least fully disposed to believe, or actually believe, ourselves to have an unfettered, non-epiphenomenal, real causally spontaneous will. Ironically, this is as true of self-styled Hard Determinists as it is of everyone else. I am absolutely sure that when these philosophers choose and act, under normal conditions, they do not actually feel or believe in accordance with their own metaphysics of free will – that they do not either really feel like biochemical puppets, moist robots, “meat-puppets,” etc., or really believe themselves to be natural automata of any kind. To be sure, the logical scope of their philosophical beliefs extends universally over all people, including themselves: but epistemically speaking, it is one thing to apply property P to everyone, including of course oneself, third-personally, and quite another thing to apply property P to oneself, first-personally. Universal instantiation is neither semantically nor epistemically equivalent to first-personal indexical predication. If everyone is P then it necessarily follows that I am P, because I am one of the people in the domain of discourse. But if everyone is P, even if I believe that everyone

329

Robert Hanna

is P, it does not necessarily follow that I really believe that I am P, because I still have to identify myself with one of the many people in the domain of discourse, and I might self-consciously or non-self-consciously refuse to do that. Otherwise put, if Hard Determinists really did believe that they themselves were natural automata (biochemical puppets, moist robots, “meat puppets,” etc.), then obviously they would seek – or at least obviously they would at least need – psychiatric help for the treatment of well-attested symptoms of schizophrenia, like the unfortunate Ketema Ross (Bass 2015; see also Mellor 1970; Frith 1992; Spence 1996). Of course, they do not really believe it; hence, they have effectively compartmentalized, segregated, and causally isolated their purported philosophical beliefs from their basic agentive commitments or “beliefs” in the specific sense of essentially non-conceptual, hierarchical-desiderative, volitional believing-in or faith (Glaube) – as it were, their heart of hearts – and are, as a consequence, about as healthy and sane as professional academic philosophers usually are. Furthermore, and more generally, I submit that no healthy, sane real human person ever really and truly believes himself or herself, in the sense of essentially non-conceptual, hierarchical-desiderative, volitional believing-in or faith, their heart of hearts, to be a natural automaton (biochemical puppet, moist robot, etc.), no matter what their free will metaphysics says. Consider, e.g., Classical Compatibilism/Soft Determinism. The very idea of “real” freedom of choice or action, when taken together with a self-consciously or reflectively fettered, epiphenomenal, will, would be as absurd and pointless as freedom of action together with either a naturally determined or a powerfully manipulated will. Indeed, as Frankfurt rediscovered and influentially pointed out, psychological freedom, per se, although it is consistent with Universal Natural Determinism (namely, when it is non-veridical psychological freedom), nevertheless remains essential to our rational human personhood, as veridical psychological freedom, at the level of effective first-order desires, together with the self-conscious or reflective awareness of having veridical psychological freedom, at the level of what Frankfurt calls decisive or self-identifying second-order volitions (Frankfurt 1988b, 1988d, 1988f ). This is because a self-consciously or reflectively fettered, epiphenomenal, will utterly defeats and undermines our belief in our own intentional agency. Now consider Hard Determinism. Virtually every theory that falls under these rubrics also has a “debunking strategy” or “error-theory,” which says that our brains mechanically create the cognitive illusion that we are really free, even though we are really natural automata. But then those theories have the following serious epistemic problem. Suppose that you hold the following view: Free will is an illusion. Our wills are simply not of our own making. Thoughts and intentions emerge from background causes of which we are unaware and over which we exert no conscious control. We do not have the freedom we think we have. (Harris 2012: 5) If that is true, then we are all natural machines with an irresistibly strong tendency to create cognitive illusions for ourselves. Therefore, under the supposition that her theory is true, any holder of such a view cannot rule out the real possibility that she has created a cognitive illusion for herself by defending Natural Mechanism together with a debunking strategy or an error-theory. But if she cannot rule this out, then she is not rationally justified in believing in her own theory. So her belief in her own theory is cognitively self-stultifying. This conclusion, in turn, debunks the would-be debunkers.

330

Will-power

Now suppose that someone were to object to that last argument by saying: This argument, purporting to show that Hard Determinism is self-stultifying, goes by very fast. Would you embrace the claim that since one cannot rule out the possibility that one is in a lifelong dream, one is not rationally justified in any of one’s external world beliefs? This is how I would reply: No, I wouldn’t embrace this claim. Indeed, I’ve published two papers on Cartesian dream skepticism, in which I explicitly refute it (Hanna 2000; Hanna 1992). Nevertheless, the crucial point in the present context is whether, as you imply, there’s a basic analogy between the Cartesian dream-skeptic’s (as it turns out, unsound) argument and my self-refutation argument against the Hard Determinist. And what I’d say to that is also no, for there’s a crucial disanalogy between what I am critically arguing here against the Hard Determinist, and what the Cartesian dream skeptic would (unsoundly, as it turns out) argue. In turn, that’s because, since the Hard Determinist is explicitly committed to a neurobiologically-based error theory that applies to everyone’s capacities for reasoning philosophically, then in an easily-accessible, nomologically-identical, nearby possible world, it also applies to the Hard Determinist’s capacity for reasoning philosophically too. Now when there are relevantly alternative skeptical possibilities, those possibilities do indeed undermine rational justification. Hence, at least so far, my self-stultification argument holds up. Interestingly and importantly, the closely related view that Derk Pereboom calls “Hard Incompatibilism” does not have the entailment that there is no genuine concept of free will (Pereboom 2001, 2014). According to the Hard Incompatibilist, real free will is possible, but actually does not exist. So while, according to Pereboom, there is no actual fact of real free will, there is nevertheless an intelligible concept and also a vivid phenomenology of real free will, and correspondingly, these can be taken to support the thesis that, in some non-trivial sense, we really are “deterministic agent-causes” (Pereboom 2015). In these respects, Hard Incompatibilism is a significantly more subtle and defensible view than Hard Determinism. But where Hard Incompatibilism has serious internal difficulties, I think, is in reconciling the explicit rejection of the fact of real free will with an explicit acceptance of the genuine concept of real free will, the vivid phenomenology of free will, and the thesis about our “deterministic agent-causal” powers. This mixed attitude of rejection/acceptance seems to me to produce serious cognitive (and practical) dissonance. According to Pereboom, I really and truly am a biochemical puppet and a moist robot: I know this via mechanistic physics and biology, and “scientific philosophy.” Yet at the same time, by virtue of the genuine concept and vivid phenomenology of real free will, I also cannot help believing outside my laboratory or study that I am (in some sense) deeply free and (morally or non-morally) responsible for my choices and actions, and therefore, somewhere else in my head and in my heart, I feel in my bones that I am not a biochemical puppet or moist robot. When I am wearing my white lab coat and doing experiments, or in my open-necked Brooks Brothers shirt and nice blue jeans teaching philosophy, I believe I am really a puppet and a robot; but when I get home, change into my comfortable T-shirt and baggy shorts, and open a beer, I believe I am really not a puppet or robot. This is philosophical schizophrenia, not a stable philosophical position.

331

Robert Hanna

More specifically, Pereboom claims that the “standard phenomenology of agency” is evidentially neutral as between Classical Libertarianism on the one hand, and Hard Determinism/Hard Incompatibilism/Soft Determinism on the other. And according to Pereboom, the “standard phenomenology of agency” is this: The phenomenology appears to reveal that in paradigm cases, actions are caused not solely by events or states, but are rather actively caused by agents themselves. (Pereboom 2015: 278) Pereboom’s evidential-neutrality conclusion may follow if we construe the “standard phenomenology of agency” narrowly enough, but two things have gone seriously wrong here. First, Pereboom has not considered the inherently evidentially richer, essentially embodied, agentive phenomenology of spontaneous, vital, non-mechanical sourcehood, natural creativity in “natural open space,” and more generally he has completely overlooked the phenomenology of agency as we find it in the work of O’Shaughnessy, Schopenhauer, Kant, Kierkegaard, and Frankfurt. More specifically, the five crucial additional elements in this inherently evidentially richer phenomenology of agency are: i

ii iii iv v

the O’Shaughnessy-inspired and Schopenhauer-inspired identification of the agent’s sourcehood with her inherently spontaneous and vital, hence non-mechanical, necessary, and complete minded animal embodiment, the Kant-inspired free-agent-as-creative-natural-artist-in-natural-open-space analogy, the Existentialism-inspired Kierkegaardian “either/or” conception of self-commitmentto-a-live-option, the Frankfurt-inspired hierarchical-desire conception of the will and personhood, and the equally Kant-inspired and Existentialism-inspired idea that free agency is inherently normatively motivated by the pursuit of principled authenticity.

Second, and correspondingly, Pereboom has not considered, or even looked-for, any view about the metaphysics of free will and practical agency that is even remotely like Natural Libertarianism; on the contrary, he has strategically boxed-in his metaphysical options by critically considering only the various versions of Classical Libertarianism as opposing doctrines. So the plausibility and soundness of Pereboom’s argument depends entirely on his having started with a needlessly poverty-stricken and boxed-in domain of phenomenological materials and metaphysical options. Therefore, to conclude this critical line of argument by modus tollens-ing Pereboom’s modus ponens: if we do not start with these needlessly poverty-stricken and boxed-in materials, but instead on the contrary seriously consider the inherently evidentially richer agentive phenomenology of essentially embodied, spontaneous, vital, non-mechanical sourcehood, natural creativity in natural open space, self-commitment-to-a-live-option, willing-as-hierarchical-desiring, and the pursuit of principled authenticity, together with the metaphysics of Natural Libertarianism, then Hard Incompatibilism is clearly false. It is also very important to point out in this connection that having a self-consciously or reflectively fettered, epiphenomenal, will is categorically not the same as the classical, Lord-Byron-style, Romantic self-consciousness of being in the grip of a grand passion that carries you away with it. The crucial contrast between these two is the inherent, categorical difference between 332

Will-power



333

Robert Hanna

Related topics See Chapters 13 (on Merleau-Ponty), 19 (Horgan and Nida-Rümelin), 20 (Smith), 23 (Strawson), and 24 (De Monticelli).

Notes 1 Strictly speaking, Davidsonian theories of action do not require that the cause of action be temporally prior, since Davidson’s theory is a third-person, interpretation-theoretic account, whereby a given belief-desire pair constitutes a reason that “rationalizes” an underlying physical event-cause, upon which the belief-desire pair are naturally or nomologically supervenient, and supervenience is a synchronous relation. Nevertheless Davidson’s and Davidsonian theories have other serious problems, including epiphenomenalism and its own deviant causal chain worry. See Hanna and Maiese (2009: ch. 3). 2 Analogously, I am committed to strong metaphysical disjunctivism about sense perception. See Hanna (2015: ch. 3). 3 On the other hand, ordinary, non-drug-induced, non-pathological dreams exhibit real causal spontaneity, as well as satisfying the other conditions for deep freedom. So, strictly speaking, ordinary dreams are what I call veridical illusions of deep freedom, because they include i ii

essentially non-conceptual veridicality about our deep freedom, together with propositional and conceptual errors about many details of the course of one’s actual life and of the actual world in which ordinary dreaming deep freedom purports to be exercised,

in a way that’s highly analogous to the familiar bent-stick-in-water or bigger-moon-near-thehorizon cases of veridical illusion in the philosophy of perception. In turn, this entails that our ordinary sleeping and dreaming life is as much a part of our free agential lives as ordinary waking life is. But the philosophy of sleep is a long story for another day – or to be more precise, for another book.

References Arpaly, N. (2007) Merit, Meaning, and Human Bondage, Princeton, NJ: Princeton University Press. Bass, C. (2015) “By Reason of Insanity,” Yale Alumni Magazine, pp. 48–53, http://www.yalealumnimagazine.com/articles/4076/by-reason-of-insanity. Evans, G. (1982) Varieties of Reference, Oxford: Clarendon/Oxford University Press. Frankfurt, H. (1988) The Importance of What We Care about, Cambridge, MA: Cambridge University Press. Frankfurt, H. (1988a) “Alternate Possibilities and Moral Responsibility,” in H. Frankfurt (1988), The Importance of What We Care about, pp. 1–10. Frankfurt, H. (1988b) “Freedom of the Will and the Concept of a Person,” in H. Frankfurt (1988), The Importance of What We Care about, pp. 11–25). Frankfurt, H. (1988c) “The Problem of Action,” in H. Frankfurt (1988), The Importance of What We Care about, pp. 42–52. Frankfurt, H. (1988d) “Identification and Externality,” in H. Frankfurt (1988), The Importance of What We Care about, pp. 58–68. Frankfurt, H. (1988e) “The Importance of What We Care About,” in H. Frankfurt (1988), The Importance of What We Care about,: pp. 80–94. Frankfurt, H. (1988f ) “Identification and Wholeheartedness,” in H. Frankfurt (1988), The Importance of What We Care about, pp. 159–176. Frith, C. (1992) The Cognitive Neuropsychology of Schizophrenia, Hillsdale, NJ: Erlbaum. Gallagher, S. (2005) How the Body Shapes the Mind, Oxford: Clarendon/Oxford University. Press. Goethe, J.W. (1990) Faust, trans. W. Kaufmann, New York: Anchor Books. Hanna, R. (1992) “Descartes and Dream Skepticism Revisited,” Journal of the History of Philosophy 30, 377–398. Hanna, R. (2000) “The Inner and the Outer: Kant’s ‘Refutation’ Reconstructed,” Ratio 13, 146–174. Hanna, R. (2006) Rationality and Logic, Cambridge, MA: MIT Press.

334

Will-power Hanna, R. (2015) Cognition, Content, and the A Priori: A Study in the Philosophy of Mind and Knowledge, aka THE RATIONAL HUMAN CONDITION, Vol. 5, Oxford: Oxford University Press. Hanna, R. (2018a) Deep Freedom and Real Persons: A Study in Metaphysics, aka THE RATIONAL HUMAN CONDITION, Vol. 2, New York: Nova Science. Hanna, R. (2018b) Kantian Ethics & Human Existence, aka THE RATIONAL HUMAN CONDITION, Vol. 3, New York: Nova Science. Hanna, R. and Maiese, M. (2009) Embodied Minds in Action, Oxford: Oxford University Press. Hanna, R. and Thompson, E. (2005) “Neurophenomenology and the Spontaneity of Consciousness,” in E. Thompson (ed.), The Problem of Consciousness, Calgary, AL: University of Alberta Press, pp. 133–162, https:// www.academia.edu/9962718/Neurophenomenology_and_the_Spontaneity_of_Consciousness. Harris, S. (2012) Free Will, New York: Free Press. Horgan, T. (2011) “The Phenomenology of Agency and Freedom: Lessons from Introspection and Lessons from Its Limits,” Humana Mente Journal of Philosophical Studies 15, 77–97, http://www. humanamente.eu/PDF/Issue15_Paper_Horgan.pdf. Ismael, J. (2016) How Physics Makes Us Free, Oxford: Oxford University Press. Kane, R. (1996) The Significance of Free Will, Oxford: Oxford University Press. Kant, I. (1996) Immanuel Kant: Practical Philosophy, trans. M. Gregor, Cambridge, MA: Cambridge University Press. Kant, I. (1997) Critique of Pure Reason, trans. P. Guyer and A. Wood, Cambridge, MA: Cambridge University Press. In-text citations include page-numbers from both the A edition (1781) and B (1787) editions of the CPR and also the Cambridge edition. Kim, J. (1993) Supervenience and Mind, Cambridge, MA: Cambridge University Press. Kim, J. (2005) Physicalism, or Something Near Enough, Princeton, NJ: Princeton University Press. Meat Puppets, The (1994) “We Don’t Exist,” https://search.yahoo.com/yhs/search?p=the+meat+puppets+we+don%27t+exist+youtube&ei=UTF-8&hspart=mozilla&hsimp=yhs-003. Mellor, C.S. (1970) “First Rank Symptoms of Schizophrenia,” British Journal of Psychiatry 117, 15–23. O’Shaughnessy, B. (1973) “Trying (as the Mental ‘Pineal Gland’),” Journal of Philosophy 70, 365–386. O’Shaughnessy, B. (1980) The Will: A Dual Aspect Theory, 2 vols., Cambridge, MA: Cambridge University Press. O’Shaughnessy, B. (2008) The Will, 2 vols., second edition, Cambridge, MA: Cambridge University Press. Pereboom, D. (2001) Living Without Free Will, Cambridge, MA: Cambridge University Press. Pereboom, D. (2014) Free Will, Agency, and Meaning in Life, Oxford: Oxford University Press. Pereboom, D. (2015) “The Phenomenology of Agency and Deterministic Agent-Causation,” in M. Altman and H. Gruenig (eds.), For Horizons of Authenticity in Phenomenology, Existentialism, and Moral Psychology: Essays in Honor of Charles Guignon, New York: Springer, pp. 277–294. Rorty, R. (1979) Philosophy and the Mirror of Nature, Princeton, NJ: Princeton University Press. Rorty, R. (1982) Consequences of Pragmatism, Minneapolis: University of Minnesota Press. Sartre, J.-P. (1966) Being and Nothingness, trans. H. Barnes, New York: Washington Square Press. Schopenhauer, A. (1907) On the Fourfold Root of the Principle of Sufficient Reason, trans. Mme. K. Hillebrand, London: George Bell and Sons, http://www.gutenberg.org/files/50966/50966h/50966-h.htm. Schuessler, J. (2013) “Philosophy That Stirs the Waters,” New York Times, 29 April 2013, http://www. nytimes.com/2013/04/30/books/daniel-dennett-author-of-intuition-pumps-and-other-toolsfor-thinking.html?emc=eta1&_r=0. Sellars, W. (ed.) (1963) “Philosophy and the Scientific Image of Man,” in Science, Perception, and Reality, New York: Humanities Press, pp. 1–40. Spence, S.A. (1996) “Free Will in the Light of Psychiatry,” Philosophy, Psychiatry, and Psychology 3, 75–90. Steward, H. (2012) A Metaphysics for Freedom, Oxford: Oxford University Press. Wittgenstein, L. (1953) Philosophical Investigations, trans. G.E.M. Anscombe, New York: Macmillan. Wollheim, R. (1984) The Thread of Life, Cambridge, MA: Harvard University Press.

335

22 THE PHENOMENOLOGY OF AGENCY AND THE COGNITIVE SCIENCES Shaun Gallagher

When engaged in action, the experience of agency includes the experience of one’s actions and of oneself. This is a relational phenomenology, that is, it is not just an experience of an action as unrelated to me, and not just an experience of myself without some kind of relation to an action or set of actions. Part of this relation involves the notion of intention, and the kind of intention in play. Both action and intention, and therefore, the experience of agency, are complex phenomena. Some aspects of action are non-conscious and sub-personal; others involve consciousness. To understand the phenomenology of agency, we need to understand the complexity involved in action and intention. Moreover, since an individual’s action, and the experience of agency, most often occur in social contexts where joint action is possible, and social practices predominate, it’s not enough to look at processes that are merely individual. In this chapter, I’ll first consider some of the phenomenological aspects of individual agency. I’ll supplement this account by drawing on the extensive cognitive science and some of the controversy that has developed around the concept of sense of agency. This will lead us to consider how agency relates to intention. Finally, I’ll introduce some important qualifications on these analyses by considering the effects of social context and cultural practices on the experience of agency.

Action and the phenomenology of agency One way to characterize the experience of agency is to distinguish it from the experience or sense of ownership (Gallagher 2000; Tsakiris et al. 2007). The sense of ownership can pertain to one’s body and movement. It’s the sense that it is my body that is moving, and this applies equally to voluntary and involuntary movements. Even if I involuntarily start to fall, this is still something happening to me – it is my movement. In contrast, the sense of agency is the sense that I am the originator of an action, or that I am controlling the action as it unfolds. A sense of agency accompanies an intentional or voluntary action, but does not accompany a reflex or involuntary movement. In this regard, the sense of agency has been regarded as a first-order or pre-reflective experience that is built into action, generated by pre-action (pre-motor) processes (Marcel 2003; Chambon et al. 2014).

336

Agency and the cognitive sciences

There is no unanimous agreement about the distinction between sense of agency and sense of ownership, however. Anthony Marcel (2003), for example, uses the phrases interchangeably. John Campbell (1999) has suggested that the sense of agency is really one form of the sense of ownership, of which there are two. On the one hand, (1) a person may regard a bodily experience as her own just because she finds it to be part of her stream of experience, or happening within the boundaries of her body. On the other hand, (2) a person can claim a bodily experience – for example the experience of a movement – as her own because she originated or caused the movement, i.e., she recognizes herself to be the agent of this movement. Some theorists refer to the latter as an experience of authorship or agency (e.g., Humphrey 1992; Bortolotti and Broome 2009; Grünbaum and Zahavi 2013). A further complication, however, is that some of these authors consider the senses of ownership and agency to involve an inference or second-order reflective consciousness, the result of judging an action to be one’s own, or attributing an action to oneself. Stephens and Graham (2000), for example, distinguish between the attribution of ownership (or subjectivity) and the attribution of agency. These attributions are the result of reflective judgments concerning whether an action seems to fit with a person’s understanding, theory, or narrative about who they are and what they are capable of doing, and, retrospectively, whether a particular action truly belongs to them. This type of second-order evaluation or judgment, which may also be prospective and involved in intention formation (see Mylopoulos 2017 for example), is frequently referenced in moral contexts in conjunction with reflective decision-making (e.g., Frankfurt 1988). It is important, then, to distinguish between a sense of agency, understood as a first-order, non-inferential, pre-reflective experience, and a sense (or attribution) of agency, as a second-order or reflective judgment (Marcel 2003; Vosgerau and Newen 2007). A full phenomenology of agency would include a description of both reflective and pre-reflective aspects. Is a minimal sense of agency anything more than a pre-reflective sense of being the originator of the action, and in control of the action as it proceeds? Elisabeth Pacherie has suggested that it may also include awareness of a goal, awareness of an intention to act . . . awareness of movements, sense of activity, sense of mental effort, sense of physical effort. . . experience of authorship, experience of intentionality, experience of purposiveness, experience of freedom, and experience of mental causation. (Pacherie 2007: 6) It’s not clear, however, that each of these aspects will show up in the first-order phenomenology of agency as such. These different aspects may rather be the product of a reflection that attempts to pick out these distinctions. As I engage in action I do not necessarily experience a sense of effort distinct from my sense of control, although I may be able to make that distinction in my reflective examination of what I do experience. As such, that distinction may not exist in my experience of agency on the level of my immersed engagement with action. This may depend on circumstances, however, and, for example, whether the agent has some degree of expertise in regard to the action. One way to think about the various components of the sense of agency listed by Pacherie is that one or more of the components may come into the phenomenological foreground under specific circumstances, but otherwise remain recessive and unnoticed, although contributing to the overall feel of what it’s like to engage in a specific action. If the angle of my

337

Shaun Gallagher

reach is odd, or if I happen to be fatigued, it may be that the feeling of effort is a larger part of my phenomenology. If I have a pain in my shoulder, I may be more aware of the precise details of my movements. If the action is important for my friend, or normatively constrained in some special way, then I may experience purposiveness in a way that I do not ordinarily experience it. In this respect, one can consider the various components of the sense of agency as organized in a kind of dynamical gestalt where some of the components remain in the background, outside of our awareness, but still perhaps contributing to the overall feel of the action, while other components come to the fore under specific circumstances, and enter explicitly into the phenomenology of agency. In this regard, however, we can think that in typical circumstances, the primary elements that make up the sense of agency pertain to pragmatic intention and goal. The sense of agency in some regard reflects the structure of action itself. Action is made up of a complex collection of physical processes, from neural activation (which may involve chemical and neural transmitter processes that work at a molecular level), to muscle activation, to basic movements of joints and muscles, that combine to form basic actions (e.g., reaching or grasping), which combine to form a complex action, which may be part of a set of actions. What precisely am I aware of when I am engaged in such action? This may vary from one action and circumstance to another, but one cannot be aware of molecular or neuronal processes, and generally speaking one is not aware of muscle activation, or, in most circumstances, any details of the movements of joints and muscles. In many cases I am aware of what I am doing and that I am doing it at the most pragmatic level of description (“I’m getting a drink”) rather than how I am doing it ( Jeannerod 2003). Thinkers from Aristotle to Anscombe agree on this: In general, as Aristotle says, one does not deliberate about an acquired skill; the description of what one is doing, which one completely understands, is at a distance from the details of one’s movements, which one does not consider at all. (Anscombe 1963: 54) Likewise, the phenomenology of a pre-reflective sense of agency, then, can be thin and recessive (Haggard 2005), and without distinctions between intentional causation, initiation, control, degrees of freedom, purposiveness, effort, and so on. This is not to say that such things cannot be found by reflecting on my action, as a phenomenologist might do. But in our everyday, natural attitude, when we are engaged in action, we tend to be agents rather than phenomenologists. The idea that action involves both non-conscious processes and processes that we can be aware of can be seen in relatively simple actions. When I reach to pick up my cup to take a drink, not only am I not aware of neuronal processes and muscle activations at the subpersonal level, I’m not aware of the detailed shaping of my fingers into a grasp that is appropriate for grabbing the cup in a specific way – in contrast to a different grasp that would be appropriate for picking up some other item, such as my pen, and also in contrast to a different grasp that would be appropriate for grabbing the same cup in order to throw it, or wash it up, rather than for purposes of taking a drink (Marteniuk et al. 1987; Jeannerod 1997; Ansuini et al. 2013). The kinematics for these different grasps are different, but I am not aware of the kinematics of my grasp per se. Such kinematic processes, however, may contribute to what I am aware of as I reach to grasp the cup. I may be aware of extending my arm and grabbing the cup – and the way these movements kinaesthetically feel (e.g., as smooth, or as awkward) may depend on the particular kinematics involved. To grab the cup from an odd angle just 338

Agency and the cognitive sciences

feels different (more awkward) than grabbing it from a more habitual angle. In that type of case, the experience of effort may come to the fore, whereas otherwise, I don’t notice it. In most cases my awareness is taken up by my intended goal and is reflected in my response to a question such as, “What are you doing?” The answer is not “activating my neurons” or “exercising my muscles,” or “reaching to grasp my cup” – it is rather, “taking a drink.” All of the details of my action, even some of which I can become reflectively aware, for example, in odd circumstances, are typically occluded, so to speak, by my pragmatic intent or goal.

The cognitive science of the sense of agency Are there implicit markers of agency? One proposal is that intentional binding (Haggard et al. 2002; Moore and Obhi 2012) operates as an implicit marker. In the case of voluntary action in contrast to involuntary movement, perceived times of the action and its effect are temporally compressed, i.e., shifted toward each other (Figure 22.1). Although the subject is unaware that an underestimation in time perception has taken place, the shift may be one element that contributes to the difference in feeling between voluntary and involuntary movement (Braun et al. 2018 indicate that the experimental literature is not conclusive on this and call for further study). Sensory attenuation is another proposed implicit marker. We experience less sensory feedback for voluntary action than for passive movements (Blakemore et al. 1999). Mylopoulos (2015) suggests that in these experiments the attenuation is actually related to action effects rather than to the motoric aspects of action, and he takes this to be an objection to counting sensory attenuation as a marker for the sense of agency. It’s not clear, however, that sensory attenuation of the action effect is not a marker for the sense of agency. A number of experiments suggest that it is (see my discussion of Grünbaum, below). When experimenters explicitly ask subjects about their experience, it is uncertain whether the results reflect what they have experienced at a pre-reflective level as a sense of agency, or an attribution or judgment of agency. In such experiments the subject may be shown a video of their own or someone else’s movement (or consequence of a movement), with or without variable time delays or postural differences. They are then asked to identify or judge whether that movement is theirs (e.g., Daprati et al. 1997). Generally speaking, the pre-reflective sense of agency has been linked to low-level sensorimotor processes; in contrast, reflective judgments about agency have been linked to higherlevel cognitive processes including background beliefs and contextual knowledge relating to action. There is not always a consistency between these levels, and they can be dissociated from one another (Moore 2016; Moore et al. 2012). For example, if an unexpected action outcome happens when I am alone in a room, I may have no pre-reflective sense of agency

Figure 22.1 The intentional binding paradigm. Subjects underestimate the actual time interval between action onset and sensory outcome (based on Moore and Obhi 2012)

339

Shaun Gallagher

for the action, but I may nonetheless attribute it to myself since no other explanation is available (Synofzik et al. 2008). Low-level sensorimotor processes that purportedly generate the pre-reflective sense of agency have been explained in terms of sub-personal, predictive comparator mechanisms in the brain. A sensory-feedback comparator is said to compare the intended or desired result of movement with sensory feedback from the actual movement. A feed-forward comparator allows for faster correction of action by comparing intended movement with efference copy (a copy of the motor command) or predicted sensory input allowing for ongoing correction during movement, and prior to actual sensory feedback (Frith 1992; Gallagher 2005a; Moore and Haggard 2008; Chambon et al. 2014). The idea that these sub-personal motor control processes generate a sense of agency is sometimes contested since these processes are unconscious and we have no access to them at a personal level. Since an agent is conscious of her intention and her action, Dan Wegner (2002) and others have proposed that our experience of agency is due to a comparison between intention and action on a conscious level. When one’s intention and action match one attributes self-agency for the action. This is a retrospective rather than predictive model, and it suggests that the experience of agency is illusory, an epiphenomenal add-on to the sub-personal control mechanisms that are the real causes of the action. Evidence for this is found in experiments that show we attribute agency for actions that are not ours when we are primed to think about that action, or are given environmental and social cues beforehand (Wegner et al. 2004; see Nahmias 2005; Carruthers 2010 for critical discussion). Under certain conditions, then, it is clear that one might be able to generate an illusory experience of agency. It is not clear, however, that the experience of agency is always illusory. Our sensorimotor systems characteristically operate in an integrative and intermodal fashion. Specific areas of the brain, such as the insula (e.g., Farrer and Frith 2002), integrate sensory afference, motor efference, and sensory cues from the environment. This type of integration, including both motor control processes and predictive processing of priors and environmental cues, can modulate intentional binding and the sense of agency. Moore and Haggard (2008) have shown that intentional binding occurs between an action (pressing a button) and an effect (sounding of a tone) when the agent is primed to expect the tone, even if she did not cause it. They also show that when the key press actually causes the tone, even if the agent’s expectation for the tone is low, the agent attributes the effect to her action, and retrospectively triggers the binding effect. This suggests that an integration of various motor control processes and perceptual cues (including perhaps feelings of fluency or interruptions of fluency [see Chambon et al. 2014; Buhrmann and Di Paolo 2017]) will provide a better explanation of the phenomenology of agency (Moore et al. 2009; David 2012). Consistent with the dynamical gestalt model mentioned above, in this integration of factors, under varying circumstances, one factor may be weighted differently than another. This view has been cashed out in a predictive processing model which assumes that the brain integrates different agency cues characterized by signal variance, and therefore some degree of uncertainty (Hohwy 2007; Kilner et al. 2007; Moore et al. 2009). Each cue is weighted probabilistically, according to its individual precision, and the brain derives an overall prediction of self- versus other-agency (Rohde et al. 2016). It is likely, for example, that proprioception is one important internal cue for the sense of agency. Proprioceptive awareness of one’s movement, however, is attenuated and not overly precise. The attenuation might be considered a “flaw” (e.g., de Vignemont 2014: 998), but only if one assumed that proprioception was meant to deliver precise, objective awareness of one’s movement. Neither proprioception, nor any one of the other agency cues, however, functions in an isolated manner by itself. Reliability is rather measured in terms of the 340

Agency and the cognitive sciences

whole system and its integrated functioning. Moreover, we should consider questions about reliability and precision in pragmatic (action-oriented) terms rather than in epistemic terms (Gallagher 2017). That is, proprioception, functioning along with other cues (in modalities such as touch, vision, interoception, etc.), provides a pragmatic sense of agency related primarily to the subject’s actions and action possibilities. The system is not set up to give the agent an exact or detailed sense of the “how” of her actions. We get sufficient precision when we need it via the mix of senses, and pragmatic estimates are good enough in most cases for motor control, and certainly for a sense of agency. Braun et al. (2018: 8) argue that this type of predictive processing cue-integration theory has three important advantages: (1) it provides a parsimonious explanation of how different agency cues can be optimally integrated to one overall agency inference mechanism; (2) it is flexible enough to integrate many different agency cues; and (3) it allows to integrate prior knowledge (top–down constraints) into the model. Given the important role played by action (explained as a form of active inference) in predictive processing, the ability of the system to track its own actions in distinction from perceived changes in the environment is central. It would seem reasonable to think that whatever system operations are involved in action, this self/non-self distinction can easily underpin an intrinsic sense of self-agency. Accordingly, Friston (2012) suggests that “implicit in a model of [active inference] sampling is a representation or sense of agency.” Despite a relative broad consensus that most of us experience “the feeling that we are in control of what we are doing most of the time [i.e., ] the normal sense of agency” (Haggard and Chambon 2012: 390), recent debates have put into question whether there is anything like a minimal or pre-reflective sense of agency when one is engaged in action. Thor Grünbaum, for example, writes: “One can accept that there are conscious mental states or processes that are proprietary to action, such as practical deliberation, practical decision-making and intending, without accepting that there is a special phenomenal feeling associated with motor control” (2015: 3314). Grünbaum’s argument focuses on the theoretical neuroscience, and specifically on the particular account of the sense of agency that considers it the product of comparator mechanisms involved in motor control. He doesn’t deny that a comparator mechanism may be involved in motor control, but he challenges the idea that comparator processes generate a distinct experience of agency. There are two points that I want to address in his argument. First, he argues that if the sense of agency is generated by such sub-personal comparator mechanisms, that means that it is intention-free. By “intention-free” I take him to mean that, on such accounts, the sense of agency is generated even if the agent has not formulated a prior, personal-level intention to act in a certain way. But such prior intention is not the only form of intention that may be involved in action (see below). Reaching for my cup of tea as I work on my computer does not require that I consciously deliberate and form a plan to do so. Without a prior intention, my action may still involve a present intention-in-action, and/or, most basically, motor intentions – a form of intention intrinsic to its kinematic properties, which is perceptually observable even in another person’s actions (Becchio et al. 2012), and kinaesthetically experienced by the agent in the way the movement feels. Furthermore, at least some comparator models include the idea that there is some functional (sub-personal) element in the system that counts as an intention, and that this intention is precisely what is compared to efference copy or sensory reafference to facilitate motor control (e.g., Frith 1992; Frith and 341

Shaun Gallagher

Gallagher 2002). In this respect it’s not clear that the sense of agency can be characterized as intention-free, unless one equates intention exclusively with prior intention. In that case, the only experience of agency would be the result of a reflective prospective intention formation, or a retrospective attribution – that is, a judgment of agency rather than a pre-reflective sense of agency. This would also make the phenomenology of agency much less embodied, since what is denied is a sense of agency generated in motor control processes, and what remains is a higher-order judgment about agency. Nonetheless, Grünbaum may be right about comparator mechanisms not generating a pre-reflective sense of agency; there are reasons to question whether comparator models of motor control offer the best explanation (see, e.g., Synofzik et al. 2008; Friston 2011; Buhrmann and Di Paolo 2017; Braun et al. 2018). Still, as already implied by more integrative models (e.g., Moore and Haggard 2008), the pre-reflective sense of agency may be generated in other ways. This leads to the second point. Grünbaum points to an important qualification involved in a number of experiments on the sense of agency. He focuses on an action-recognition experiment by Daprati et al. (1997). Subjects are asked to perform a hand movement and monitor it on a computer screen, which shows either their own hand movement, or a hand movement made by someone else. They are then asked whether what they saw was their action or not. Typically, subjects mistook the actions of the other’s hand as their own in about 30% of the cases; schizophrenic subjects who had a history of delusions of control and/or hallucinations misjudged 50–77% of the cases. The problem, as Grünbaum notes, is that on each trial the subject is in fact engaging in the action of moving his own hand, and, accordingly, should have a pre-reflective sense of agency for moving their hand. I note that the same objection can be raised in regard to other experiments, for example, Farrer and Frith (2001; see Gallagher 2005b, 2012 for discussion). Grünbaum rightly notes that in such experiments, rather than reporting a sense of agency based on motor control processes (since hypothetically, given the subject’s hand movement in each trial, such motor control processes would remain constant across all trials), subjects were simply reporting differences in what they were monitoring on the screen. I think this is right, but the conclusion is not what Grünbaum proposes, that they are therefore not reporting a sense of agency. The alternative conclusion is that the pre-reflective sense of agency is more complex than an experience that is generated by sub-personal motor control processes. Based on these experiments, we can say that the sense of agency involves at least two aspects – one having to do with the control of bodily movement in action, and one having to do with the intentional aspect of the action, i.e., what the action accomplishes in the world (Gallagher 2005b, 2012; Haggard 2005). The first aspect would be tested by asking the subjects whether they were the one’s moving their hands. The second is tested by asking them whether what they see (on the computer screen) is what they are doing or what they intended to do. Both points raised by Grünbaum lead to considerations about intention and how intention relates to the experience of agency. I’ll discuss these considerations in the next section.

Agency and intention Although agency usually involves an intended action, there are some instances where the sense of agency diverges from intention and follows unintended consequences. Marcel (2003) gives the example of treading on someone’s toe. We apologize for doing so because we have a sense of agency for the action – we did not intend to do it, but we have no doubt that this was something that we did – that we originated the action that led to this result even if we did not feel entirely in control of the outcome. 342

Agency and the cognitive sciences

If, however, the phenomenology of agency with its “complex and nonunitary phenomenal structure” (Braun et al. 2018) typically does involve some form of intention, then the relation between intention and agency needs to be explained. There are no free-floating intentions that are unrelated to a subject (agent) of intention and to an object of intention. A prior intention, expressed as “I intend to do A” (or a collective intention, “We intend to do A”) where “A” is some action, is something like a foreshadowing of “I do thus and so in order to A,” where “thus and so” signifies some kind of basic activity (Hornsby 2013) by an agent. In addition, intentions are integrated with actions by specification of the goal: “I A in order to do X,” where X is some intended goal. Elisabeth Pacherie (2006, 2007) outlines a three-fold demarcation of intentions which helps to explain how the various aspects of action and intention interconnect. A first type of intention, and perhaps the standard way of thinking about intention as something prior to action, is what she designates D(istal) intention. D-intentions are formed prior to action and represent the action as a whole. Typically, they are detached from the precise situation of the action and they specify types rather than tokens of actions. Their content is conceptual and descriptive. For example, I may decide to purchase a new energy-efficient automobile next month. Or I may plan to go grocery shopping later today. Nothing needs to be specified about how or where or with whom I will do these things, however, so the D-intentions remain somewhat abstract. D-intentions are also “subject to distinctive normative pressures for consistency and coherence: in particular, they should be means-end coherent, consistent with the agent’s beliefs and consistent with other intentions he or she may have” (Pacherie 2007: 3; see Harman 1976; Bratman 2007: 26). D-intentions are typically at play in prospective intention formation, where one reflectively decides on or plans out a course of action. They may result from relatively abstract and conceptual considerations, or may be formulated spontaneously, as the result of situated reflection. They may also be part of what is at stake in retrospective attribution where an agent acknowledges or endorses her past action as something she intended or not. D-intentions can be conceived of as operating on the narrative timescale. The second type of intention is the present or proximate intention (P-intention) and is equivalent to what John Searle (1983) calls “intention-in-action.” P-intentions are operative at the time of action and may serve to implement action plans inherited from D-intentions. When the time comes for me to purchase my new car, I have to take action, the details of which are specified by particular and present circumstances – where I am located, where the car dealership is located, how I get from A to B, how I proceed in my negotiations with the dealer, and so forth. At each step I determine how I put my D-intention into action. My D-intention must broadly govern my P-intention as I decide in the moment to go to a particular dealer, and how I will get there. P-intentions anchor the action plan both in time and in the situation of action and thus effect a transformation of the descriptive contents of the action plan into perceptual actional contents constrained by the present spatial as well as non-spatial characteristics of the agent, the target of the action, and the surrounding context. (Pacherie 2007: 3) Note, however, that not every intentional action requires that there is a D-intention. Many of my intentional actions are unplanned. In some circumstances (for example, when watching a football match on television), I may “absent-mindedly” find myself heading for the fridge to get a beer. Although I did not engage in an explicit decision-making process, or 343

Shaun Gallagher

plan anything out in this regard, my negotiation of the environment which helps me to get to the fridge is seemingly informed by my P-intention to get a beer. In order to purchase my car or get my beer, of course, I typically have to move my body, which may itself be postured in different ways. I may have to get up from my seat, navigate through doorways, walk across rooms, etc. This is, of course, what we consider to be taking action. If I don’t move, I never get to the point of fulfilling the P-intention. Pacherie introduces an additional form of intention to address this requirement – the M(otor) intention: The final stage in action specification involves the transformation of the perceptualactional contents of P-intentions into sensorimotor representations (M-intentions) through a precise specification of the spatial and temporal characteristics of the constituent elements of the selected motor program. (Pacherie 2007: 3). Is it possible that M-intentions could carry the action to an intended goal without the help of D- or P-intentions? Once initiated, for example, an action may be carried along to completion without much monitoring or sustaining effort. Pacherie notes that the guiding and monitoring control functions that are usually tied to P-intentions may be taken over by M-intentions. In the case of well-practiced or easily accomplished habitual actions, for example, environmental affordances may solicit and sustain the action such that the control mechanisms might be fully governed by M-intentions. Driving your car is a good example. In such habitual or situated forms of coping do we lose the sense of agency? There is some evidence that the answer is no. Dijksterhuis et al. (2006) have shown that more unconscious (or less consciously deliberative) strategies deliver better results than more conscious deliberation even in complex cases that call for decision, such as buying a car. It has also been shown that conscious evaluative deliberation about the same object can lead to less consistent choices over time (Levine et al. 1996; these results are controversial; see Nieuwenstein and Van Rijn 2012 for some qualifications). Although the experimenters did not ask about sense of control or agency in the contrasting cases, there was no indication that at the point of decision people felt less in control using unconscious strategies. Typically, we might think, P-intentions in some way get cashed out in M-intentions, in the same way that D-intentions get cashed out in P-intentions. M-intentions are clearly tied to sub-personal processes of motor control, and in this sense are coupled with the prereflective sense of agency generated by such processes. On a most basic level, they operate on an elementary timescale of milliseconds – which is the same timescale where one finds intentional binding. P-intentions may be coupled with what we called the intentional aspect of the action, i.e., what the action accomplishes in the world. As such it may remain pre-reflective, especially if the action runs smoothly; or it may rise to the level of reflective consciousness, especially if the action is complex or difficult, or if something goes wrong and some strategic attention to what I am doing is required. Accordingly, all three types of intention may contribute to the experience of agency. As we saw, Pacherie claims that the experience of agency may contain an experience of intentional causation, a sense of initiation and control, which is further analyzable into more specific experiences. If the phenomenology of agency on the pre-reflective level remains thin, i.e., short-lived and experientially recessive, the full phenomenology may be complex and include the results of ongoing prospective, perceptual, and retrospective monitoring,

344

Agency and the cognitive sciences

reflecting different types of intention (Gallagher 2012). In addition, Pacherie mentions a long-term experience of agency: a sense of oneself as an agent apart from any particular action, i.e. a sense of one’s capacity for action over time, and a form of self-narrative where one’s past actions and projected future actions are given a general coherence and unified through a set of overarching goals, motivations, projects and general lines of conduct. (2007: 6) It is important to point out that this long-term sense of agency, along with the actiondependent experiences of agency that may be tied to D-, P-, and M-intentions, may be modified by factors that extend beyond just brain and body.

Sense of agency and social context Although the question about how the sense of agency is generated goes beyond the phenomenological issue, I think that the full story also goes beyond what experimental cognitive science has so far told us about sub-personal processes or reflective judgments. This is not to deny the role of sub-personal processes. If something does go wrong with sub-personal motor control processes, then we can expect some change in the sense of agency. The phenomenology may be different if, for example, motor control processes are disrupted, as may be the case in some symptoms of schizophrenia, such as delusions of control (Frith 1992; Frith and Gallagher 2002). That issue aside, however, because the embodied agent is always situated in a social and cultural environment, or as phenomenologists tend to put it, the agent is always in the world, a number of other factors may contribute to the constitution of the sense of agency – and our immediate body self-awareness more generally. How do other people and social forces affect the experience of agency? Prospective and retrospective dimensions of intention formation and our attribution and evaluation of actions are often shaped by others, and by the situations in which we encounter others. What my friends consider an appropriate choice may influence my decision to buy a certain kind of car, for example. Thinking of intention formation and motivation strictly in internalist terms of simply having the right mental state (e.g., a belief ) about a particular action plan (e.g., Nagel 1970) ignores the role of social context. Deliberations are often shared deliberations, and intentions often take shape in our conversations with others. Accordingly, deliberation and intention-formation are not merely a set of mental states or propositional attitudes, or causal brain states, if they are often co-constituted with others. Phenomena such as peer pressure, implicit or explicit social referencing, or our habitual behavior when in the presence of others may detract from or enhance one’s feeling of agency. Some actions, as in the case of joint actions, are distributed across more than one agent. On the one hand, the distributed nature of such actions does not necessarily reduce the individual’s sense of agency, although one can understand such agency “as made possible by the way in which it is embedded within a larger social matrix” (Goodwin 2018: 440). On the other hand, one’s sense of agency could be dramatically attenuated under certain conditions (such as military training, or in circumstances involving rage), and in ways that may put into question whether an agent (in any subjective sense) is in control of ensuing actions (Protevi 2009: 152). Consider some extreme cases, like compulsive or addictive situations, hysteria, or conversion disorder. In addictive behavior, for example, the loss of a sense of agency for one’s

345

Shaun Gallagher

actions is not just the result of chemically induced dependency. Compulsive drug-related behaviors correlate with neither the degree of pleasure reported by users nor reductions in withdrawal symptoms as measured in placebo studies. Robinson and Berridge (2000) propose an “incentive-sensitization” model in which pathological addiction correlates highly with the salience of socially situated drug-related behaviors and stimuli. The agent’s perception of his social world can be altered so that some circumstances and some specific people become hedonically significant. Brain regions that mediate incentive sensitization are the same areas that process action specification, motor control, and social cognition and that correlate with intentional deliberation, social navigation, and action (Grant et al. 1996; Breiter et al. 1997; Wang et al. 1999). The perceptual salience of the social situation, accordingly, contributes to intention formation and the experience of agency, sometimes enhancing but also (as in extreme addictive behavior) sometimes robbing the agent of any sense of control. More generally, normative constraints and cultural practices can modulate our autonomy and have an effect on our long-term sense of agency. Our actions may be limited by what is deemed acceptable behavior within specific institutions and subcultures. Consider also that class, race, and gender issues may not only limit the kind of behavior in which I feel comfortable engaging, but also affect my pre-reflective sense of agency. For example, in regard to the idea mentioned above, that in most cases my awareness is taken up by my intended goal and that the details of the “how” of my action are typically occluded by my pragmatic intent or goal, we can note some qualifications or limitations concerning the claim about “most” or “typical” cases. It’s possible to criticize the narrow focus of studies of such things as motor control (body-schematic) processes and minimal aspects of the sense of agency that do not take into account social and cultural contexts. Experiments in psychology and cognitive science, as is well known, tend to study populations of so-called WEIRD outliers (westernized, educated (college student) populations from industrialized, rich democracies; cf. Henrich et al. 2010). Most of these subjects are Caucasian and, at least in the past, were mostly drawn from male populations. It’s possible, then, that cognitive science tends to universalize a white-male model of agency and body awareness more generally, and thus fails to take into account the effects of race and gender norms (e.g., Grosz 1994; Young 2005; Fanon 2008; Weiss 2015). Frantz Fanon’s criticism of Merleau-Ponty’s analysis of body schema is still relevant today. Fanon suggests that how the sense of agency and body self-awareness is said to operate for the “normal” subject, i.e., as non-objectified and as recessively in the background, is impossible for those whose bodies are deemed inherently inferior, that is, for those who are ruled out, from the outset, from achieving the status of ‘normal’ subjects. Being viewed and treated as a ‘normal’ subject [is] an inherited privilege that white bodies enjoy and that non-white bodies do not. (Weiss 2015: 86) In specific circumstances, certain groups of people may be compelled to attend to their movements in ways that bring the “typically” recessive aspects of agency or how they are moving to attention (e.g., Yancy 2014). Unfortunately, there are few cross-cultural studies that target the sense of agency. One study, by Barlas and Obhi (2014), indicates that both Western and non-Western groups showed similar performance on reflective judgments of agency but differences in intentional binding (suggesting differences in the pre-reflective sense of agency) for consonant (pleasurable) outcomes (stronger in Western agents than in Won-western agents). Cultural background can differentially influence low versus high levels of the sense of agency (also see 346

Agency and the cognitive sciences

Markus and Kitayama 2010). Kabeer (1999: 438), specifically in regard to the circumstances of women in different cultures, points to numerous social situations and practices that can modulate one’s autonomy and sense of agency (also see Hanmer and Klugman 2016; Cheong et al. 2017). For example, Ellen Judd shows that Chinese women (in rural China) who demonstrate high degrees of agency, and perceive themselves as such, “can only express and activate this agency provided that doing so does not place them in competition with men” (1990: 48). One might think that questions about autonomy in social, cultural, and normative contexts are somewhat divorced from more basic questions about a sense of agency described in terms of motor control processes. There is much more to discuss in this respect, but I’ll conclude by simply pointing to some experiments that suggest that socio-cultural factors go all the way down and have an effect on basic motoric and perceptual experience. In several experiments Soliman and Glenberg (2014; Soliman et al. 2013) have shown the effects of cultural and religious differences on spatial perception and reaction times. In one study, for example, these effects were most pronounced when subjects self-identified as culturally interdependent versus culturally individualist. I think their conclusion can extend to basic conceptions of autonomy and the sense of agency (see Gallagher 2020): [C]ulture enters the scene not as a self-contained layer on top of behavior, but as the sum of sensorimotor knowledge brought about by a bodily agent interacting in a social and physical context. As such, culture [permeates] the web of sensorimotor knowledge, and can only be arbitrarily circumscribed from other knowledge. . . . [Cultural biases are] acquired and maintained through immersion in environments with different patterns of interpersonal interactions. . . . These interactions ‘tune’ the sensorimotor system. . . By tuning, we mean a process of neuroplasticity, or learning. (Soliman and Glenberg 2014: 209–210).

Conclusion That the experience of agency is a complex phenomenon is evident not only in phenomenological and empirical studies of action and intention, but also in considerations that document different experiences resulting from differences in culture, gender, or race. For a full understanding of the pre-reflective sense of agency and the reflective judgments that we make about our agency, it seems clear that an interdisciplinary approach is required. Not only a strict phenomenology, but psychological and neuroscientific studies, as well as social and cultural studies can throw some light on this experience. Other chapters in this volume have more to say about social context, autonomy, and the role of action-affordances.

Related topics See Chapters 19 (Horgan and Nida-Rümelin), 20 (Smith), 21 (Hanna), and 23 (Strawson).

Acknowledgments The Humboldt Foundation’s Anneliese Maier Research Award and the Australian Research Council (ARC) grant, Minds in Skilled Performance. DP170102987, supported the author’s research on this chapter. 347

Shaun Gallagher

References Anscombe, G.E.M. (1963) Intention, Oxford: Basil Blackwell. Ansuini, C., Cavallo, A., Bertone, C., and Becchio, C. (2013) “The visible face of intention: Why kinematics matters,” Frontiers in Psychology 5, 815. Barlas, Z. and Obhi, S.S. (2014) “Cultural background influences implicit but not explicit sense of agency for the production of musical tones,” Consciousness and Cognition 28, 94–103. Becchio, C., Manera, V., Sartori, L., Cavallo, A., and Castiello, U. (2012) “Grasping intentions: From thought experiments to empirical evidence,” Frontiers in Human Neuroscience 6, 1–6. Blakemore, S., Frith, C., and Wolpert, D. (1999) “Spatiotemporal prediction modulates the perception of self-produced stimuli,” Journal of Cognitive Neuroscience 11, 551–559. Bortolotti, L. and Broome, M. (2009) “A role for ownership and authorship in the analysis of thought insertion,” Phenomenology and the Cognitive Sciences 8 (2), 205–224. Bratman, M. (2007) Structures of Agency, New York: Oxford University Press. Braun, N., Debener, S., Spychala, N. et al. (2018) “The senses of agency and ownership: A review,” Frontiers in Psychology 9, 535. Breiter, H.C., Gollub, R.L., Weisskoff, R.M., Kennedy, D.N., Makris, N., Berke, J.D., et al. (1997) “Acute effects of cocaine on human brain activity and emotion,” Neuron 19 (3), 591–611. Buhrmann, T. and Di Paolo, E. (2017) “The sense of agency – A phenomenological consequence of enacting sensorimotor schemes,” Phenomenology and the Cognitive Sciences 16, 207–236. Campbell, J. (1999) “Schizophrenia, the space of reasons and thinking as a motor process,” Monist 82 (4), 609–625. Carruthers, G. (2010) “A problem for Wegner and colleagues’ model of the sense of agency,” Phenomenology and the Cognitive Sciences 9, 341–357. Chambon, V., Sidarus, N., and Haggard, P. (2014) “From action intentions to action effects: How does the sense of agency come about?,” Frontiers in Human Neuroscience 8, 320. Cheong, Y.F., Yount, K.M., and Crandall, A.A. (2017) “Longitudinal measurement invariance of the women’s agency scale,” Bulletin of Sociological Methodology 134 (1), 24–36. Daprati, E., Franck, N., Georgieff, N., et al. (1997) “Looking for the agent: An investigation into consciousness of action and self-consciousness in schizophrenic patients,” Cognition 65 (1), 71–86. David, N. (2012) “New frontiers in the neuroscience of the sense of agency,” Frontiers in Human Neuroscience 6, 161. de Vignemont, F. (2014) “A multimodal conception of bodily awareness,” Mind 123: 989–1020. Dijksterhuis, A., Bos, M.W., Nordgren, L.F., van Baaren, R.B. (2006) “On making the right choice: The deliberation-without-attention effect,” Science 311, 1005–1007. Fanon, F. (2008) Black Skin, White Masks. Trans. R. Philcox, New York: Grove Press. Farrer, C. and Frith, C.D. (2002) “Experiencing oneself vs. another person as being the cause of an action,” NeuroImage 15, 596–603. Frankfurt, H.G. (1988) The Importance of What We Care about, Cambridge, MA: Cambridge University Press. Friston, K. (2011) “Embodied inference: Or ‘I think therefore I am, if I am what I think’,” in W. Tschacher and C. Bergomi (eds.), The Implications of Embodiment, Exeter: Imprint Academic. Friston, K. (2012) “Embodied inference and spatial cognition,” Cognitive Processing 13 (S1), S171–S177. Frith, C.D. (1992) The Cognitive Neuropsychology of Schizophrenia, Hillsdale, NJ: Lawrence Erlbaum Associates. Frith, C.D. and Gallagher, S. (2002) “Models of the pathological mind,” Journal of Consciousness Studies 9 (4), 57–80. Gallagher, S. (2000) “Philosophical conceptions of the self: Implications for cognitive science,” Trends in Cognitive Science 4 (1), 14–21. Gallagher, S. (2005a) “Dynamic models of body schematic processes,” in H. De Preester and V. Knockaert (eds.), Body Image and Body Schema, Amsterdam: John Benjamins, pp. 233–250 Gallagher, S. (2005b) How the Body Shapes the Mind, Oxford: Oxford University Press. Gallagher, S. (2012) “Multiple aspects of agency,” New Ideas in Psychology 30, 15–31. Gallagher, S. (2017) “Self-defense: Deflecting deflationary and eliminativist critiques of the sense of ownership,” Frontiers in Human Neuroscience 8, 1612. Gallagher, S. (2020) Action and Interaction, Oxford: Oxford University Press. Goodwin, C. (2018) Co-operative Action, Cambridge, MA: Cambridge University Press.

348

Agency and the cognitive sciences Grant, S., London, E.D., Newlin, D.B., et al. (1996) “Activation of memory circuits during cueelicited cocaine craving,” Proceedings of National Academy of Sciences of the United States of America 93 (21), 12040–12045. Grosz, E. (1994) Volatile Bodies: Toward a Corporeal Feminism, Bloomington: Indiana University Press. Grünbaum, T. (2015) “The feeling of agency hypothesis: A critique,” Synthese 192, 3313–3337. Grünbaum, T. and Zahavi, D. (2013) “Varieties of self-awareness,” in K.W.M. Fulford, M. Davies, R. Gipps, G. Graham, J. Sadler, G. Stanghellini, and T. Thornton (eds.), The Oxford Handbook of Philosophy and Psychiatry, Oxford: Oxford University Press, pp. 221–239. Haggard, P. (2005) “Conscious intention and motor cognition,” Trends in Cognitive Sciences 9 (6), 290–295. Haggard, P., Clark, S., and Kalogeras, J. (2002) “Voluntary action and conscious awareness,” Nature Neuroscience 5 (4), 382–385. Haggard, P. and Chambon, V. (2012) “Sense of agency,” Current Biology 22 (10), R390–R392. Hanmer, L. and Klugman, J. (2016) “Exploring women’s agency and empowerment in developing countries,” Feminist Economics 22 (1), 237–263. Harman, G. (1976) “Practical reasoning,” Review of Metaphysics 29, 431–463. Henrich, J., Heine, S.J., and Norenzayan, A. (2010) “The weirdest people in the world?,” Behavioral and Brain Sciences 33 (2–3), 61–83. Hohwy, J. (2007) “The sense of self in the phenomenology of agency and perception,” Psyche 13 (1), 1–20. Hornsby, J. (2013) “Basic activity,” Aristotelian Society Supplementary 87 (1), 1–18. Humphrey, N. (1992) A History of the Mind, New York: Simon and Schuster. Jeannerod, M. (1997) The Cognitive Neuroscience of Action, Oxford: Blackwell Publishers. Jeannerod, M. (2003) “Consciousness of action and self-consciousness,” in J. Roessler and N. Eilan (eds.), Agency and Self-Awareness, Oxford: Clarendon Press Judd, E.R. (1990) “ ‘Men are more able’: Rural Chinese women’s conceptions of gender and agency,” Pacific Affairs 63 (1), 40–61. Kabeer, N. (1999) “Resources, agency, achievements: Reflections on the measurement of women’s empowerment,” Development and Change 30 (3), 435–464. Kilner, J.M., Friston, K.J., and Frith, C.D. (2007) “Predictive coding: An account of the mirror neuron system,” Cognitive Processing 8 (3), 159–166. Levine, G.M., Halberstadt, J.B., and Goldstone, R.L. (1996) “Reasoning and the weighting of attributes in attitude judgments,” Journal of Personality and Social Psychology 70, 230–240. Marcel, A. (2003) “The sense of agency: Awareness and ownership of action,” in J. Roessler and N. Eilan (eds.), Agency and Awareness, Oxford: Oxford University Press, pp. 48–93. Markus, H.R. and Kitayama, S. (2010) “Cultures and selves: A cycle of mutual constitution,” Perspectives on Psychological Science 5 (4), 420–430. Marteniuk, R.G., MacKenzie, C.L., Jeannerod, M., Athenes, S., and Dugas, C. (1987) “Constraints on human arm movement trajectories,” Canadian Journal of Psychology 41: 365–378. Moore, J. and Haggard, P. (2008) “Awareness of action: Inference and prediction,” Consciousness and Cognition 17, 136–144. Moore, J.W. (2016) “What is the sense of agency and why does it matter?,” Frontiers in Psychology 7, 1272. Moore, J.W., Middleton, D., Haggard, P., and Fletcher, P.C. (2012) “Exploring implicit and explicit aspects of sense of agency,” Consciousness and Cognition 21, 1748–1753. Moore, J.W. and Obhi, S.S. (2012) “Intentional binding and the sense of agency: A review,” Consciousness and Cognition 21, 546–561. Moore, J.W., Wegner, D.M., and Haggard, P. (2009) “Modulating the sense of agency with external cues,” Consciousness and Cognition 18, 1056–1064. Mylopoulos, M. (2015) “Agentive awareness is not sensory awareness,” Philosophical Studies 172 (3), 761–780. Mylopoulos, M. (2017) “A cognitive account of agentive awareness,” Mind & Language 32 (5), 545–563. Nagel, T. (1970) The Possibility of Altruism, Oxford: Clarendon Press. Nahmias, E. (2005) “Agency, authorship, and illusion,” Consciousness and Cognition 14 (4), 771–785. Nieuwenstein, M., & van Rijn, H. (2012) “The unconscious thought advantage: Further replication failures from a search for confirmatory evidence,” Judgment & Decision Making 7 (6), 779–798. Pacherie, E. (2006) “Towards a dynamic theory of intentions,” in S. Pockett, W.P. Banks and S. Gallagher (eds.), Does Consciousness Cause Behavior?, Cambridge, MA: MIT Press, pp. 145–167.

349

Shaun Gallagher Pacherie, E. (2007) “The sense of control and the sense of agency,” Psyche 13 (1), 1–30. Protevi, J. (2009) Political Affect, Minneapolis: University of Minnesota Press. Robinson, T. and Berridge, K. (2000) “The psychology and neurobiology of addiction: An incentive-sensitization view,” Addiction 95 (8s2), 91–117. Rohde, M., Van Dam, L.C.J., and Ernst, M.O. (2016) “Statistically optimal multisensory cue integration: A practical tutorial,” Multisensory Research 29, 279–317. Searle, J. (1983) Intentionality, Cambridge, MA: Cambridge University Press. Soliman, T., Gibson, A., and Glenberg, A.M. (2013) “Sensory motor mechanisms unify psychology: The embodiment of culture,” Frontiers in Psychology 4, 885. Soliman, T.M. and Glenberg A.M. (2014) “The embodiment of culture,” in L. Shapiro (ed.), The Routledge Handbook of Embodied Cognition, London: Routledge, pp. 207–220. Stephens G.L. and Graham G. (2000) When Self-Consciousness Breaks, Cambridge, MA: MIT Press. Synofzik, M., Vosgerau, G., and Newen, A. (2008) “Beyond the comparator model: A multifactorial two-step account of agency,” Consciousness and Cognition 17, 219–239. Tsakiris, M., Schütz-Bosbach, S., and Gallagher, S. (2007) “On agency and body-ownership,” Consciousness and Cognition 16 (3), 645–660. Vosgerau, G. and Newen, A. (2007) “Thoughts, motor actions, and the self,” Mind and Language 22, 22–43. Wang, G.J., Volkow, N.D., Fowler, J.S., et al. (1999) “Regional brain metabolic activation during craving elicited by recall of previous drug experiences,” Life Sciences 64 (9), 775–784. Wegner, D.M. (2002) The Illusion of Conscious Will, Cambridge, MA: MIT Press. Wegner, D.M., Sparrow, B., and Winerman, L. (2004) “Vicarious agency: Experiencing control over the movements of others,” Journal of Personality and Social Psychology 86, 838. Weiss, G. (2015) “The normal, the natural, and the normative,” Continental Philosophy Review 48, 77–93. Yancy, G. (2014) “White gazes: What It feels like to be an essence,” in E.S. Lee (ed.), Living Alterities, Albany: State University of New York Press, pp. 43–64. Young, I.M. (2005) On Female Body Experience, Oxford: Oxford University Press.

350

Phenomenology of Agency 2: Aspects of Agency

23 THE PHENOMENOLOGY OF FREE AGENCY Galen Strawson

The experience of freedom You set off for a shop on the evening of a national holiday. You intend to buy a cake with your last ten-pound note in order to supplement the good supply of food you already have. When you get to the shop it’s about to close, along with all the other shops. As it happens, there’s one cake left; it costs ten pounds. On the steps of the shop someone is holding out an Oxfam tin, collecting money for Famine Relief. You stop. It is absolutely clear to you—it seems absolutely clear to you—that it’s entirely up to you what you do next. It is (it seems) perfectly clear to you that you are truly, radically free to choose what to do, in such a way that you’ll be wholly responsible for whatever you do choose. What could be more plain? You can put the money in the Oxfam tin, or go in and buy the cake, or just walk away. You’re not only completely free to choose. You’re not free not to choose. Standing there, you may believe determinism is true.1 You may believe that in five minutes’ time you’ll be able to look back on the situation you are now in and say, of what you will by then have done, ‘It was determined that I should do that’. But even if you do wholeheartedly believe this—right now—it doesn’t seem to be able to touch your current sense of the absoluteness of your freedom and moral responsibility; not in any way at all. Such situations of choice and action, large and small, morally significant and morally neutral occur constantly in human life. Shall I have white wine or red? Shall I go to Vietnam or Canada? Shall I go to The Elephant Room or The Continental Club? The experience of freedom of choice and action lies at the very heart of the human experience of agency. There are interesting exceptions, both good (a sense of inevitability in artistic creation) and bad (experience of compulsion in cases where one is doing something one believes to be wrong and believes that one could possibly not do). But the normal case is clear: in the normal case—and I’m going to stick to the human case—experience of freedom of action is—from a young age—essentially constitutive of the overall experience of agency. This is not to say that it is vivid in every case. This experience of freedom exists in many degrees, from one-year-old Louis’s experience of being in control of his limbs, to eight-year-old Louisa’s sense of freedom of choice between two different flavors of ice cream, to thirty-year-old Lucy’s experience of being free in a way that makes her wholly morally responsible for what she does. I’m going to put 352

The phenomenology of free agency

little Louis aside and focus principally on experience of freedom in agency—in deliberation, choice, decision, passage à l’acte—that involves experience of responsibility, moral or not. The subject is vast, and I can examine only one part of it. But it is I think a central part. People mean different things by ‘the experience of freedom’. Spinoza thinks that freedom is consciousness of necessity, and presumably thinks the same about experience of freedom (Krishnamurti holds a similar view). Some have in mind an experience of liberation— liberation of mind, or perhaps liberation from jail—that has nothing necessarily to do with the experience of agency in performing a particular action. These things are not my present subject.

Cognitive phenomenology vs. sense-feeling phenomenology The experience or phenomenology2 of free agency is more than a matter of sense-feeling experience. This is apparent even when one takes the term ‘sense-feeling’ in the widest possible way to cover not only all exteroceptive sensory experience and all interoceptive (kinaesthetic, proprioceptive, somatosensory) feelings, but also all moods and emotional feelings— all elements of all moods and feelings—that can be taken to be non-cognitive. One way to express this point is to say that the experience of freedom essentially involves cognitive phenomenology in addition to sense-feeling phenomenology. I introduced the term ‘cognitive phenomenology’, at the beginning of Freedom and Belief (1986a), precisely in order to be able to characterize the phenomenology of agency. Some such term seemed necessary because I was working in the analytic-philosophy tradition, in which it was standardly assumed that all phenomenology is wholly a matter of sense-feeling phenomena. Taking the term ‘cognitive phenomenology’ to cover all aspects of phenomenology—including of course conative phenomenology—that go beyond mere sense-feeling phenomenology,3 I said that the book was centrally concerned with what one might call the cognitive phenomenology of freedom . . . with our beliefs, feelings, attitudes, practices, and ways of conceiving or thinking about the world, so far as these involve the notion of freedom. It is concerned with the experience we have of being free agents, and of being truly responsible for what we do in such a way that we can be truly, ultimately, sans phrase deserving of (moral) praise and blame. It considers the causes, the character, and the consequences of this experience. (1986a: v)4 Many analytic philosophers still reject the idea that there is any such thing as cognitive phenomenology, even when they allow that there is such a thing as phenomenology at all (some, madly, do not). But one has to allow that the phenomenology of free agency essentially involves cognitive (sc non-sense-feeling) phenomenology in addition to sense-feeling phenomenology as soon as one allows that there is such a thing as the phenomenology of free agency at all. It’s enough to consider the most basic characterization of the experience of free agency. It’s the experience, no less, the lived, felt conviction, no less, that it is up to one what one does. Even this simplest expression of the experience of freedom has an irreducibly propositional, hence conceptual, hence cognitive, content. It’s marked here by the word ‘that’, but doesn’t depend on it; the experience is equally well described by saying that it’s experience of its being entirely up to one what one does. It’s an interesting question to what extent non-human animal agents—dogs and dolphins, say—may experience something like this.5 353

Galen Strawson

In the normal case, the human experience of freedom in the living moment of agency or choice flows out of a settled, standing sense of being a free agent, a steady background feeling of being, generically, a free agent. And this persisting experiential set is active not only in the positive experience of its being entirely up to one what one does in a particular case, but also in the negative experience of lack of freedom in action, experience of constraint, constriction, of its not being up to one what to do, in a particular case. In fact the generic sense of oneself as a free agent can be uniquely vivid in such negative cases. One is sharply aware of one’s usual or generic free agency precisely in experiencing oneself as not free in a particular case.6 The experience is also vivid when one struggles not to do something one wants to do. The basic or core experience of up-to-me-ness is already and essentially a matter of cognitive (= non-sense-feeling) phenomenology, although sense-feeling phenomenology is of course also involved, and it carries in the normal adult human case something richer: the conviction that one is fully responsible, and indeed fully morally responsible (if it is a moral matter), for what one does. Plainly this can’t be supposed to be merely a matter of sensory phenomenology. ‘Yes. It’s obvious that the adult human phenomenology of agency involves cognitive (=non-sense-feeling) phenomenology as well as sense-feeling phenomenology’. I agree. It’s extraordinary that so many analytic philosophers have supposed that all phenomenology properly so-called is (and can only be) sensory phenomenology. But facts are what they are, and one does for this reason need to insist, however briefly, on the reality and importance of cognitive phenomenology understood in the current extremely broad sense—even if one ignores the considerable number of analytic philosophers who, directly or indirectly, deny the existence of phenomenology altogether. There is “no banality so banal that no philosopher will deny it” (Antony 2007: 144). This is part of the reason why “philosophy is often a matter of finding a suitable context in which to say the obvious” (Murdoch 1964: 33). The usefulness of the distinction between cognitive phenomenology and sense-feeling phenomenology isn’t put in question by the fact that there may be border disputes. Certainly cognitive phenomenology is not only a matter of linguistic understanding-experience—the experience of understanding what people say when they speak. And this is not just because it’s equally essential to the experience of private conscious thought, as well as the human experience of agency. The point is far wider. All perceptual experience (as opposed to mere sensory experience) also essentially involves cognitive phenomenology, because it essentially involves the deployment of concepts. Suppose you look up and see tables and chairs, a window, a tree. You naturally, automatically, immediately, and involuntarily see them as tables and chairs, a window, a tree. This is not a merely sensory—sense-feeling—occurrence.7

‘Up-to-me-ness’, radical freedom, and ultimate responsibility One of the most interesting questions, when one considers the experience of radical freedom, of being radically free in one’s actions—I’ll call this RF for short—and the associated experience of ultimate responsibility, of being ultimately responsible for one’s actions—I’ll call this UR for short—is whether it is possible for fully self-conscious conceptually sophisticated intentional agents like ourselves to entirely lack any sense of themselves as RF and UR while still genuinely experiencing themselves as agents.8 If you think as I do that freedom of action of a sort that suffices to ground UR is provably impossible,9 you may think that it must be at least be possible for such creatures not to believe that they possess it, because it must at least be possible for such sophisticated agents to believe the truth. In fact, though, this is not obvious, as the story with which I began aims to show. 354

The phenomenology of free agency

We may begin by considering pathological cases—extreme depersonalization, for example, as it occurs in severe clinical depression. In these cases, however, it may be said that people cease to experience themselves as genuine agents at all, even as they continue to be aware of themselves as opening doors and brushing their teeth. In Freedom and Belief I introduce a large cast of different possible agents, in order to test the idea that one could have a genuine sense of ‘up-to-me-ness’ without necessarily having a sense of RF or UR. There is ‘Theoria’ (Chapter 12), also known as the ‘Spectator Subject’, who is profoundly and curiously affectively detached from her own actions (she is related to Meursault, the hero of Camus’s novel L’Etranger). There are the ‘Natural Epictetans’ (Chapter 13), whose experience of choice and action is always utterly effortless, in such a way that they never have any sort of experience of difficult choice or indecision. There is the ‘being of limited conception’, also known as ‘Stolidus’ (§14.8.1). Stolidus can genuinely grasp itself as facing choices, but seems nonetheless to lack any sense of itself as RF or UR, and is perhaps kin to Fido the dog (who faces a vivid choice in Chapter 8), or one of Fido’s even more intelligent future descendants. Then there’s the ‘Genuine Incompatibilist Determinist’, also known as ‘Moira’ (§14.8.3), a supremely sophisticated intentional agent who is fully self-consciously self-aware, who is not affectively detached from her life and actions in the way the Spectator Subject is, and who is utterly convinced of the metaphysical impossibility of RF or UR. Can she, and her sisters Pepromene, Eimarmene, and Chreousa (Freedom and Belief, Appendix G) genuinely manage not to experience themselves as RF, when facing choices like the one described in §1—or indeed far more agonizing and consequential choices? (If Moira agrees to submit to ten years of torture—torture of a kind that she knows she will bitterly regret submitting to as soon as it starts (it leaves no time for moral self-congratulation)—she will save ten other people from the same fate.) Well, that is the question. Kant’s views about the phenomenology of agency are of great interest in this connection. He holds that self-conscious agentive creatures like ourselves (‘rational beings’) can’t help thinking of themselves as RF and UR, and that the reason this is so is that they experience themselves as subject to the moral law. Experience of moral obligation is fundamental to— essentially constitutive of—experience of RF and UR, on Kant’s view. “The moral law”, he says, “is the only condition under which freedom can be known” (Kant 1788: 4 n). Kant not only claims that consciousness of the moral law is necessary for knowledge of RF and UR. He also claims that it’s sufficient: for “if there were no freedom, the moral law would never have been encountered in us” (ibid.). Experience—awareness—of moral obligation is in his terms the ratio cognoscendi of freedom, the means by which we know we have RF. The concept of the freedom of the will10 does not precede the consciousness of the moral law in us but is deduced [i.e. derived by us] from [our experience of ] the determinability of our will (Willkür) by this law as an unconditional command. (Kant 1793: 45 n.; cf. 1788: 30; 1785: 95/447; 1781–1787: Bxxxii–iii) Is this true? Kant is plainly right that experience of moral obligation is deeply bound up with experience of freedom in the normal human case. I don’t, however, think that experience of radical freedom necessarily involves experience of moral obligation. Consider Zibidi, a solitary, highly intelligent, thoroughly rational, fully self-conscious, purposive intentional agent from Aldebaran. It seems that Zibidi can have experiences of such a kind that it can form the idea of its own RF, and perhaps can’t help forming this idea, without having any distinctively moral sense at all. Zibidi regularly has vivid experience of 355

Galen Strawson

being able to do only one of two overwhelmingly desirable but incompatible things, and of deliberating furiously about which to do—experience of precisely the kind that seems likely to give rise to the characteristic absolute sense of being able to choose and act freely— although it has not acquired any trace of any specifically moral habit of thought. If one considers the matter generally for a moment, and independently of Kant, I think it becomes clear that a grasp of moral notions isn’t strictly necessary for having the idea of oneself as RF—even if it’s sufficient. It remains true that the human phenomenology of agency is deeply moralized, deeply penetrated by moral beliefs, emotions, attitudes, and practices— not only in the case that principally concerns me here, the case of one’s experience of one’s own agency, but also and equally in the case of one’s experience of other people’s agency.

Compatibilist and incompatibilist elements in the experience of freedom 1 Theories of the metaphysics of free agency divide naturally into compatibilist and incompatibilist. Compatibilists think that genuine free agency is wholly compatible with determinism. Incompatibilists deny this; they think that the truth of determinism renders genuine free agency impossible (whether or not the falsity of determinism can help). And some incompatibilists are libertarians; they think that determinism is false, and that we are genuine free agents, and indeed that we have RF and UR. This theoretical division is worth mentioning because it has a counterpart in the phenomenology of freedom of action. There are important elements in the thoughts and feelings that constitute the phenomenology of agency that may be said to be compatibilist in character and implication in the following sense: they don’t necessarily involve any phenomenologically live sense of RF or UR of a sort that is (by common consent) incompatible with the truth of determinism. I’ll say that they are compatibilist*. So too there are elements in the phenomenology of agency that may be said to be incompatibilist and more specifically indeed libertarian in character in the following sense: they do involve some phenomenologically live sense of RF or UR. I’ll say that they are libertarian*. Philosophers and psychologists have conducted a number of experiments designed to establish whether people in different cultures are more naturally compatibilist in their conception of freedom or more naturally incompatibilist or libertarian. They’ve had mixed results, and this, no doubt, is because different experimental set-ups, different questionnaires, and so on elicit different aspects of our experience of freedom, some of which are, precisely, compatibilist*, and some of which are incompatibilist* or libertarian*. I’ll consider some of these shortly. First, though, I think it’s worth conducting a small thought experiment. It consists in applying the belief in determinism constantly and unremittingly to the present course of one’s life. One does one’s best, for a minute or two, to think rapidly of every small action one performs or movement one makes as determined—as not ultimately determined by oneself (perhaps as ultimately determined by things that happened long ago). Every time one finds the thought slipping away one refreshes it and reruns it. This (I propose) is likely to have the effect of erasing any sense of the presence of a freely deciding and acting ‘I’ in one’s thoughts. For—so it appears—there is simply no role for such an ‘I’ or self to play. It may be strangely, faintly depressing, or it may give rise to a curious, floating feeling, a feeling not so much of impotence as of radical uninvolvement in the passing show of one’s own psychophysical being. I take this to indicate a respect in which one’s natural pre-theoretical sense of agency has a strongly incompatibilist* and indeed 356

The phenomenology of free agency

libertarian* cast, even though it also has strong compatibilist* elements. One naturally and unreflectively conceives of oneself as possessing RF (and indeed UR) of a kind that is simply incompatible with the truth of determinism. One’s everyday ‘I’, with its standard everyday sense of its RF, fades or disappears in the thought experiment because such an ‘I’ is simply not possible on the terms of the thought experiment. And this sense of ‘I’ is central to the human phenomenology of agency: the sense of self and the experience of agency are inextricably entangled.11

Compatibilist and incompatibilist elements in the experience of freedom 2 In the last section I proposed that there are respects in which our natural sense of self and of freedom is profoundly libertarian*, i.e. deeply imbricated with a sense of RF and UR. I also proposed that there are respects in which it is entirely compatibilist*: there are aspects of our general sense of ourselves as free agents that don’t seem to be put in question in any way by the fact that RF and UR are impossible if determinism is true. But what is perhaps most striking is the way in which natural compatibilist* components of our sense of free agency seem in practice to underlie or underwrite the incompatibilist* and libertarian* components (the belief in RF or UR) even though compatibilism is incompatible with belief in RF. A fully naturalistic explanation of our deep sense of RF must connect it tightly with our primordial sense, massively and incessantly confirmed since infancy, of our ability to do what we want to do in order to (try to) get what we want, by performing a vast variety of actions, great and small, walking where we want, making ourselves understood, picking up this and putting down that. We pass our days in more or less continual and almost entirely successful self-directing intentional activity, and we know it. Most of these actions are routine or trivial, more or less thoughtlessly performed, but this doesn’t diminish the importance of the experience of their performance as a source of the sense of RF—radical self- determinability—that we ordinarily have. Even if we don’t always achieve our aims, when we act, we almost always perform a movement of the kind we intended to perform, and in that vital sense (vital for the sense of RF, of radically self-determining self-control) we are almost entirely successful in our action. This experience is central to our natural compatibilism*. It gives rise to a sense of freedom to act, of complete self-control, of responsibility in self-directedness, that is compatibilistically unexceptionable and completely untouched by arguments against RF based on the impossibility of UR. And yet it is—I suggest—precisely this compatibilistically unexceptionable sense of freedom and efficacy that is one of the fundamental bases of the growth in us of the compatibilistically impermissible or libertarian* sense of RF. To observe a child of two fully in control of its limbs, doing exactly what it wants to do with them, and to this extent fully free to act in the wholly compatibilist sense of the phrase, and to realize that it is precisely such unremitting experience of self-control that is the deepest foundation of our naturally incompatibilistic* or libertarian* sense of RF, ultimate-responsibility-entailing self-determination, is to understand one of the most important facts about the genesis and power of our ordinary libertarian* sense of freedom (and of self ). One reason why we advance from the compatibilist* to the incompatibilist* or libertarian* sense of freedom is perhaps negative. Ignorant of the causes of our desires, we don’t normally experience our character, desires, or pro-attitudes as determined in us in any way at all, let alone in any objectionable way; as Spinoza remarked. We don’t think back behind ourselves as we now find ourselves. And even if a desire is experienced in its importunacy 357

Galen Strawson

as somehow foreign, alien, imposing itself from outside the self, as it were, this probably only serves, by providing a contrast, to strengthen our general sense that our desires and proattitudes are not determined in us. For if a desire or pro-attitude is experienced as imposing itself, then there must be some other pro-attitude in the light of which the first is experienced as imposing itself, and the second one will presumably not also be experienced as an imposition. It will presumably be a pro-attitude one ‘identifies’ with and apprehends as part of oneself, and acquiesces in. A very great deal is locked up in this acquiescence. For although it’s highly unlikely to involve any explicit sense that one has been in any way actively self-determining as to character, it does nevertheless seem to involve an implicit sense that one is, generally, somehow in control of and answerable for how one is; even, perhaps, for those aspects of one’s character that one doesn’t particularly like. As for those pro-attitudes and aspects of one’s character that are welcome to one, it’s as if the following ghostly subjunctive conditional lurks in one’s attitude to them: if per impossibile I were to be (had been) able to choose my character, then these are the features I would choose (would have chosen). This, I suggest, contributes importantly to the sense of responsibility for themselves (and hence for their agency) that most people have, more or less obscurely, more or less constantly. It’s hardly surprising that the ghostly subjunctive conditional confirms the central acceptable status quo. For the ‘I’ that features in the conditional is in fact constituted, considered as something with pro-attitudes that imagines choosing its pro-attitudes, by the very proattitudes that it imagines choosing. I think that this implicit sense of being somehow self-determined in respect of one’s agentive being (it’s strictly implicit—it’s really just a Spinozan shadow, a lack of any sense of not being self-determined) is one of the deep reasons why we have (or don’t fail to have) a libertarian* sense of RF. But I don’t think it’s the principal reason. The principal reason was given in the first section; it concerns the nature of our experience of choice in particular cases. It’s simply that we are, in the most ordinary situations of choice, unable not to think that we will be truly or absolutely responsible for our choice, whatever we choose. And here our natural thought may be expressed as follows: even if my character is indeed just something given (a product of heredity and environment, or whatever), I’m still now able to choose (and hence act) completely freely and truly responsibly, given how I now am and what I now know; and this is so whatever else is the case—determinism or no determinism. This is what we find, I think, when we consider the case of the choice between the cake and the donation to Famine Relief. The case is relatively dramatic, as remarked in the first section; but choices of this general type are common. They occur frequently in our everyday lives, and seem to prove conclusively that RF is both possible and real. The argument that RF is impossible seems powerless when faced with this conviction, which is reinforced by the point just considered, according to which something in itself negative—the absence of any general sense that our desires, pro-attitudes, character, and so on are not ultimately self-determined—is implicitly taken as equivalent to some sort of positive self-determination. We certainly don’t ordinarily suppose that we’ve gone through some sort of active process of self-determination at some particular past time. Nevertheless, it seems accurate to say that we do unreflectively experience ourselves rather as we would experience ourselves if we did believe that we had engaged in some such activity of self-determination. This, I propose, is one of the foundation stones of the general human phenomenology of free agency. There are many complexities here, but the main features of the development of our libertarian* sense of RF and UR out of our unremitting and compatibilistically speaking 358

The phenomenology of free agency

unexceptionable sense of complete self-control may be summarized as follows. (1) We tend to think that we have a will (a power of decision) distinct from all our particular motives. (2) In all ordinary situations of choice, we think that we’re absolutely free to choose whatever else is the case (even if determinism is true, for example), and are so just because of the fact of our full appreciation of our situation. (3) In some vague and unexamined fashion, we tend to think of ourselves as in some manner responsible for, answerable for, how we are (see further Strawson 1986a: ch. 6). All these aspects of the sense of RF or UR directly concern only one’s experience of oneself and one’s own agency. It’s one’s commitment to belief in one’s own radical efficacy, control, self-determination, and total responsibility (in normally unconstrained circumstances), rather than one’s commitment to holding others responsible and treating them as proper objects of reactive attitudes, that is primarily unrenounceable. For what on earth is one to think that one is, or is doing, if one thinks one is not and cannot really be responsible at all for what one does? Here I disagree with those who follow P. F. Strawson in thinking that our overall experience of RF and UR is grounded primarily in our experience of other people, and in particular in our experience of other people as being proper objects of the ‘reactive attitudes’, rather than, more simply, in our experience of our own agency. On Strawson senior’s view there’s a powerful and essential element of intersubjectivity in the constitution of our overall experience of agency; and it is surely true that our experience of others as agents plays a large part in our overall phenomenology of agency. I think nevertheless, and once again, that the deepest source of one’s sense of RF and UR lies in one’s experience of oneself and one’s own agency, one’s core sense of oneself as a self-determining planner and performer of action, someone who can create things, make a sacrifice, do a misdeed. Obviously one’s sense of oneself as agent and one’s sense of others as agents coevolve. Even so I think that one’s experience of one’s own agency is primary. This, again, is not to deny that one’s experience of oneself is deeply determined by one’s interaction with others (one’s awareness of others’ positive and negative attitudes to one’s actions is also crucial). It’s simply (1) to consider two things that develop in us in the course of our social development—our sense of ourselves as truly responsible and our sense of others as truly responsible, (2) to claim that the nature and causes of these two things can profitably be distinguished, and (3) to claim that the former is more fundamental than the latter, so far as our general commitment to belief in RF is concerned.12 The ordinary human phenomenology of agency, free agency, is in any event fundamental to the ordinary human phenomenology of self. It’s utterly fundamental to most people’s basic sense of self, of what they are, of what it is to be a person. There is also, however, vast and extraordinary variation, both when one considers different people and when one considers the same person at different times. And the notion of ordinariness is to that extent somewhat fragile. One of the more striking interpersonal variations has to do with the experience of reasoning, thought, and judgment: of what it is to ponder, try to puzzle something out, come up with new ideas. When I turn my mind to some issue, I bring it to presence in conscious thought, usually in silent inner speech, and, for the rest, simply wait for something to happen—pop up, spring to mind in the tumbling conscious process (it is not, for me, phenomenologically, any sort of flow). Something strikes me; it occurs to me that p, dawns on me, suddenly hits me. At no time do I have any trace of what Wegner calls the “emotion of authorship” with respect to what actually comes up (Wegner 2002: 325–326). It’s as Merleau-Ponty says (he is describing writing a book): “I struggle blindly on until, 359

Galen Strawson

miraculously, thoughts and words become organized by themselves” (Merleau-Ponty 1945: 429). It’s not at all a matter of agency, phenomenologically speaking. There is agency in the business of pushing my mind back to the issue when it starts to wander, but none of the actual occurrence— emergence —of content carries any trace of a feeling of agency. Others do claim to have such an experience, and although I can’t imagine what it might be like, I’m happy to accept that it does occur. “The tendency to attribute control to self is a personality trait”, as Wegner also says, and is much stronger in some than others.13 I’m in the Rimbaud camp: “It’s false to say ‘I think’: one ought to say ‘it thinks me’ ”; “I am an other . . .. This is obvious to me: I am a spectator at the unfolding of my thought; I watch it, I listen to it” (Rimbaud 1871: 251).

Related topics See Chapters 19 (Horgan and Nida-Rümelin), 13 (on Merleau-Ponty), 21 (Hanna), and 26 (Figal).

Notes

360

The phenomenology of free agency

References Antony, L. (2007) “Everybody Has Got It: A Defense of Non-Reductive Materialism,” in B. McLaughlin and J. Cohen (eds.), Contemporary Debates in Philosophy of Mind, Oxford: Blackwell. pp. 143–159. Dennett, D. (1984) Elbow Room: The Varieties of Free Will Worth Wanting, Cambridge, MA: MIT Press. Kant, I. (1788/1956) Critique of Practical Reason, trans. by Lewis White Beck, Indianapolis, IN: Bobbs-Merrill. Kant, I. (1785/1960) Fundamental Principles of the Metaphysic of Morals, trans. by T.K. Abbott 1960, Indianapolis, IN: Bobbs-Merrill. Kant, I. (1793/1960) Religion Within the Limits of Reason Alone, trans. by T.M. Greene and H.H. Hudson, New York: Harper and Row. Merleau-Ponty, M. (1945/1962) The Phenomenology of Perception, trans. by C. Smith, London: Routledge and Kegan Paul. Murdoch, I. (1964/1970) “The Idea of Perfection,” in I. Murdoch (ed.) (1979), The Sovereignty of Good, London: Routledge and Kegan Paul, pp. 1–45. Rimbaud, A. (1871) “Letter to Georges Izambard, 13 May 1871; letter to Paul Demeny, 15 May 1871,” in A. Rimbaud (ed.) (1972), Oeuvres Completes, Paris: Gallimard, pp. 251–252. Strawson, G. (1986a) Freedom and Belief, Oxford: Clarendon Press. Strawson, G. (1986b) “On the Inevitability of Freedom (from a Compatibilist Point of View),” in G. Strawson (ed.) (2008), Real Materialism and Other Essays, Oxford: Oxford University Press, pp. 307–318. Strawson, G. (1994) “The Impossibility of Moral Responsibility,” Philosophical Studies 75, 5–24. Strawson, G. (2003) “Mental Ballistics: The Involuntariness of Spontaneity,” in G. Strawson (2008), Real Materialism and Other Essays. Strawson, G. (2011) “Cognitive Phenomenology: Real Life,” in T. Bayne and M. Montague (eds.), Cognitive Phenomenology, Oxford: Oxford University Press, pp. 285–325. Wegner, D.M. (2002) The Illusion of Conscious Will, Cambridge, MA: MIT Press.

361

24 THE PHENOMENOLOGY OF RATIONAL AGENCY Roberta De Monticelli

All life is position-taking. (Husserl 1911/2002: 290)

The state of the art Phenomenology is experiencing a renaissance thanks to its powerful contributions to the embodied/enactive approach to consciousness and cognition, which has become a prominent paradigm in contemporary consciousness studies. The embodied-enactive perspective puts the perceiving subject back into the world, stressing the dynamic reciprocity between embodied agents and the environments with which they interact (Varela et al. 1991; Gallagher 2005; Hanna and Maiese 2009; Colombetti 2011; Bower and Gallagher 2013). Nevertheless, contemporary phenomenology still lacks a satisfactory overall account of normativity and rationality, up to the standards set by classic works in phenomenology (see this Handbook, Part A). This contribution aims to bridge the gap between work on the embodied mind and rational agency or personhood. The first part of the present contribution addresses the state of the art in contemporary debates. The second offers some relatively original developments toward a full-fledged phenomenology of rational agency. More specifically, the proposed theory of acts should be read as a genetic phenomenology of embodied and individualized personhood. For we are probably born to become rational agents, more or less reasonable, accountable, and morally sensible persons, but we are definitely not born rational (or responsible) agents, capable of giving reasons for our actions. Still less are we born “pure” or disembodied moral agents.

Phenomenology A short clarification about how to understand the term “phenomenology” is in order here, especially as the ambiguity of this technical word is bound up with that of another crucial term, “intentionality.” Intentionality is widely understood as “a specific property which, if instantiated, makes minds of or about objects and facts” (Salice 2018: 604). This aboutness which distinguishes conscious mental states from physical states is undoubtedly what 362

The phenomenology of rational agency

Brentano had in mind when he first introduced intentionality as the fundamental concept of empirical psychology in 1874 (Brentano 1874/1995). “Aboutness” remains a core feature of the richer Husserlian notion, although not its only feature, as developed by Husserl in the Fifth Logical Investigation and taken up by most early phenomenologists. The additional core feature of Husserl’s enriched notion of intentionality is positionality or position-taking (Stellungnehmen) and concerns the subjective (or “noetic”) pole of an intentional relation. By contrast, aboutness focuses on the objective (or “noematic”) pole of intentionality, or what the noesis is about. We shall unpack the concept of positionality in the course of this contribution, since it is due to this second core feature that a phenomenological account of consciousness and action, as opposed to a psychological one, is already an account of reason. Positionality, in fact, accounts for the normativity to which our consciousness is subject even in basic perceptual and emotional experiences, as we shall see in full detail. This suggests an initial gloss for the term “phenomenology” as used here, albeit a negative one: it does not mean psychology, if psychology is about mental facts. Phenomenology is a method of philosophy. Adopting the phenomenological stance toward any object requires clarifying how that object appears from an appropriate intentional, that is, first-personal perspective, e.g., a perceptual one, if it is the object of a perception, or an emotionally characterized one, if it is the object of an emotional experience, and so on. A phenomenology of rational agency requires adopting the perspective of an agent intending to act in a certain way and thereby uncovering the factors that determine whether acts count as right or wrong, in a variety of different senses (e.g., as useful, efficacious, convenient, expedient, just, elegant, appropriate). A phenomenologist adopts this perspective by putting herself “ideally” in the place of such an agent. To do that “ideally” means to “bracket” whatever is contingent for an actual subject, e.g., the particular person I am, focusing instead on whatever necessarily pertains to agency as such. As a piece of ideally examined life, a phenomenology of agency, unlike its counterpart in psychology, has to ask, moreover, about the very sources of normativity, in relation to which actions appear as right or wrong. The truncated conception of intentionality as the property of aboutness enjoyed by conscious mental states explains not only the current lack of distinction between psychology (as an empirical study of mental facts) and phenomenology (as an inquiry into the essential features of whatever object can be presented in a direct or intuitive mode, from a first-personal, idealized perspective). It also encourages a narrowing of the sense of “phenomenology” into a specialized part of psychology that deals with the analysis of “phenomenal consciousness,” leaving intentionality for separate treatment. This further restriction, definitely postBrentanian, underlies current usage of the word in the contemporary academic discourse. Its meaning is now shaped by a distinction (or a “gap”) between intentionality as aboutness (i.e., as what makes a mental state a representation of something in the external world) and “qualia” or “phenomenal” consciousness, a distinction that motivates the “hard problem” of consciousness, the idea that consciousness is impenetrable by any functional account distinguishing our conscious minds from functionally well-equipped zombies (Chalmers 1996; for an argument against separating intentionality and phenomenality see Horgan and Tienson 2002). A phenomenology of agency in this further restricted sense would describe what it is like to act or to be active by leaving the objective nature of action and agency entirely out of consideration. That approach could – even if it need not – be a kind of phenomenalism, perfectly compatible with an “eliminative” or a reductive approach to (phenomenal) consciousness, albeit not implying it. “Phenomenology” would then describe the “mere” appearance of agency and not how agency really works. Or else, it could endorse a dualism of sorts, severing any ontological bond between the outer and inner world, objectivity and 363

Roberta De Monticelli

subjectivity, mind and body. Actions as “lived” would certainly “appear” as steered and even constituted by what an agent “intends” to do, but as seen objectively, from a third-person perspective, they would be nothing but natural events (Davidson 2001). Yet what an agent intends to do (in the usual sense of doing something on purpose) will tell us in most cases which action is being performed (e.g., bribing somebody versus repaying a debt). True enough, intentionality (as aboutness and positionality) does not imply “intending” in the usual sense: a perception has intentionality without having a purpose and even actions are not done on purpose sometimes – such as stepping on somebody’s foot. Yet voluntariness seems to be essential to acting as such: if you unwillingly step on somebody’s foot, you were actually trying to do something else, e.g., going your way. Such that actions are not separable from intentions (even though intentions can be left unrealized). Indeed, identifying intended actions is necessary for our first being able to ascertain one of the most striking, dramatic (and philosophically interesting) phenomena of human agency, namely, the discrepancy between what we intend to do and what we actually bring about, whether as individual agents or collectively. To sum up, what I will develop here is a full-fledged phenomenology of (rational) agency, distinct from both a psychology of agency and a mere description of agential phenomenal consciousness. A full-fledged phenomenology of agency proceeds from the bottom up by capturing the essential features of voluntariness and its normative constraints beginning from the most basic phenomena of conscious life, like perception and emotion, and thereby bridging the explanatory gap between embodied consciousness and personhood, i.e., rational agency.

Rational agency A rational agent is an agent capable of acting based on reasons, including values of all sorts, e.g., hedonic, vital, economic, moral, legal, political, epistemic, aesthetic, and religious ones. Indeed, all sorts of values can be reasons for action. Rational agency isn’t restricted only to “instrumental” or strategic rationality and valuebased rationality. That is, it includes more than – in Kantian terms – conditional and unconditional reasons for acting, economic calculation and commitment to ideals or moral duty. It also includes making things with words, thereby complying (or not) with various syntactic, semantic, pragmatic constraints; making goods with things, such as building houses and artifacts with stones or wood; making sense of facts, whether explaining them or sublimating them in the light of art; and so on. Or, at least, there are no obvious reasons for resisting such a wide conception of the domain of rational agency. This domain is, in fact, the one outlined by Husserl in his Prolegomena, his attempt to inquire into the sources of normativity for all the “practical” and “normative” disciplines, from logics to ethical, legal, political theories, and from aesthetics and the theories of the arts to all possible technologies (Husserl 1900–1901). As soon as we recognize a plurality of (spheres of ) values in view of which people act, we shall have to admit all sorts of corresponding norms by which their deeds show up as right or wrong, in many different ways, e.g., as a good piece of reasoning, apt professional conduct, a morally justified deed, correct city planning, as well as the opposite of all these. In fact, if values are reasons for action, and if there is, as we shall argue, objectivity and corrigibility in value experience, grounded largely in emotional experience, then a full-fledged phenomenological approach to rational agency promotes a redefinition of reason, questioning the traditional opposition between emotion and reason while integrating 364

The phenomenology of rational agency

emotional sensibility as a part, and even a fundamental part, of rationality (Hanna and Maiese 2009: ch. 5; De Monticelli 2019). Last but not least, we could not be rational agents unless we were capable of highly irrational actions. Rationality is not only a functionally based disposition but also a willingness to take ownership of one’s actions by giving (good) reasons for them. It involves freedom to violate norms or to reject values of all sorts. Only rational agents can lie, i.e., take advantage of pragmatic constraints, such as the rule by which an assertion counts as the expression of a belief, by violating them. Only rational agents – that is, persons – act based on value judgments and value priorities. As such, only persons can become criminals. Only persons, moreover, can go mad in the psychiatric sense.

Practical intentionality Against this background, having highlighted the variety of ways that we can act on desireindependent reasons (on all sorts of duties, universal or particular, moral or deontological, and professional, and on all sorts of obligations, legal norms, political strategies, personal commitments, or according to all sorts of rules, of etiquette, cultural, epistemic, pragmatic, aesthetic, technical, etc.), one feels the inadequacy of the so-called “Classical Model of Practical Rationality.” That was in fact the mainstream view until the 1980s. According to it, an intentional action is understood as an event which is causally determined by mental states like beliefs and desires or, more generally, by preceding motivational and cognitive states. This model incorporates a very classical form of compatibilism. Voluntary actions are determined by causes, exactly as any other event in nature. The only difference between the former and the latter is the kind of cause. Voluntary actions have psychological causes, such as beliefs and desires, rather than physical or biological causes. This model thus involves an account of free will resembling the standard empiricist account found in such historically disparate figures as Locke and Davidson (Davidson 1971, 2001). Prima facie, provided one can distinguish inner compulsion and external constraint from psychological determination, it would seem there is no harm in describing free or voluntary actions as causally determined by the relevant states of belief and desire, i.e., the sort of “psychological” or mental causes called “motives” or “reasons.” John Searle, who is famously opposed to this model, takes it to be a representation of human rationality as essentially the rationality of apes with some added complexity. Human rationality is distilled to no more than the following sort of happening: you are thirsty, you see a bottle of what you take to be water before you, so you decide to drink the water (Searle 2001). This is definitely a reasonable decision: but would the decision not to drink it, because – say – you choose to quench a child’s thirst instead, be irrational? In recent decades a shift occurred within contemporary action theory away from that simple belief-desire (BD) model to the belief-desire-intention (BDI) model of practical intentionality. Adding to Searle’s skepticism about BD, Michael Bratman showed how much is lost in translation when intentions – that is, the conative states attributed to the agent of intentional actions – are reduced to beliefs and desires (Bratman 1987, 2007). There are at least three features of intentions that cannot be accommodated by the sort of practical syllogism leading you from a desire and a belief to an action as on BD. The first is commitment, i.e., a sort of self-obligation that is revocable if it is only intrapersonal and that is distinct from mere desire. Making a decision imposes a commitment on the decision-maker, generating an “ought” from the decision. The second feature of intentions absent from BD is planning, however vague, about what is needed to reach a goal (Bratman 2014: 15ff.), meeting the 365

Roberta De Monticelli

constraints characteristic of the practically possible, e.g., practical non-contradiction, temporal irreversibility. Importantly, such constraints do not reduce to beliefs and desires (Bratman 2000). The third feature is causal self-referentiality (Searle 1983: 86ff.). This requires that the intended state of affairs be realized by the subject, thus narrowing, as BD fails to do, the range of possible intentions held by an agent. John Austin had already shown in a celebrated paper that you can do something intentionally without doing it on purpose, and that neither sort of case requires deliberate action, thereby showing, among other things, that an intention can be a mere part of a purpose, a means to an end, e.g., when, say, “I needed money for the horse races, so I dipped into the till, intending all the while to put it back as soon as I had collected my winnings.” There my intended theft is just part of a plan whose purpose is not to steal (Austin 1970: 275). With Bratman, we can generalize the idea, by taking intentions in general to be parts of plans, which in turn reflect policies, or standing practical commitments, like the policy of writing at least a page each day (Bratman 2000: 33–34). Another component of the BDI is reflexivity. Before Bratman’s intervention, Harry Frankfurt had identified personhood or rational agency with a reflexive capacity that human agents – as opposed to “wantons” or non-human animals – have of endorsing their motives for action in second-order desires consisting of the will to be, or not be, determined by first-order desires. The first-order desire actually causing the action is thereby an instance of “free” will only if it is endorsed by a second-order desire (Frankfurt 1982). Bratman seems to give reflexivity an even more important role, distinguishing weak reflexivity, i.e., the Frankfurtian capacity to have higher-order attitudes for or against first-order desires, from strong reflexivity, i.e., the same capacity but insofar as it is embedded in a policy and therefore part of a plan organizing future life (Bratman 2001: 103). A BDI-consistent model of this kind nevertheless suffers from at least two shortcomings. First, it grants only a very restricted scope to rational agency. Small children can feel responsible for what they do and be asked to explain or to justify their deeds before they can plan actions or adopt policies. Agents have, in addition, a (more or less adequate) responsiveness to given data and circumstances not yet captured by their plans, policies, or ends. If, for instance, you come across a scoundrel assaulting a girl, you may feel it as your duty – but not exactly as your wish – to rescue her at the risk of your safety: can morality be irrational? The second problem is much more general. It is shared by all “causal” theories of action, according to which an action is voluntary only if it is caused by relevant mental states of an agent. BD and BDI may differ on the sort of states that are ultimately required, as Davidson’s model differs from Searle’s or Bratman’s. But on all of those views the mental state of the agent rather than the agent himself is (causally) “responsible” for the action performed. This description is utterly inadequate to the full-fledged phenomenology of intending to do something. Whether or not my intention of getting up in the morning is reducible to my desire to get up, whether or not it is part of my plan for the day or a policy for my wakeful time to do so, the intention itself is not sufficient to bring about the getting up, unless I myself perform the act, which I may indefinitely put off as I lay in bed all day with a growing feeling of guilt (Hornsby 2004; Searle 2004; for a discussion of agent causation, see Chapter 19 in this volume).

A theory of acts How are we to give account of the role of the acting subject? Can we do better than to characterize an (intentional) action as an action “caused” by a conative state? 366

The phenomenology of rational agency

The notion of a state certainly plays a central role in any broadly naturalistic account of our life, whether mental or non-mental. Contemporary philosophy of mind has by and large adopted this notion in order to account for our mental life. In fact, according to a widely accepted jargon, mental life is a sequence of mental states (e.g., beliefs, desires, emotions, intentions). The nature of a sequence of mental states is generally described either as a stream, having a temporal order of succession, or as a succession of states each of which stands in causal relations with the states preceding and succeeding it. Yet, if mental is merely a succession of mental states, then the mental life of a person is even less than an imitation of ape rationality. It does not obviously set itself apart from the sort of life characteristic of an ant, or even the “life” of a Turing Machine, which may also be defined as a causal succession of states. This suggests a way to define ontological naturalism about persons: it is (or it is based on) the reduction of acts to states.1 Phenomenologically, this description is not at all true to the life of persons. It fails to capture our mental life as we know it “from the inside,” and it fails even more profoundly to capture our grasp of other persons’ meaningful behavior. Far from being a mere flow of states of consciousness (a description fitting dreams), a wakeful life and any span of it looks much more like a series of acts linked together by a relation of motivation that is not obviously causal, as we shall see. Note that this use of “act” is more inclusive than what is usually meant by “action.” Consider an example. A friend comes in. I perceive her and feel joy. This response is an act – and not just a state that I may or may not endorse, according to the classic Kantian picture: for my joy is an appraisal of my friend’s importance to me. This joy might motivate me to stop what I’m doing, run up to her, and hug her. But surely this joy will not motivate me to do those things without my “consent.” If I were to see her while in the midst of a public talk, I would not endorse this desire (one more act), as I would prefer not to interrupt my presentation. To cut a long story short, many mental “states” don’t seem capable of “causing” the following ones without the subject’s endorsement. This endorsement, by which the subject takes a stance concerning the state (either assenting to it or withholding assent), makes an actual motive out of an otherwise merely possible one. Without this endorsement, no possible motive could become an actual one. We can define a possible motive as a motive lacking causal efficacy in the absence of an endorsement, even if an endorsement is no sufficient condition of causal efficacy. This seems to be the essential difference between causality and motivation. (For a Frankfurt-style account of this difference, see Hanna and Maiese 2009: ch. 4.) Endorsing a mental state (or the opposite stance) is an instance of positionality, a “position” being the essential or distinctive feature of an act in a strict sense, conferring “act-uality” on it. Now we can better see, maybe, how the standard notion of intentionality is faulty. Intentionality has not only a first-person perspective but also a first-person actuality. Living as a subject (and experiencing oneself as such), even pre-reflectively, is responding more or less adequately to the surrounding world, which presupposes neither concept mastery nor taking on propositional attitudes. Positionality changes a reaction into a response, and a stimulus into an object, a possible truth-maker for propositional acts (see below).

Kinds of acts So far, I have argued that neglecting positionality, thereby reducing personal acts to mental states, results in a subjectless account of conscious life. Agency is an essential component of subjectivity. The usual opposition between the firstand third-person perspective is better understood in terms of an opposition between the 367

Roberta De Monticelli

engaged and the merely observational perspective, or between being committed and merely having a point of view (Bilgrami 2010: 25). An incomplete theory of intentionality blurs this second and more fundamental opposition. However, we should avoid conceiving of positional acts as a strange form of “mental” action, indulging in a form of Cartesian dualism as if there were some spectral agent or ego over and above the person. We can do that by conceiving positional acts as a special case of acts, involved or presupposed by acts of all the other kinds. On a provisional analysis, surely in need of elaboration, this word, “act,” denotes actions enjoying one or more of these features: (1) punctuality, as in shooting, jumping (in contrast to running), highlighting the present occurrence of an action, as in “caught in the act of . . .”; (2) some sort of self-manifestation, and in particular, self-commitment, either relative to one’s future behavior, to others, or to both. Notice that an implicit reference to the acting subject’s power of initiative (Spiegelberg 1986) is present in both cases (in contrast to the “actions” of a machine), while “acting” in the sense of “playing a role,” e.g., on stage, further extends the idea of a subject and its dispositions being manifested. Both nuances are present, too, in the legal sense of “act” (e.g., a “jobs act”), which refers to the product of legislation (being in this case the equivalent of the Latin participle actum). To sum up, we can subdivide these two overlapping types of act into still further types: 1 2

3 4

Punctual actions (as opposed to temporally extended actions and activities); Actions manifesting attitudes or dispositions, possibly with positive or negative value (e.g., an act of friendship, an act of courage); possibly ritual acts (e.g., of worship, of faith); Speech acts, that is, doing things with words, as with assertions and questions (most of which are also social acts); Social acts, including speech acts like performatives, commissives, and directives (which generate social institutions, reciprocal contractual bonds, roles, and deontic powers, i.e., obligations and rights (Searle 2010)), as well as institutional acts like sentences, laws, government decisions, which have legal and political ramifications.

Classes 3 and 4 are particularly interesting for a theory of rational agency, since they contain most of the desire-independent reasons for action (e.g., pragmatic commitments, selfobligations, legal obligations, commands) acknowledged by philosophers who, like Searle, naturalize intentionality but not rationality. Here positionality as expressed by an act’s illocutionary force changes, to use the Husserlian terminology, the “quality” of the act, if not its “matter.” That is, it enables changes in the kinds of things one can do with words (e.g., make statements, pose questions, venture hypotheses, offer prayers, give commands, make promises).

Making decisions Are there pre-reflective and non-linguistic positional acts? Perceptions, emotions, and even a large class of decisions appear to qualify as such. (Vis-à-vis the list of kinds of act just enumerated, we might assign them the ordinal number 0. For a similar focus on a fundamental category of essentially embodied acts, see Hanna and Maiese 2009). Adopting the Cartesian and Brentanian terminology, we could call them “mental” acts, alongside propositional acts like judgments.2 That language is misleading, however, if what is meant is a sort of “inner”

368

The phenomenology of rational agency

or “mental” action (a curious kind of causally inert action). Gilbert Ryle’s criticism of the very idea of mental actions is compelling: Nobody ever says . . . he has performed five quick and easy acts of the will, and two slow and difficult ones between breakfast and lunch. (Ryle 1949/2000: 64) Here Ryle is making fun of the idea that there is a little inner agent, a “ghost” hidden inside our body. But if we reject the idea of the inner agent and regard a decision as the actual exercise of that ordinary capacity to make decisions that is called “will,” Ryle’s sarcasm loses its bite. We can count decisions. We can even say that some are difficult, others not so much, and so on. Husserl himself warned against the mythology of mental actions.3 Nevertheless, it would be absurd to deny that decisions exist or that we “make” decisions. In fact, a decision is a very specific positional act, one of endorsing a possible motive for action and making it an actual one, as described above. That is the very nature of practical intentionality, as masterfully analyzed by the most insightful phenomenologist concerned with understanding the will, Alexander Pfänder (Pfänder 1911, see Chapters 2 and 3 in this volume). A theory of the will, capable of accounting for a decision’s specific Aktcharakter, as opposed to the mere content (Materie) of an “intention” or a purpose (which may or may not be endorsed and made effective), appears to be virtually absent from discussions within contemporary philosophy of mind, as is the distinction between (positional) act and state (yet see Hanna and Maiese 2009 for an exception).

Exploring the world, learning from experience, doubting: validity claims As has been rightly observed, Husserl’s distinction between the quality and matter of an act bears a certain resemblance to the current distinction between the attitude (or mode) and content of mental states (Gallagher and Zahavi 2008: 115). Yet, there is much more to the act’s quality than there is to a psychological attitude. Quality includes a specific validity claim, by which any positional act falls “under the jurisdiction of reason” (Rechtsprechung der Vernunft) (Husserl 1913/1983: 159) or is subject to possible assessment. This point needs clarification. We remarked that the truncated concept of intentionality current in contemporary philosophy of mind encourages a subjectless view of the ground level of conscious life, failing to take care of its commitments. But how can a passive and unquestioned sequence of perceptions and emotions, without commitments to their veridicality or appropriateness, still be called “experience”? How could one ever discover perceptual or emotional illusions, if perceptions and emotions did not “claim” to be veridical or appropriate, thereby eliciting our commitments? How could experience be fallible, and how could we “learn from experience,” that is correct past experiences by new ones, if not by assessing (or rejecting) the corresponding claims of validity? Positionality is precisely what accounts for the naïve, “unquestioned” claim of validity that perceptions and emotions contain. We may see this point better by contrasting perceptions and emotions with acts of imagination, which essentially do not have any claim of truthfulness (at least not in this basic sense), which do not raise any doubt about their adequacy, and which typically have a “neutralized” positionality (in Husserl’s terminology).

369

Roberta De Monticelli

Ordinary perceptions and emotions, however, are corrigible. A claim of validity can be canceled by a modified position (e.g., in one of perplexity, doubt, rejection). Hence, if by “experience” we mean not just the causal impact of external reality on an organism, but something we can learn from, something which is or can fail to be veridical or appropriate, something which can provide evidence for our judgments of fact and value, then we must take positionality into account. In summary, the difference between acts and states is the difference between experience as evidence for (possible) true statements (of fact or value) and experience as the effect of causal impact of reality on an organism. But the experience through which we explore the world, for example, in the act of looking around or directing our gaze on a particular object, the experience we “learn from,” is of the first kind. It is always more or less adequate and could not be subject to critical doubt without positionality. Even playing baseball, an intentional activity which is not primarily assessable for truth, does presuppose perception and evaluation and their corresponding claims of validity. As a final observation concerning the distinction between acts and states, we may say that states are merely the effects of the world’s causal impact on an organism, whereas (basic) acts are adequate or inadequate responses to reality. Hence, positionality is the foundation of normativity. (For a similar conclusion, tying this foundation of rational normativity to pre-reflectively conscious, essentially embodied “caring,” see Hanna 2015.)

A hierarchy of acts In the preceding sections, we have found persons to be the subjective pole of the intentional relation by discovering the place of agency and reason in experience. We have in a sense de-naturalized intentionality by showing how mental life is subject to the “jurisdiction of reason.” Yet it is highly unlikely that we are born conscious of this subjection, of our fallibility and free will. How, then, do we acquire full-fledged rational agency and come to exercise the capacities that go along with personhood? In the following sections, I shall outline a genetic phenomenology of rational agency, as constituted by and through one’s acts of position-taking. Positionality is to the subject of an intentional relation what the mode of presentation is to its object. It is how the subjective pole of an intentional state is given or made present to itself, i.e., the way persons experience themselves as such, as subjects. From a full-fledged notion of intentionality we learn that, as persons, we come to a full-fledged being only on the basis of those acts through which “we” learn to respond appropriately in the long apprenticeship of reality and value characteristic of the neotenous creatures we are. The following outline sketching the hierarchy of acts will suggest how and why that is so. The basic level of our entire personal life comprises what we may call basic acts containing first-order positions. There are two classes of such basic acts (well intertwined in their actual occurrence): cognitive and emotional, or, more simply, perception and emotion. Basic cognitive acts, or perceptions, are characterized by first-order “doxic” positionality (doxa, Greek for “belief ”). Basic emotional acts, however, feature “axiological” positionality (axios, Greek for “valuable”). Doxic positionality consists in recognizing a perceived thing’s existence. It is a kind of assent or denial, although it is immediate, not reflective (as is, e.g., propositional belief ). It is not a “judgment,” if we mean by that the illocutionary act of stating a propositional content, even if not worded or voiced. A perception can be illusory as such, in its own content. But it could not do so if there were no doxic position, which, by contrast, is absent in

370

The phenomenology of rational agency

acts of imagination or day-dreaming (where veridicality is not an issue). Doxic positionality confers on perceptions their distinctive (defeasible) claim of veridicality. First-order axiological positionality (what Husserl calls Wertnehmung) involves the acknowledgment of the positive or negative salience or value of a given thing or situation. Emotions generally include such positions or “valences” and can be appropriate or inappropriate (imagine, e.g., feeling terror in the presence of a peaceful kitten). The negative axiological stance (including the “flight response”) of the inappropriate experience is, in that case, clearly wrong in some sense. First-order positions are not freely taken. I cannot help but endorse the existence of what I see or touch. I cannot choose to take the opposite axiological stance when I come across an object of fear or horror. This holds even if a thing’s existence turns out, in the further course of experience, to be illusory (something perceived as a living thing turns out to be a scarecrow), or if an object of fear turns out not to warrant a fearful response after all. For this to happen – for the stance taken to be modified retroactively as “crossed off,” a mode of perceptual doubt, or, in short, a new stance – there must be an antecedent stance. When addressing intentionality, Husserl uses the term “Akt” as more or less synonymous with “intentionales Erlebnis.” Yet, to identify act and intentional lived experience would not be entirely satisfactory, because no act can be reduced to the lived or conscious experience of it. Acts – even “mental” acts – can exceed their conscious aspects. Like anything effective, an act in part transcends consciousness. There is more to it than is experienced, as we shall see shortly.

First degree of personhood’s emergence: facing objectivity A person is a subject that “emerges” from a sequence of biological and mental states by virtue of the positional component of perceptions, emotions, and behaviors insofar as these are subject to normativity (being right or wrong). Such basic pre-reflective, norm-driven experience stands out from behavior that is simply adaptive or biologically driven (e.g., for the satisfaction of needs). One might wonder what makes experience “norm driven.” Grasping this is crucial for correctly understanding our notions of emergence and subject. A stream of psychological states through which an animal – say, a dog, a dolphin, a chimpanzee – interacts with its natural and social environment is not sufficient for the constitution of a personal subject. Animals seem to experience reality as what resists their drives and desires or satisfies them, but not as what reveals doxic and axiological positions as wrong or right. We first learn that positions are right or wrong by sharing the habits and norms of the life-community in which we are born. Learning from positional acts, knowing what to do next – these are paradigmatic forms of norm-driven behavior. Although some non-human animals could be subject to perceptual and emotional illusions and maybe capable to correct them, only humans seem to have developed into “normative animals.” As a matter of fact, social learning may be part of reaching adulthood for (some) non-human animals. What seems to be characteristic of humans, though, is cultural or norm-based learning (Tomasello 1999: 37–39), typically requiring joint attention and cooperative dispositions, as a condition to introduce coherence, organization, and order in even the most basic responses to the environment. In that way, “meaningful” structures of behavior gradually emerge from relatively unorganized sequences of reactions to inner and outer stimuli. Reinforcing right responses, discouraging wrong ones, jointly executing right positional acts, and the like are ways in which care-givers and the community provide the

371

Roberta De Monticelli

foundations of an emerging subjectivity. Thus, a form of shared intentionality is necessary for the emergence of the infant as a subject of a motivational chain of acts from a mere flow of states, as Max Scheler first pointed out (Scheler 1923/2008).

Second degree of personhood’s emergence: managing experience But we can go further. We also manage the states arising in this contact with reality. In basic acts we experience reality as objective and experience as well our grasp of it as limited and fallible. But regardless of whether we get things right or wrong, the use we make of prior basic positional acts is, within certain limits, up to us. It is in our power to further expose ourselves to reality, that is, to accept or reject data as actual motives of subsequent life (experiences and actions). This observation leads us to recognize what we may call a second degree of personhood’s emergence. We manage our states by a second class of acts, involving second-order positions, that is, positions that we take relative to basic acts and their correlates (states of affairs). As opposed to basic acts, these “managing acts” are, in a broad sense, free. A first-order stance denying reality to what I perceive to be real is not in my power. Yet, resisting the motivational weight of a perceived fact (or a felt value) is in my power. I can receive a piece of bad news, or learn about a very painful fact, and, above and beyond that, I can let myself be motivated by it. I can also “repress” it, by ignoring it, not allowing it to motivate my further acts, e.g., my emotions, thoughts, decisions, or behavior. I may look away, or “neutralize” the first-order position. With this act, I can manage my experience by regulating my exposure to further experience. To endorse or ignore basic inputs is to take up a second-order position. Second-order positions count as acts that are free in a broad sense. By “in a broad sense” I mean to emphasize a typical feature of this second class of acts. Such acts are neither necessarily nor entirely conscious. We can manage our passivity in the dark, as it were, like when we repress grief, thereby forestalling the possibility of working through that grief, yet without admitting to ourselves that we are doing so. This is, as it were, the “gray area” of spontaneity, where we act without explicitly assuming responsibility for our acts and where most of our life is spent.

Third degree: the emergence of personal identity To better grasp the essential features of the third and final class of acts, we might think of the just-discussed second degree of personhood as the management of one’s passivity. This paradoxical-sounding expression reminds us that experience is never completely “passive.” Otherwise we could not even say that we “grow up” through experience. Yet, the “path” that each of us takes through the world, so to speak, by managing passive motivations and regulating their influence on further experience, need not be a series of choices, conscious or otherwise. Doubtless, what we call personality and character traits manifest themselves in second-order acts. These can always be clarified in retrospect, and alternative possible plots can be brought to consciousness, although that is not necessary. However, by regulating exposure to the onslaught of information we are struck with in basic acts, we undeniably exert a power of some sort, attesting to an efficacy entirely absent in basic acts. We do or do not authorize a given experience to exert its motivational force on us in the further course of experience. Further experience, though, does not necessarily mean further action. By tacitly endorsing or ignoring data and states as motives in our ongoing life, we do not necessarily engage in active 372

The phenomenology of rational agency

or overt behavior. To avoid working through bereavement, for example, is not active behavior. The fully conscious management of motives comprises only a subclass of free acts. Such acts are, in fact, authorizations to proceed, licensing one to make something of the factual and axiological data of a given experience. In that way we give ourselves license to actively behave in such and such a way, for example, to get involved in reading about medicine or philosophy, or, at a more basic level, to run up the hill for the pleasure of it. In either case, we may thus take the first step in acquiring a habit or skill. Our acts shape us. By making something out of the data of a given experience we also make something out of ourselves. These are the acts we may deem free in the strict sense. This is the exact sense in which a phenomenological theory of freedom goes well beyond the kind of compatibilism involved in the Classic Model of Rationality (the BD-model discussed in Section 3). Those acts by which we endorse (or reject) data as motivations or reasons for action are acts that are free in the strict sense. These acts are essentially – even if only to a minimal degree – commitments to one’s future behavior. What characterizes this class in its most exemplary instances is the engagement of one’s future self, which can take the form of obligations we impose on ourselves with respect to ourselves or others. Decisions are paradigmatic instances of the former, promises of the latter. Acts that are free in the strict sense are self-constitutive acts. By endorsing a reason for action, I not only make a commitment involving my future self, I also accept responsibility for what I shall be. In this sense, decisions are paradigmatic instances of self-constitutive acts, even if we might, by further analysis, discover that the essential nature of a decision is better clarified by taking it to be a sort of promise made to oneself. A decision truly engages one’s future self, and, conversely, one bears responsibility for one’s past decisions, especially in so far as one is responsive to other people’s expectations. We may ultimately discover, as Nietzsche first suggested, that personal responsibility was, genealogically speaking, linked to social acts of promising, before becoming the amazing power of self-imposed obligation that we now attribute to our (free) will.

Conclusion. Actuality, subjectivity, and personhood – the individualized rational agent Phenomenology encourages the de-naturalizing of intentionality through the embodiment of rationality. No perplexity needs arise from biology, pace Searle. At the same time, rational agency cannot be just a function of language and social institutions, as though it were nothing “before” them, again pace Searle. Nor does it show up only at the highest level of planning, pace Bratman. Austin is right to hold that “intentionally” does not necessarily mean “on purpose” or “deliberately.” But Austin still did not see that one can act spontaneously without intending what one does and that even non-intending admits degrees. Acting freely, too, admits degrees, pace Frankfurt. “Actual” experience is proto-agency, and the bedrock of personhood. “All human living is position-taking,” says Husserl, though that need not mean taking a side. Stances can be (socially) questioned and corrected, but what makes them correct (or not) is not social agreement, but reality. The consistency of a mind is founded on the minimal adequacy of positional acts. In a sense, our openness to truth is constitutive of personhood. We may sum up the argument outlined above in three claims: (S) A person is a subject of acts. Here the notion of personhood is not presupposed but rather explained by the notion of an act. Our living is not reducible without remainder to “personal” living. Digestion, for instance, does not qualify as an act, while eating and – even more so – sex do. 373

Roberta De Monticelli

(A) To live as a person is to emerge from one’s states by one’s acts. Recapitulate our analysis, without basic acts we cannot stand up as subjects facing (each other and) objects, exploring the world and learning from experience. Only by managing experience do we in a way rise above it as, to a greater or lesser degree, directors of our lives. But only self-constitutive or self-committing acts are sources of personal identity through time. Unlike Boethius, who defined personhood as the individuality of a reasonable nature, contemporary philosophers ignore the problem of individuation. Phenomenology, however, teaches us that rational agency is not only embodied, but highly individuated. (F) A certain subset of acts is a necessary condition for the emergence of an individual personality, namely, the one consisting of free acts. At its most basic level, positionality grounds subjectivity. At its highest level, it shapes a person’s identity, which solidifies in habit. Between those levels, personality or character is manifested. Our acts shape us. Personhood is constituted not only in the Husserlian sense of being experienced as exercising positionality, but also in the sense of being shaped by that exercise. The exercise of positionality is what makes the self.

Related topics See Chapters 2 (on Pfänder and Husserl), 3 (on Pfänder), 7 (on Reiner), 19 (Horgan and Nida-Rümelin), 20 (Smith), 21 (Hanna), and 25 (Drummond).

Notes 1 Mental states are what they are in virtue of the causal and functional role they have in our mental life, independently of how this role is physically implemented. This idea has been fundamental to standard cognitive science, since it was the cornerstone of the project of naturalizing our mental life lying at the heart of this paradigm, dominated by such figures as Quine, Dennett, Fodor, Dretske, and the Churchlands. In a nutshell, the idea is to show that mental states can be properties of the world as studied by the natural sciences. 2 This language has its roots in St Augustine, De Trinitate XI, II, where he speaks of the “act of seeing” that “keeps the sense of the eye in the object seen.” Brentano refers to Descartes’ two classes of mental acts (iudicii, volitiones) as corresponding to his two classes of Akten (Urteile and Akte des Interesses/ Gemütsbewegungen), both based on simple “representations” (Descartes’ ideae). Mental acts involve one of two opposite stances or position takings: acknowledgment/denial (Anerkennung/Leugnung) and like/dislike (Liebe/Hass). Curiously, Brentano does not incorporate these central features of his theory of positionality into his theory of intentionality (cf. Brentano 1874/1995: 198). 3 “In talking of ‘acts’, however, we must steer clear of the word’s original meaning: all thought of activity must be rigidly excluded” (Husserl 1901/2001: 393/102). In a footnote to this passage Husserl approvingly quotes Natorp against a “mythology” of mental actions and operations: “We too reject the ‘mythology of activities’: we define ‘acts’ as intentional experiences, not as mental activities” (Husserl 1900–1901/2001: 354).

References Austin, J. (1970) “Three Ways of Spilling Ink,” in J.O. Urmson and G.J. Warnock (eds.), Philosophical Papers, second edition, Oxford: Clarendon Press. (Reprinted from The Philosophical Review 75/4, 1966, pp. 427–440). Bilgrami, A. (2010) “The Wider Significance of Naturalism – A Genealogical Essay,” in M. De Caro and D. Macarthur (eds.), Naturalism and Normativity, New York: Columbia University Press, pp. 23–54. Bower, M. and Gallagher, S. (2013) “Bodily Affects as Prenoetic Elements in Enactive Perception,” Phenomenology and Mind 4, 78–93. Brentano, F.C. (1874/1995) Psychology from Empirical Standpoint, trans. by A.C. Rancurello, D.B. Terrell and L.L. McAlister, London: Routledge.

374

The phenomenology of rational agency Bratman, M.E. (1987) Intention, Plans, and Practical Reason, Cambridge, MA: Harvard University Press. Bratman, M.E. (2000) “Reflection, Planning, and Temporally Extended Agency,” The Philosophical Review 109, 35–61, in M.E. Bratman (2007, Structures of Agency: Essays, pp. 21–46. Bratman, M.E. (2001) “Two Problems about Human Agency,” Proceedings of the Aristotelian Society 2000–2001, pp. 309–326, in M.E. Bratman (2007), Structures of Agency: Essays, pp. 89–105. Bratman, M.E. (2007) Structures of Agency: Essays, New York: Oxford University Press. Bratman, M.E. (2014) Shared Agency. A Planning Theory of Acting Together, Oxford: Oxford University Press. Chalmers, D. (1996) The Conscious Mind: In Search of a Fundamental Theory, Oxford: Oxford University Press. Colombetti, G. (2011) “Varieties of Pre-Reflective Self-Awareness: Foreground and Background Bodily Feelings in Emotion Experience,” Inquiry 54 (3), 293–313. Davidson, D. (1971) “Agency,” in R. Binkley, R. Bronaugh and A. Marras (eds.), Agent, Action, and Reason, Toronto, OH: University of Toronto Press. (Reprinted in Davidson 2001, pp. 43–62). Davidson, D. (2001) Essays on Actions and Events, Oxford: Oxford University Press. De Monticelli, R. (2019) “Values, Norms, Justification and the Appropriateness of Emotions,” in T. Szanto and H. Landweer (eds.), The Routledge Handbook of Phenomenology of Emotions, London, New York: Routledge, pp. 275–287.Frankfurt, H. (1982) “Freedom of the Will and the Concept of a Person,” in G. Watson (ed.), Free Will, Oxford: Oxford University Press, pp. 81–95. Gallagher, S. (2005) How the Body Shapes the Mind, Oxford: Clarendon Press. Gallagher, S. and Zahavi, D. (2008) The Phenomenological Mind: An Introduction to Philosophy of Mind and Cognitive Science, New York: Routledge. Hanna, R. (2015) Cognition, Content, and the A Priori, Oxford: Oxford University Press. Hanna, R. and Maiese, M. (2009) Embodied Minds in Action, Oxford: Oxford University Press. Horgan, T. and Tienson, J. (2002) “The Phenomenology of Intentionality and the Intentionality of Phenomenology”, in D. Chalmers (ed.), Philosophy of Mind. Classical and Contemporary Readings, Oxford: Oxford University Press, pp. 520–533. Hornsby, J. (2004) “Agency and Action,” in J. Hyman and H. Steward, H. (eds.), Agency and Action, Cambridge, MA: Cambridge University Press, pp. 1–23. Husserl, E. (1900–1901/2001) Logical Investigations, English trans. by J.N. Findlay, London and New York: Routledge. Husserl, E. (1911/2002) “Philosophy as a Rigorous Science,” trans. by M. Brainard, in The New Yearbook for Phenomenology and Phenomenological Philosophy 2, London/New York: Routledge, pp. 249–295. Husserl, E. (1913/1983) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy I. Introduction to a Pure Phenomenology, English trans. F. Kersten, in E. Husserl (1983), Collected Works. Vol. II, The Hague, Boston, Lancaster: Martinus Nijhoff. Pfänder, A. (1911) “Motive und Motivation,” in A. Pfänder (ed.), Münchener Philosophische Abhandlungen: Theodor Lipps zu seinem sechzigsten Geburtstag gewidmet von früheren Schülern. Leipzig: J.A. Barth, pp. 163–195. Ryle, G. (1949/2000) The Concept of Mind, London: Penguin Modern Classics. Salice, A. (2018) “Practical Intentionality. From Brentano to the Phenomenology of the Munich and Göttingen Circles,” in D. Zahavi (ed.), Oxford Handbook of the History of Phenomenology, Oxford: Oxford University Press, pp. 604–622. Scheler, M. (1923/2008) The Nature of Sympathy, trans. by P. Heath and W. Stark, New Brunswick and London: Transaction Publishers. Searle, J. (1983) Intentionality: An Essay in the Philosophy of Mind, Cambridge, MA: Cambridge University Press. Searle, J. (2001) Rationality in Action, Cambridge, MA: MIT Press. Searle, J. (2004) Liberté et Neurobiologie, Paris: Grasset et Fasquelle. Searle, J. (2010) Making the Social World: The Structure of Human Civilization, Oxford: Oxford University Press. Spiegelberg, H. (ed.) (1986) “Initiating: A Phenomenological Analysis,” in Steppingstones – Toward an Ethics for Fellow-Existers, Dordrecht: Nijhoff. Tomasello, M. (1999) The Cultural Origins of Human Cognition, Cambridge, MA: MIT Press. Varela, F., Thompson, E. and Rosch, E. (1991) The Embodied Mind: Cognitive Science and Human Experience, Cambridge, MA: MIT Press.

375

25 ACTING, CHOOSING, AND DELIBERATING John J. Drummond

Tom has Parkinson’s disease and experiences tremors of the bodily extremities. Susan blushes when embarrassed. Are these bodily events actions?1 It seems not, since they are something that happens to people, something they undergo rather than something they do. These bodily events are neither initiated by the person nor under the person’s control, but when we do something, we think we initiate and control the bodily events. Hence, it seems that an action is, minimally, a bodily performance that is (i) initiated by the agent and (ii) under the agent’s control. Moreover, it also seems to be part of the sense of an initiated action that it is aimed at some end, at realizing an (apparently) good (desirable) state of affairs. Hence, it seems also that for a physical performance to be an action it must be (iii) aimed at an end. So, for example, when I unthinkingly reach for my glass to take a sip of water or when I take a break from writing and stand before an open refrigerator trying to remember what I came to get, I initiate and control the action, and the action is aimed at the end of quenching my thirst or satisfying my hunger, even if I am not explicitly aware of the end while reaching for the glass or walking to the refrigerator. With these preliminaries in mind, this chapter first distinguishes three senses of “action.” The second section turns to the most complete sense of action—chosen actions—and the third section to the question of deliberation and its relation to choosing. I shall frame the discussion around some Husserlian distinctions, but I aim neither to expose nor interpret his views. My aim is, rather, to sketch a position that allows for the multiple possibilities that arise when we consider actions.

Acting An action satisfying the three aforementioned conditions—I shall call it a “voluntary action,” although in a slightly different sense from Aristotle [Nic. Ethics: 1111a22–25, Aristotle 1984])—grasps a present situation as bad, deficient, unsatisfying, or unacceptable in some respect and evaluates an envisioned situation as good and as realizable by a determinate bodily performance. The envisioned good may be a condition or state in myself, such as being an honest person, and it is realized in performing the action—telling the truth—itself. Or the envisioned good may be some product or external state of affairs that is realized as an effect of my action, say, making an omelet. Husserl calls the (implicit or explicit) volitional intention 376

Acting, choosing, and deliberating

that initiates and governs a voluntary action the fiat, the “let it be done” (Husserl 1988: 107). The transition from the evaluation of the envisioned end as (apparently) good to the fiat is mediated by desire. The envisioned good disclosed by the evaluation is taken as desirable, and the fiat directs us to that desirable good as something to be realized through a bodily performance. The fiat, in other words, emptily intends the envisioned, realizable good and is inseparable from the performance that realizes (or attempts to realize) it (Melle 1997: 180). At every moment of the action, we have a partial fulfillment of the fiat, and the action continues just so long as the fiat is operative. As Ullrich Melle (1997: 181) puts it, “during an action we have an intention of the will that constantly maintains itself as an empty intention that simultaneously constantly acquires the fullness of satisfaction through a constantly creative positing.” The phenomenological notion of the fiat, the empty volitional intention as fulfilled in a temporally extended action, recognizes that the action explained is part of a more comprehensive activity that unfolds over time. Michael Thompson’s claim that the fundamental mode of explaining action is “naïve action explanation” (Thompson 2008: 86) involves the same recognition. A naïve action explanation explains breaking eggs as follows: “I am breaking eggs because I am making an omelet” (Thompson 2008: 85). Thompson contrasts such naïve action explanations with “sophisticated” (philosophical) explanations that appeal to wants or desires as the explanans of a voluntary action (Thompson 2008: 86–87), for example, “I am breaking eggs because I want to make an omelet.” Thompson does not deny that sophisticated explanations can be given for actions, but he thinks them secondary to and dependent upon naïve explanations. A difficulty in Thompson’s account from a phenomenological perspective is that phenomenology is concerned not with explanation but with the description of the intentional structure of action, and naïve action explanations mask that structure. The fiat, I have said, is founded upon an evaluative sense, which is, in turn, founded upon the non-axiological properties of the valued object (e.g., a thing, situation, or event) (see, among others, Husserl 1970b: 636–639, 1988, 252; Stein 1989: 100–101; Pfänder 1913: 340; Brentano 1995: 60–61; Drummond 2013: 252). In the case of evaluation, an emotion—the evaluating experience—is motivated by the thing’s non-axiological properties and the ascription of the value-attribute to the thing is justified by those same properties. Fear, for example, apprehends the dog as dangerous. Why? Because the dog is charging me directly while baring its teeth and growling. Those non-axiological features are the (motivating) reasons for my experiencing fear and the ( justifying) reasons for my sense that the dog is dangerous. Similarly, in the case of action, and precisely because the fiat is motivated by a desire engendered by the evaluation, the search for the reasons for the action would move “downward” toward the intentionally motivating layers of sense. So, contra Thompson, the naïve explanation “I am breaking the eggs because I am making an omelet” includes as part of its meaning that I want an omelet. The naïve explanation, while parsimonious, gains its explanatory force from the fact that I desire an omelet. Intentional motivation, at least initially, moves “upward” to more complex meanings: non-axiological properties found an affective, evaluative meaning rooted in those properties, and the value-attributes disclosed by the emotion found, with the mediation of desire, the fiat and the actions undertaken to fulfill it. Many have inferred from the relation between evaluation and voluntary action that the feeling- and emotion-experiences in which we value objects contain in themselves inclinations and motivations to act in certain ways and thereby serve to explain and justify our actions, at least in part. It is an undeniable, but contingent, fact that emotions motivate 377

John J. Drummond

desires and actions; it is not the case that desire with its attendant tendency to action belongs essentially to every emotion-type or to every emotional episode. The view that emotions incline us to act presupposes that the desiderative belongs to the emotive. Linda Zagzebski (2003: 116), for example, speaks of the affective dimension of an emotional experience—the bodily feelings that belong to the emotion—as “pushy.” Her basic idea is that pleasurable feeling states “push” us toward actions that will maintain or increase the pleasure, while painful feeling states “push” us toward actions that will lessen or end the pain. While it is clear that desire presupposes an emotive evaluation of a (apparent) good, the emotion, as I shall argue, does not necessarily involve desire and its “pushiness.” For Zagzebski, it is the subjective condition that motivates action, but I agree with many of the early phenomenologists that it is the value disclosed in the emotional experience—the (apparent) good or (apparent) bad—that “pulls” us toward action. Robert Roberts too believes that emotions in general, even if not universally, incline us to act. Roberts views emotions as “concern-based construals” (Roberts 2003: 64, 2013: 46), and he distinguishes a basic concern tied to the emotion’s affective dimension and a consequent concern or motivation tied to action (Roberts 2003: 144, 2013: 116). He defines fear, for example, as follows: “X presents a threat to Y of a significant degree of probability; may X or its threatened consequences for Y be avoided” (Roberts 2013: 113). The basic concern in fear is the avoidance of danger; this motivates, on Roberts’s view, the consequent concern to do something that will remove the danger. The consequent concern motivates the action. Against views that claim that the tendency to action is an aspect of an emotion, I would note that there are emotions whose “performance dimension” terminates in the expression of the emotion. Emotions involve undergoing physiological changes and sensing these in bodily feelings (Drummond 2020), but our bodies express our emotional experience in ways that go beyond these undergoings. There is, in other words, emotion-motivated behavior that falls short of action. Emotion-motivated behaviors are (i) initiated by us, but (ii) only sometimes under our control, and (iii) lacking an end. Emotion-motivated behaviors are physical enactments of the emotions they express. Whether or not they are under our control I shall call them “incomplete actions.” Consider, for example, Joe who is angered by someone’s comment in a meeting. Joe undergoes an acceleration of his heart rate, a rise in the skin temperature of his fingers, a tightening of the muscles in his abdomen, jaw, and around the eyes, and the clenching of his fingers. In these none of the conditions for action are satisfied. Suppose, however, that Joe also completes the clenching of his fingers and pounds his fist against the table. In forming and pounding his fist, Joe initiates the action, although, assuming he is not being theatrical, it is not clear whether he controls it. Whereas the physiological changes Joe undergoes are intrinsic to the emotion, the physical performance he undertakes in pounding his fist on the table is not intrinsic to the emotion. On the assumption that Joe is not being theatrical, he has no desire to pound the table. There is no end he seeks to realize in pounding the table; Joe is not acting for a reason. He just reacts—emphatically—to the comment made by his colleague. Emotion-motivated behavior, in brief, is incompletely acting; it consists wholly in—that is, is exhausted by—expressing the emotion. Another reason to doubt that emotions incorporate a desiderative moment is that some emotions, for example, awe and wonder, do not incline us to act. They involve bodily expressions similar to surprise reactions but not desire; they tend instead to arrest action. Joy too seems to terminate in bodily expressions, that is, in emotion-motivated behaviors: smiling, jumping for joy, raising both arms after scoring a goal, punching the air, and so forth. But they are not aimed at an end. Nor does desire seem a component of emotion when we 378

Acting, choosing, and deliberating

are “struck” by value (Mulligan 2009: 154). Such experiences are often expressed by an exclamation or interjection, such as “How sad” upon hearing of a colleague’s serious illness, or “How vulgar” upon witnessing someone’s behavior, or “Such generosity” upon hearing that someone has made a magnificent gift to an institution. The sadness, disgust, and admiration underlying such exclamations neither arouse desire nor motivate actions arising from the emotion itself. When emotion does incline us to act, the relation between the emotion and the action is made complex by the mediating role of desire. In the examples above, the emotions do not motivate action insofar as they do not motivate desire. When, however, an emotion does motivate a voluntary action as, say, in fear’s motivating flight, it first motivates the desire that, in turn, motivates the volition. That is, the desire motivates both the fiat that emptily and implicitly intends the end and the action that is the processual fulfillment of the fiat. The good as realizable is the object of a desire, but it is the volition, the fiat, that initiates the action that will realize that good. In such a case, in brief, the voluntary action is (i) initiated by the fiat, (ii) controlled by the fiat, and (iii) aimed at an end. Husserl stresses the fact that the desire or wish is not the willing: “Mere wishing is not willing; it contains nothing of the practical modalities and is not itself a practical act, an act of the will in the broadest sense” (Husserl 1988: 103; see also Melle 1997: 179). I can wish for the impossible or for what is past (or both, as in “I wish I did not have that second piece of pie at dinner last night”), but the wishing that is relevant for the will and for action is a wishing for what is practically possible, what is realizable in action. This wishing takes the form of desire. While the willing is founded on and inseparable from desire (Husserl 1988: 105), it is not reducible to it. As Michael Bratman argues, volitional intentions are subject to characteristic norms of consistency and means-end coherence. Roughly, she is rationally required not both to intend A and intend B if she believes that A and B are not co-possible; and, again roughly, she is rationally required to be such that if she intends E and believes M is a necessary means to E then she intends M. (Bratman 2012: 73; see also Husserl 1988: 221–224) Such norms, however, do not apply to desires. It is not irrational to desire both A and B even when knowing that their objects are non-compossible or to desire E while not desiring M. Moreover, it is also a volitional norm that when one wills to do A, and nothing interferes, one does A. We do not resist doing A once we have willed it, but we do regularly resist some desires. In summary, then, we can characterize the intentional structure of voluntary actions as follows: a b

c d

Subject S perceives (remembers, imagines, judges) an object O (a thing, situation, event, etc.) as having non-axiological properties x, y, and z; S’s perceiving (remembering, imagining, judging) O as x, y, and z motivates an intentional feeling or emotional experience that evaluates O as having the value-attribute b (bad, deficient, unsatisfactory); S envisions P (which might be a modification of or replacement for O) as having nonaxiological properties d, e, f; S’s envisioning P as d, e, f motivates an intentional feeling or emotional experience that prefers P (to O) as having the value-attribute g (good); 379

John J. Drummond

e f g

S’s preferring P as g motivates a practicable desire for P; S wills (the fiat) the voluntary action H realizing P; and S (voluntarily) performs H realizing P (a performance that over time fulfills the fiat).

It is important to stress that the multiple aspects of voluntary actions are not sequential. It is not the case that (a) through (g) follow one another in time. The motivating and founding relations obtaining among the various aspects are relations of sense, of meaning or significance, that have experiential import. The experience analyzed is the action as a whole, and (a) through (g) are internal, structured aspects of the voluntary action that realizes P.

Choosing Voluntary action, the progressive completion of a volitional intention ( fiat) that initiates and controls a physical performance aimed at a desired end, executes what Husserl in 1914 (1988: 107) calls an action-will (Handlungswille), a notion that parallels John Searle’s concept of intention-in-action (Searle 1983: 84, 2001: 44). Chosen actions differ from voluntary actions in that they not only aim at an end but are undertaken in the light of that end. In choosing, in other words, the agent has an explicit volition to attain the end. I choose to eat as a snack a piece of fruit (rather than, say, a chocolate sundae) for the sake of my health. It is not merely that I, as a matter of desire, prefer the fruit to the sundae and (merely) voluntarily eat the fruit; I choose the action with an eye toward the end of health (Aristotle, Nic. Ethics: 1111b26–27). The eating of the fruit is (iv) chosen as conducive to the end (Aristotle, Nic. Ethics: 1112b12, 1113a3–12; see also Sokolowski 1985: 11). Chosen action involves a decision, that is, choice and deliberation and, to that extent, reflection and reason. Let us reconsider Joe’s emotion-motivated behavior of pounding his fist on the table in anger. But now let us assume that he is being theatrical in order to ensure that his anger is noted. Joe is now acting for a reason. He is, however, not only acting toward an end; he undertakes the (chosen) action of pounding his fist in the light of the end of ensuring that his anger does not go unnoticed by the one who made the comment that angered him. To take another example, Bobby, a basketball coach, is angered by what he takes to be a bad call by a referee that could be crucial to the outcome of the game. He picks up his chair and tosses it out on the floor. At this point, his “action” might, like Joe’s original, non-theatrical pounding of his fist on the table, simply be an emotion -motivated behavior—an incomplete action— that emphatically (and dangerously) expresses his anger. Joe and Bobby in these examples are “acting” (behaving, incompletely acting) in anger. Suppose, however, that when Bobby throws his chair onto the floor, he aims it at the referee hoping to hit him in the legs and injure him. Bobby is now choosing and acting in the light of an end: to exact revenge. Bobby is now acting from anger and for the posited end of injuring the referee. With these ideas in mind, we can outline the intentional structure of chosen actions: 1 2

3

Subject S perceives (remembers, imagines, judges) an object O (a thing, situation, event, etc.) as having non-axiological properties x, y, and z; S’s perceiving (remembering, imagining, judging) O as x, y, and z motivates an intentional feeling or emotional experience that evaluates O as having the value-attribute b (bad, deficient, unsatisfactory); S envisions P (which might be a modification of or replacement for O) as having nonaxiological properties d, e, f; 380

Acting, choosing, and deliberating

4 5 6 7 8 9

S’s envisioning P as d, e, f motivates an intentional feeling or emotional experience that prefers P (to O) as having the value-attribute g (good); S’s preferring P as g motivates a practicable desire for P; S posits P as the end to be realized; S posits H as the chosen action (best) conducing to P; S wills (the fiat) the voluntary action H as realizing P; and S (voluntarily) performs H as realizing P (a performance that over time fulfills the fiat).

The fundamental difference between chosen action and voluntary action is located in (6), (7), and (8). There are two positings [(6) and (7)] present in chosen action that are not present in voluntary action. The consequence of this difference is that the agent in chosen action [(8)] wills the action (recognized as conducive to the end) be done as conducing to the end, whereas the agent in voluntary action wills that the object of desire be realized by acting. The positings in chosen action explicitly and thematically focus the end and the conducing action in a manner that the intentional structure of voluntary action does not. Moreover, when choice intends a future chosen action—the kind of willing Husserl had in mind when he spoke of a resolve (Entschlußwille) or a plan (Vorsatz) to do something—the prior [(8)] intention-to-do-in-the-future takes the form of intention-in-action when the action commences. Searle (1983: 94) argued that the prior intention-to-do causes the intention-inaction that in turn causes the bodily movements that together with the intention-in-action make up the action. John McDowell (2011: 3) argues against Searle that the prior intention-to-do becomes the intention-in-action. Whereas it is a feature of Searle’s view that the object of the intention-in-action is the bodily movements that belong to the action, McDowell argues that when one starts to do what one had a prior intention-to-do, e.g., to cross the street when the light turns green, one’s intention, now in action, is still directed at crossing the street, not at the limb movements that need to happen if one is to do that. In saying what the prior intention was an intention to do, one mentions a time: when the light turns green, in my example. When its time comes what was a prior intention takes a new shape as an intention in action, provided the agent does not forget the intention, knows the time has come, is not prevented from acting accordingly, and does not change her mind. (McDowell 2011: 3) McDowell focuses our attention on the unity of the overarching intention controlling both the prior intention—the choice that posits H as the action conducing to the end and the fiat regarding H—and the intention-in-action: at any time in the course of crossing the street, what one still needs to do is to go on from the point one has reached to the other side of the street. But the same structure fits projects with more complex shapes, such as baking a cake or building a house. In these cases, when we spell out what is required if one is to finish what one is doing, at the moment at which one is some way into doing it, we need to introduce things to do with more complex relations to what one has already done. (McDowell 2011: 6–7) The intention-to-do as transformed into the intention-in-action continues to control and govern the performance of a chosen action throughout the course of the action, that is, as 381

John J. Drummond

long as the action continues. In phenomenological terms, at every moment of a chosen action we have a partial fulfillment of the fiat, and the futural character of a prior intention-to-do— whether that prior intention is continuous or non-continuous with the chosen action— endures in the intention-in-action but diminishes as the action moves toward completion. Searle’s and McDowell’s distinction between prior intention and intentions-in-action appeals to the temporal distinction between intending a future act and intending while acting, as does Husserl’s distinction between resolve (Entschlußwille) and action-will (Handlungswille). My distinction between voluntary action and chosen action, by contrast, is based on the explicit positings of the end of an action and of the action’s conduciveness to that end. While the notion of voluntary action coincides with those of intention-in-action and Handlungswille, the notion of chosen action does not perfectly coincide with those of a prior intention and Entschlußwille. Choice, just as much as simple volition, can be concurrent with the action—a point that reemphasizes the idea of the identity of the intention in chosen actions and the intention-in-action—or it can be prior to the action in which case the intention in the chosen action becomes the intention-in-action.

Deliberating Choosing, we have said, involves deliberation and, therefore, reflection and reason. It is an activity proper to human persons. But what is it to deliberate, and how is it related to chosen actions? We have also said that choice can be occurrent—which I understand as concurrent or immediately continuous with the chosen action—but it can also be prior to action as in instances of resolving or planning to do something in the future. Must, however, deliberation be concurrent with the choosing? When reflecting on features of the criminal code of the late German Empire, Adolf Reinach noted “remarkable antinomies” (Reinach 1989: 279) in our understanding of deliberation, one of which is that meritorious actions count as less meritorious when done “without any deliberation” and also when done “only after long deliberation.” If, in other words, the agent takes no time to deliberate, the agent acts impetuously; if the agent takes too much time in deliberating, the agent is insufficiently attuned to what is right to do in certain kinds of circumstances. How do we reconcile the conflicting claims that there must be some deliberation for chosen actions, but not too much? Deliberation is the reflective reasoning about ends and conducing actions that underlies items (6) and (7) in the outline of the intentional structure of chosen action, the moments specifically concerned with choice. Having evaluatively identified choiceworthy ends, the agent deliberatively determines which end(s) will be pursued given her abilities, interests, and concerns. Deliberation also determines what action (best) conduces to the chosen end given the particulars of the situation in which the agent is to act. Persons are embodied, social, practical, and minded beings. A minded—better, m inding— being is first of all an intentional being (Drummond 2012: 25, 2019), and I use the term “intentional” here in its broad sense of “directedness to” rather than its specifically volitional sense. We can speak of a pre-reflective directedness both to objects in the world and to the flow of experiences to which the experiences of objects belong. The latter encompasses the momentary experience’s retentional directedness to prior experience as well as an anticipatory or envisioning directedness to yet-to-come experiences. It accounts, in short, for our first-personal, pre-reflective self-awareness (Drummond 2006: 217–218). We can also speak, however, of the directedness of empty intendings toward fulfillment, and here we find the teleological dimension of mind. In our discussion of Handlungswille and intention-in-action, we have seen the interconnection of the teleological sense of 382

Acting, choosing, and deliberating

directedness and the sense of directedness as object- and self-awareness (Drummond 2006: 218). The Handlungswille is (non-thematically and implicitly) directed to an end to be realized as the object of (my) desire, and this intention is gradually fulfilled in the action initiated and governed by the fiat. For the purpose of discussing deliberation, we can summarize the intentional structure of the choice involved in chosen action as follows: C is a choice that issues in action H as (i) having end P, (ii) grounded in the valuation of P as good (or apparently good) and desirable and of the action H as good (or apparently good) in its own right, and (iii) conducive to realizing P. Deliberation, then, is an exercise of reason that bridges the axiological and practical spheres. It begins in axiological reason’s evaluation both of the choiceworthiness of the end we pursue in action and of the (proposed) actions conducive to it, and it terminates in practical reason’s determination of the ends to be pursued and the actions to be undertaken. There is, however, one constraint on our deliberation about ends. The telos of intentional experience, we have argued elsewhere, is truthfulness, that is, apprehending things and states of affairs as they truly are, having appropriate affective and evaluative attitudes toward them, and acting rightly in response to and on the basis of our truthful cognitions and attitudes (Drummond 2010: 413). The achievement of evidenced truth in all the domains of reason is the full, autonomous exercise of reason (Husserl 1989a: 33). An autonomous rational agent “decides” for herself what is a true cognition, an appropriate emotion or attitude toward things, and a right action (Husserl 1989b: 281–282). Better, an autonomous rational agent reflectively and consciously appropriates as her own conviction what is evidently experienced (Drummond 2010: 415). In doing so, she becomes self-responsible—responsible for her beliefs, attitudes, and actions, for who she is—and therein flourishes as a human person. The presence of self-responsible truthfulness as a fixed, albeit formal, end for human agency provides the context for understanding the habituation of judgments, beliefs, and attitudes that constitute non-occurrent deliberation concerning the ends we shall adopt and our reasoning concerning the actions most conducive to those ends. The reflective, deliberative agent evidentially and self-responsibly confirms or disconfirms the evaluations of ends as well as the principles of action embedded in the evaluative and moral concepts that we, by and large, inherit. An autonomous rational being strives toward evidentially fulfilling the passively acquired judgments, beliefs, and emotional attitudes that make up her traditional or cultural inheritance. Her understanding of these concepts is further affected by her experience of how she and others have acted in situations to which these concepts are relevant. At the moment of action features of the situation evoke and bring into play these concepts, the agent’s previous appropriations or rejections of them, and her understanding of what actions are well suited to the situation. When she gains the evidence that allows her self-responsibly to appropriate, revise, or reject these acquisitions, she adopts them, their revisions, or their contraries or contradictions as convictions. Such convictions inform her subsequent judgments, valuations, and choices. The self-responsible agent who without occurrent deliberation can act rightly and toward good ends has already weighed competing goods, and she has already considered the rightness or wrongness of actions conducing to these ends. She has done so in a reflective, deliberative, and self-responsible activity accomplished over time in such a way as to dispose her toward a certain kind of action in certain kinds of circumstances. These appropriated convictions— “habitualities,” as Husserl calls them (Husserl 1970a: 66–67)—make up our 383

John J. Drummond

dispositions to pick out what is salient in certain kinds of situations, to have certain kinds of attitudes toward them, and to act in typical ways. This process of appropriation or rejection in the light of direct evidence is the ongoing process of deliberation that allows us to understand how an agent can act rightly without occurrent deliberation and why too much deliberation reveals an undeveloped or underdeveloped sense of the good and the right. Since these previous judgments were not originally intended for the present circumstances, non-occurrent deliberation yields only a sense of the specific kinds of action that might be undertaken in the present. Where deliberation is non-occurrent, in other words, prior deliberation yields the sense of a specific action—that is, a certain kind of action—as appropriate for the particular circumstances in which the agent finds herself. The particular determination is undertaken on the basis of my perception of the current situation. But all that is needed for that to occur is the current perception that is informed by my already achieved sense of what ends are desirable in this kind of situation and of what actions are appropriate to achieving them (Cooper 1986: 22–24, 58). Let us consider an example. Jane is on her way home from work and walking toward the bus stop. Seeing her bus starting to pull away from the curb, she, without occurrent deliberation, begins to run or to wave her arms (or both) in the hopes that the bus driver will see her, stop the bus, and wait for her to board. Is Jane’s action voluntary or chosen? Both are possible. While the lack of occurrent deliberation suggests she is acting (merely) voluntarily, she might already be aware of the “correctness”—in this case, the efficacy—of such actions. If her sense of the efficacy of the actions is simply assumed from prior observations but is not reflected upon and deliberated in the past, then she is acting voluntarily; she is just doing “what one does” in such a situation. But if her sense of that efficacy involves prior reflection and deliberation, her action would be chosen. When Jane, informed by past experience but not by past deliberation and choice, just runs and waves her arms there is only a Handlungswille and intention-in-action. If, by contrast, she is knowingly acting to get home and runs and waves her arms because she knows, given where she is in relation to the driver’s field of vision, that running and waving her arms best conduces to attracting the bus driver’s attention, then she is choosing on the basis of her non-occurrent, prior deliberation. That prior deliberation specified for her the type of things she could do to attract the driver’s attention, while the perception of her relation to the driver’s field of vision shapes her choice from the available alternative actions. Karl Mertens has argued that Husserl’s account of Handlungswille relies on “a nondescriptive differentiation between will-intention and its corresponding fulfilling action” (Mertens 1998: 133). Husserl, on Mertens’s view, understands volitional intention on the model of empty intention and its fulfillment, where the empty intention involves a positing of a determinate state of affairs to be realized. This structure is clear in willing future action (Entschlußwille), but it is not in Handlungswille. On Mertens’s view, “although will and action in Husserl’s analysis of volitional action are inseparable, the distinction between intention and fulfillment of action-will can be pointed out in reflection” (Mertens 1998: 132). Mertens acknowledges that “we often constitute the determination of an action-will only after the performed action. But this constitution should not be confused with a mere description of the action as it is performed” (Mertens 1998: 133). Hence, Mertens argues, Husserl is guilty of reading the structure of the future intention of Entschlußwille back into Handlungswille. The view I have outlined avoids this issue, and I think Husserl too can avoid it. The key is to distinguish thematic (explicit) and unthematic (implicit) volitional intendings. This is the basis on which I have distinguished chosen action from voluntary action. Chosen action involves thematic positings of the end of the action and of the action as conducing to that end; 384

Acting, choosing, and deliberating

voluntary action does not. In voluntary action there is an unthematic volitional directedness rooted in the desire for a certain state of affairs—a desire that motivates action—as opposed to chosen action’s thematic volitional directedness to the state of affairs as an end to be realized. Moreover, Husserl distinguishes determinate from indeterminate intentions. While this distinction is important in discussions of cognition—the fulfillment of a cognitive intention relative to a practical interest involves gaining enough determination in the fulfilling sense as is demanded by our interest in the object—it is of crucial importance in volitional intentions. The performance of a willed action takes some time, during which the willing is progressively fulfilled. Just as perceptual experience revises and more precisely determines our sense of the perceived object, volitional intentions involve a certain amount of indeterminacy regarding the details of the action to be performed and of the final result. This, I maintain, is true of both prior volitional intentions and intentions-in-action. On the view suggested here, voluntary action is occurrently always a Handlungswille and intention-in-action. However, in some cases, that intention-in-action is the becoming of a prior deliberation; in that case, the action has the intentional structure of a choice, although its first-personal character is different from what we find in an Entschlußwille wherein the nonoccurrent deliberation and choice is passively re-appropriated rather than actively and occurrently made. In other cases, however, there is only the Handlungswille and the intention-in-action, which are informed by past experience, although not by past deliberation and choice. If someone were to demand from Jane reasons for acting as she did, she could offer them only as an ex post facto deliberation identifying the motives and justification for her action, but this does not change the character of the action as performed from a voluntary action to a chosen action. Mertens also claims that even if Husserl could overcome the first problem by more carefully distinguishing the ex post facto rationale of an action from the structure of the action itself, his account makes it “impossible for a volitional action to surprise its agent” (Mertens 1998: 133). Any action is, in a sense, creative, since it brings into existence a new state of affairs. Three considerations speak for the possibility of surprise. First, in fulfilling a volitional intention, as in fulfillment in general, there can be resistance or recalcitrance arising from features of the situation or the things involved in the action. Second, recalling that in intention-in-action the fiat is gradually fulfilled over the course of the action, we can note that when the empty intention is also indeterminate to some degree, the action not only fulfills but also further determines the nature of the intention-in-action, the fiat itself, as the action unfolds. Third, there can be actions whose end is simply the activity itself and its development, for example, Jackson Pollock’s “drip painting” or, as it is sometimes called, “action painting.” As Harold Rosenberg (1952) put it, “What was to go on the canvas was not a picture but an event . . . The big moment came when it was decided to paint. . . . Just TO PAINT.” In such cases, of course, creativity in the result is just the point, but there remains an unthematic and indeterminate willing of that result along with the intention-in-action itself. The volitional intention is to let the process develop and determine its own direction and result. These considerations show that there can be frequent “course changes” or further determination in acting or a willing of the activity itself as the end, and these can lead to both surprise and an unforeseen creativity.

Resolve To resolve or to plan to do something (Entschlußwille) intends the realization of an end or purpose in the future. An agent might, for example, will an end that can be attained only by repetition of the same action. Eileen practices the piano, for example, in order to become a 385

John J. Drummond

good piano player. As a young child just beginning to play the piano, Eileen did not realize that end in the very act of practicing. Nor did she realize it as immediately contiguous with any particular instance of practicing. It is only by repeatedly practicing over a long time that Eileen becomes a good piano player, and it is only by continuing to practice that Eileen remains a good piano player. The commitment to realizing that end informs and renews the present commitment to practice. Another possibility occurs when I deliberate and choose now to act later. A student, for example, decides as a sophomore that she wants to apply to graduate schools. She will not actually apply until her senior year. Her choice to apply to graduate school does not immediately produce an application. In this sense, her volitional intention remains empty, not to be fulfilled until she actually applies. Nevertheless, the future she foresees with this decision affects her present in a variety of ways. She must plan and do those things that will put her in a position to apply to graduate schools: choose a major and the right courses within the major, gather information about the schools to which it would be best to apply, consider which of the papers she has written would provide a good basis for the writing sample she must include in her applications, and so forth. Although her action of applying to graduate school is in the future at the time she decides to apply, her resolve to undertake that action in the future shapes her present planning. More importantly, these possibilities reflect the structure that is at work in the commitments around which we organize our lives: commitments as a spouse, parent, philosopher, teacher, citizen, and so forth. For the committed agent, practical reason recalls the effective force of past reflection, deliberation, and action into the present just insofar as that force is relevant to determining what I should do here and now. For the committed agent, the present is conceived in relation to the future-oriented commitments that organize her life, among which the priorities she establishes might sometimes shift as the circumstances of her life change. Resolve and planning of this sort are a way of willing what we shall be(come). And for the flourishing agent, these choices are constrained by and simultaneously realize the overarching, second-order good of truthful self-responsibility.2

Related topics See Chapters 2 (Husserl), 20 (Smith), and 26 (Figal).

Notes 1 This chapter will focus on actions involving a bodily performance. As will become clear in the discussion of the role of intentions in acting, this does not mean that actions lack a mental aspect. Nor does it mean that there are no purely mental actions as, for example, contemplating. 2 I am grateful to anonymous referees for helpful comments on an earlier version of this chapter.

References Aristotle (1984) The Complete Works of Aristotle: The Revised Oxford Translation, ed. J. Barnes, 2 vols., Princeton, NJ: Princeton University Press. Bratman, M. (2012) “Time, Rationality, and Self-Governance,” Philosophical Issues 22, 73–88. Brentano, F. (1995) Psychology from an Empirical Standpoint, trans. by A. Rancurello, D.B. Terrell and L. McAlister and ed. L. McAlister, London: Routledge. Cooper, J. (1986) Reason and Human Good in Aristotle, Indianapolis, IN: Hackett. Drummond, J. (2006) “The Case(s) of (Self-)Awareness,” in U. Kriegel and K. Williford (eds.), SelfRepresentational Approaches to Consciousness, Cambridge, MA: The MIT Press, 2006, pp. 199–220.

386

Acting, choosing, and deliberating Drummond, J. (2010) “Self-Responsibility and Eudaimonia,” in C. Ierna, H. Jacobs and F. Mattens (eds.), Philosophy, Phenomenology, Sciences: Essays in Commemoration of Edmund Husserl, Dordrecht: Springer, pp. 411–430. Drummond, J. (2012) “Intentionality Without Representationalism,” in D. Zahavi (ed.), The Oxford Handbook of Contemporary Phenomenology, Oxford: Oxford University Press, pp. 115–133. Drummond, J. (2013) “The Intentional Structure of the Emotions,” Logical Analysis and the History of Philosophy/Philosophiegeschichte und logische Analyse 16, 244–263. Drummond, J. (2019) “Intentionality and (Moral) Normativity,” in J. Marsh, I. McMullin and M. Burch (eds.), Normativity, Meaning, and the Promise of Phenomenology, London: Routledge, pp. 101–119. Drummond, J. (2020), “The Varieties of Affective Experience,” in T. Szanto and H. Landweer (eds.), The Routledge Handbook of the Phenomenology of the Emotions, London: Routledge, pp. 239–249. Husserl, E. (1970a) Cartesian Meditations: An Introduction to Phenomenology, trans. by D. Cairns, The Hague: Martinus Nijhoff. Husserl, E. (1970b) Logical Investigations, trans. by J.N. Findlay, London: Routledge and Kegan Paul. Husserl, E. (1988) Vorlesungen über Ethik und Wertlehre, 1908–1914, ed. U. Melle, Dordrecht: Kluwer. Husserl, E. (1989a) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Second Book: Studies in the Phenomenology of Constitution, trans. by R. Rojcewicz and A. Schuwer, Dordrecht: Kluwer. Husserl, E. (1989b) Aufsätze und Vorträge (1922–1937), ed. T. Nenon and H. Rainer Sepp, Dordrecht: Kluwer. McDowell, J. (2011) “Some Remarks on Intention in Action,” The Amherst Lecture in Philosophy 6, 1–18. Melle, U. (1997) “Husserl’s Phenomenology of Willing,” in J. Hart and L. Embree (eds.), Phenomenology of Values and Valuing, Dordrecht: Kluwer, pp. 169–192. Mertens, K. (1998) “Husserl’s Phenomenology of Will in His Reflections on Ethics,” in N. Depraz and D. Zahavi (eds.), Alterity and Facticity: New Perspectives on Husserl, Dordrecht: Kluwer, pp. 121–137. Mulligan, K. (2009) “On Being Struck by Value: Exclamations, Motivations, and Vocations,” in B. Merker (ed.), Leben mit Gefühlen, Paderborn: Mentis, pp. 141–163. Pfänder, A. (1913) “Zur Psychologie der Gesinnungen,” Jahrbuch für Philosophie und phänomenologische Forschung 1, 325–404. Reinach, A. (1989) “Die Überlegung; ihre ethische and rechtliche Bedeutung,” in K. Schuhmann and B. Smith (eds.), Sämtliche Werke: Textkritische Ausgabe in 2 Bänden, Erster Band, Munich: Philosophia, pp. 279–311. Roberts, R. (2003) Emotions: An Essay in Aid of Moral Psychology, Cambridge, MA: Cambridge University Press. Roberts, R. (2013) Emotions in the Moral Life, Cambridge, MA: Cambridge University Press. Rosenberg, H. (1952) “The American Action Painters,” ARTnews 51/8, 22–50. Searle, J. (1983) Intentionality: An Essay in the Philosophy of Mind, Cambridge, MA: Cambridge University Press. Searle, J. (2001) Rationality in Action, Cambridge, MA: The MIT Press. Sokolowski, R. (1985) Moral Action: A Phenomenological Study, Bloomington: Indiana University Press. Stein, E. (1989) On the Problem of Empathy, trans. by W. Stein, Washington, DC: ICS Publications. Thompson, M. (2008) Life and Action: Elementary Structures of Practice and Practical Thought, Cambridge, MA: Harvard University Press. Zagzebski, L. (2003) “Emotion and Moral Judgment,” Philosophy and Phenomenological Research 66, 104–124.

387

26 INVOLUNTARINESS Actions and their context Günter Figal

Aristotle on action Philosophical theory of action has a charter not difficult to discover. The first page of Aristotle’s Nicomachean Ethics (NE) includes all essentials that have been dominating pertinent philosophical discussions till today. Stating the teleological character of actions, of production as well as of action in the narrower sense, Aristotle makes clear how activities can be determined such as to become actions. Only because actions have a particular aim, they are not just indiscernible moments in a continuous stream of life that are inseparable from each other like waves. Their respective aims make them intelligible; a particular action is defined by its aiming at a particular aim. Aims also function as unifiers of different effects and accomplishments performed by someone. Regarded as contributions to the realization of a particular aim, different accomplishments can be coordinated in order to complement each other. Since every contribution to realizing a particular aim is part of a particular action, the question ‘What are you doing?’ can best be answered by designating the aim one has intended and, if necessary, by explaining how particular accomplishments fit into the action they belong to. Such fitting is neither by nature nor does it just happen. Rather, it is the result of intelligently planning and efficiently coordinating different accomplishments so that they form an action as consistent as possible, or, to be brief, the result of technical and practical knowledge. Knowing for instance how to make a bowl or how to give a lecture, one is able to choose and coordinate different accomplishments as means for a particular aim, and, realizing them, to reach the aim in respect to which they could be understood as means. The teleological structure as sketched applies both to productions, the aim of which is their respective product, and to actions in a more specific sense. Such actions can be characterized as aiming at results that are no products and mostly belong to the sphere of interpersonal life. Giving a lecture or a public speech, striving for a position, shopping or taking a train in order to get to work – all these are actions that, as intentional activities, are determined by their aims. Aristotle’s theory of action just explained has clear advantages. Giving a criterion that allows singling out distinctive actions from the stream of active life, Aristotle makes impressively clear why and how actions can be decided, planned, performed, explained, and, if necessary, justified. Thus he reveals the specific rationality of actions and, along with this, 388

Involuntariness

outlines the scope of practical rational discourse. And, finally, he establishes a basis for a clear understanding of human responsibility, freedom, and maturity. Persons cannot be regarded as responsible for the stream of life, but only for particular actions they have, more or less clearly, intended and initiated and have thus performed voluntarily. Deciding to pursue particular aims and pursuing them voluntarily, persons reveal their free will and, in this sense, their freedom. And, finally, doing deliberately and thoughtfully what they voluntarily do, persons prove to be mature human beings. Persons who act voluntarily and are thus, as they should be, in accordance with their practical rational faculties present themselves as the free, mature, and responsible persons needed for a good social and political life. The sketched ideal of a free, mature, and responsible personality based on the very possibility of intentional and purposeful action cannot seriously be doubted without contesting the possibility of rationally guided social life. Since there cannot be any good reason for abandoning rationality, such a contest would have no rational ground. Nonetheless, Aristotle’s conception of action and, going along with it, the ideal of free, mature, and responsible personality is not sacrosanct. It can be revised and refined, and it may even need refinement, not least in order to protect it against a problematic tendency of one-sidedness: rational persons manifesting their rationality by pursuing, explaining, and, if necessary, justifying intentional actions may be inclined to neglect that not all conditions of their activities can be adequately understood as fitting into the scheme of aim-oriented action. Such neglect may lead to the assumption that, at least in principle, everything could be taken as a possible means for a person’s aims and, in case of appropriateness, should be brought under control. Everything withdrawing from control by rational persons would, in contrast, appear as marginal or as disturbing, if not as adverse. In the latter case, it would be a challenge or a threat to rational persons. Aristotle surely is not disposed to arguing in favor of a human life dominated in such a one-sided way by will and power. In contrast to Sophistic positions, maintained for instance by Kallikles in Plato’s dialogue Gorgias, Aristotle does not ignore the limits of rational action. Rather, he recognizes that things can happen to persons they are unable to technically or practically master, and accordingly just have to undergo. Aristotle calls the experience of such things ‘involuntary’ and thus contrasts it with the voluntary character of actions. Nevertheless, Aristotle does not definitely exclude the mentioned one-sidedness. According to his conception, human life is essentially active so that actions are more significant for human life than its involuntary aspects. Since Aristotle nevertheless discusses the voluntary (τὸ ἑκούσιον) in relation to the involuntary (τὸ ἀκούσιον) and thus considers the limits of voluntariness, he offers a good starting point for examining how the relation between the voluntary and the involuntary can be described most adequately. For the sake of such a description, one should find out how in detail Aristotle’s conception of action and, going along with it, the ideal of free, mature, and responsible persons is in need of revision, and how it can be revised.

Aristotle on the voluntary and the involuntary Defining the involuntary as something done under constraint (βιᾳ) or through ignorance (δι’  ἄγνοιαν, NE III.1: 1109b 35–1110a 1), Aristotle implicitly argues for understanding involuntariness within the scope of aim-oriented action. Thus he excludes everything non-voluntary the understanding of which is independent from any reference to action. Such are physical processes as well as spontaneous but unplanned gestures or facial expressions. The reason for this may be obvious in the first case and comprehensible in the latter. Physical 389

Günter Figal

processes like heartbeat do not contrast with voluntary actions so that it would make no sense to call them ‘involuntary’. They just happen, and they are not expected to be specifically relevant to actions. This is slightly different with gestures and facial expressions. They could be voluntarily performed, for instance by an actor or by someone just imitating someone else’s behavior, whereas they normally are not intended. However, gestures and facial expressions are not lacking voluntariness for whatever reason. They can be understood as spontaneous and unplanned movements without any reference to the voluntariness of actions. The same holds true for emotions going along with actions. They, too, are not lacking voluntariness, and even if they hinder persons to consequently perform what they are aiming at, they are by nature unavoidable and accordingly cannot be in contrast to voluntary actions. As Aristotle says, they are human, and thus it would be strange to count them among the involuntary (NE III.3: 1111b 1–3). So, according to Aristotle’s argument, only what is in contrast to actions can be called ‘involuntary’. Actions would be just voluntary, if they were not impaired by a cause ‘outside’ action and without any contribution of the acting person (NE III.1: 1110b 2–3). Negating voluntariness by constraint, at least partly, such effects are measured by voluntariness and therefore justly called ‘in-voluntary’. For instance, a person carrying a vase and slipping on a wet floor so that the vase falls to bits does not slip and destroy the vase voluntarily. Slipping and dropping the vase just happens to this person. It was not intended, but rather happened impairing an intention. Basically the same holds true for involuntariness through ignorance. Actions missing their respective aims because the acting person had no chance to understand these aims adequately are not voluntary. The failure going along with these actions was not intended, and necessarily so, because, as Plato’s Socrates maintains, failure essentially cannot be intended.1 As to this, Aristotle agrees with Socrates; but Aristotle in addition differentiates between actions that are ‘not voluntary’ and those that are involuntary in a strict sense. If persons just are ignorant of what they are doing, they do something they had not intended. However, if persons understand that their actions conflict with their intentions and, because of this, feel concerned and regret what they have done, they have done something involuntarily. So persons damaging a camera because they do not know how to use it correctly have caused the damage, just not voluntarily. Someone, however, affronting someone else mentioning the other’s professional success without knowing that the addressed person recently lost her job, does this involuntarily. Speaking about the other’s professional success was meant as a compliment, and only happened to be an insult. In this case, persons would regret a comment and thus indicate that the comment’s effect was unintentional. The distinction between the voluntary and the involuntary as it has been explained and illustrated so far should be plausible. One should add, however, that Aristotle does not suggest a neat contrast – as if persons would either do something voluntarily or undergo something involuntarily or would either have clear knowledge guiding their actions or be totally ignorant. Aristotle rather makes clear that practices can have both voluntary and involuntary aspects. Though briefly, he discusses what he calls “mixed actions” (μικταὶ πράξεις) and illustrates this designation by the example of a ship crew that, getting into a sea storm, would throw a portion of the load overboard in order to save the ship and their own lives (NE III.1: 1110a 8–12). Though the crew have of course not intended and initiated the sea storm, their throwing load overboard is not immediately caused by the sea storm. What they do is a possibly well-considered intentional action that, ruling out other options, has been decided on. On the other hand, however, the crew would not have considered throwing load overboard

390

Involuntariness

without the sea storm. So, the sea storm, as an event just happening to the ship crew, conditions their action, albeit not in every respect. Aristotle’s example is completely in line with his general conception of the voluntary and the involuntary. A sea storm is an event disturbing the normal practice of sailing and thus causing persons to do what they otherwise would not have done. It is a kind of accident that could not have been avoided and as such cannot be brought under control. The ship crew could only react to it, and, in doing so, at least partly re-establish voluntary control under adverse conditions. Thus Aristotle’s example confirms the assumption according to which the voluntary has priority over the involuntary, and the involuntary aspects of actions only appear as restriction. Moreover, the example illustrates the mentioned one-sidedness of understanding rational persons primarily as agents, whose actions, if ever possible, should be without any interventions from ‘outside’. This understanding, however, is in need of revision, if such interventions are not at any rate disturbing, but rather supportive or even necessary. Then, involuntariness would still be regarded with respect to voluntariness. However, instead of negating voluntariness, the involuntary would complement the voluntary. As a consequence, understanding the rationality of persons should include the recognition of this complementarity.

Heidegger on involuntariness as complement of action How can involuntariness complement voluntary actions? Heidegger was first to make a decisive step toward an answer to this question. Human ‘Dasein’ as conceived in Being and Time is not only characterized as “existence” (Existenz), i.e. as openness to possible projects and thus to intentional, aim-oriented actions, but also as “facticity” (Faktizität), which essentially implies that a human being can understand itself “as bound up in its ‘destiny’ with the being of those beings which it encounters within its own world” (Heidegger 1977: 75, 2010: 56). In order to ‘encounter’ something whatsoever in the world, one has to understand oneself as being essentially connected with things that can be encountered. Or, to make this more concrete with respect to aim-oriented actions: in order to discover something as a possible means for realizing one’s aim, one must have an understanding of things as allowing for such a discovery. Since nothing is as such a means for realizing particular intentions, all things can only be discovered as means if they have been understood in their primary being, which as such precedes voluntary action. Practical considerations concerning possible means for realizing a particular aim can only distinguish between things appropriate and things inappropriate as means if things are basically understood as what they are. As a consequence, such a basic or primary understanding is relevant for the rationality of aim-oriented, voluntary action and nonetheless beyond such action. In this sense, it is involuntary. Heidegger has developed his conception of primary understanding by explaining the “being of innerwordly things” (das Sein des innerweltlichen Seienden, Heidegger 1977: 112, 2010: 82).2 He determines this “being,” which is the being of utensils or “things at hand” (Zeug) as “Bewandtnis” and thus introduces an expression hardly translatable, but nevertheless explicable. The meaning of the expression is twofold, and in order to grasp Heidegger’s point, both aspects must be taken into account: “Bewandtnis” means “property,” whereas the verb corresponding to the noun “Bewandtnis,” i.e. “bewenden lassen,” means “to let a matter rest.” Taking the two meanings together one can determine “Bewandtnis” as a property of something that is only to be understood if one lets it rest – as a property not to be explicitly referred to, but to be grasped, as it were, tacitly. In order to understand “innerworldly

391

Günter Figal

things” practically one must not be attentive to them, but rather allow them to remain as inconspicuous as it is appropriate for things of that kind. For instance, using a tool in order to produce something, one is attentive to one’s work, not to the tool. The meaning of “Bewandtnis” just explained is not yet complete. It applies only to the particular use of something that must be “let rest” in order to be used. The meaning, to introduce Heidegger’s terminology, is only “ontic,” and not “ontological.” However, as Heidegger argues, for “letting rest” a particular thing, an ontological “letting rest” that applies to “innerworldly things” in general is required. Understood ontologically, “letting rest” is a “previous setting free” (vorgängige Freigabe) allowing all things to be open possibilities for any particular practical understanding. Things are accessible only as such prior possibilities for action. Speaking of a “previous setting free” Heidegger does not evoke an enigmatic kind of action that precedes all particular activities. As ontological “letting rest,” this release is not an action at all, but rather a prior recognition of things as the things they are. This recognition can be called “involuntary,” if involuntariness is not understood in line with Aristotle, as impairment of voluntariness, but as its enabling. Then, “ontic” involuntariness is a refraining from intentional reference, and, “ontologically” understood, it is a refraining that goes along with an intuitive understanding of all things in their potentiality preceding all particular practical determinations. However, with his considerations concerning Bewandtnis, Heidegger only touches upon such a conception of a prior involuntary understanding of things. Instead of conceiving the potentiality that is the correlate of prior understanding as an objective character of things and as such beyond all kinds of practical discovery, Heidegger reduces this potentiality to the practical meaning of things as their being “at hand” (Zuhandenheit). As a consequence of Heidegger’s view things are not really “set free,” but rather predetermined by human practice. Accordingly, the involuntariness of “setting free” is predominated by the voluntariness of actions. Involuntariness is only a necessary condition for voluntariness, and measured by voluntariness. It is not its complement or even counterpart so that it could reveal the limits of practical understanding. So, Heidegger affirms the dominance of practical understanding, as becomes especially clear with his conception of the “ontic” version of “letting rest,” which is nothing but a necessary condition of the efficient use of a thing whatever. For the author of Being and Time, all things are modeled on the paradigm of utensils and tools. Things, whatever they are, belong to the scope of human practice. Heidegger has revised this conception developing a more radical version of “previous setting free” in his essay On the Essence of Truth (Vom Wesen der Wahrheit).3 According to this version, things are not “set free” as beings “at hand,” but rather just as unconcealed beings. Thus the prior setting them free is an engagement “with the disclosedness of beings” (Heidegger 1993: 125). However, whereas Heidegger’s earlier version is too determinate, the later is too unspecific and, because of this, not applicable to reference to particular beings. As Heidegger maintains, any “letting be” of particular beings conceals the “previous setting free” of being as a whole (Heidegger 1993: 129–130). An attempt nevertheless to apply this “previous setting free” to particular beings in order to understand an involuntary “letting be” of such beings proves to be a failure. Unconcealment is the essence of all beings accessible for reference. Their unconcealment is no criterion for distinguishing between beings determined by voluntary actions and those one should involuntarily “let be” what and how they are. However, despite this result, one should not neglect the importance of Heidegger’s discovery. With the idea of a prior, and involuntary “setting free” of all things, Heidegger has changed the discourse on human practice in such a way that further clarifications of the 392

Involuntariness

constitutive role of involuntariness have become possible. In order to conceive involuntariness as a serious complement of voluntary action, one only has to develop Heidegger’s idea independently from his presupposition of a world dominated by human practice and also to overcome his all too general considerations on “unconcealment.” In contrast, one may assume that human practice is only possible in a world not dominated by it and, accordingly, find ways for a “setting free” of things that is concretely relevant for practice. How can such a “setting free” take place amidst actions, not as their integral part, but rather revealing how actions belong into a non-practical context? And how can a “setting free” apply not only to things but also to persons so that not only voluntary production but also voluntary action in the narrow sense could be complemented by involuntariness?

Strategies of involuntariness in the arts and the involuntariness of understanding In order to answer these questions one should first concentrate on productive activities, because in this case the complementary character of involuntariness for voluntary production can well be illustrated and even explained by an illuminative example. This example is artistic production, especially painting and, even more, ceramic art.4 Artistic production differs remarkably from ‘normal’ production. Though artworks often are crafted in a way similar to the production of things without any aesthetic significance, artworks considerably differ from such things in their originality and individuality. Accordingly, the production of artworks must not be oriented to a schematic form so that products more or less resemble each other. Artworks are as such not supposed to be examples of a particular kind of things with characteristic properties that make them identifiable and, if they are for use, functional. Rather artworks are made for appearing; they are supposed to appear in a way that would draw attention again and again. Artworks must not appear just as planned and constructed artifacts, but rather as if they were independent from the artists who produced them and also from a calculable process of production. As Kant puts it in the Critique of the Power of Judgment (§45), artworks, admittedly being products, should look “like nature.” Such originality of artworks cannot voluntarily be achieved. Artworks that show all too obvious marks of the artist’s intention to produce something ‘original’ are for the most part of inferior aesthetic quality. Accordingly, artists, and especially modern artists, have taken the original character of artworks seriously and developed what can be called ‘strategies of involuntariness’. Many artists have made attempts not to control every aspect of their production, however, without leaving production just to chance. Rather, they wished to introduce the involuntary into their practice and thus to establish an interplay of voluntariness and involuntariness. At least in the first instance, this may sound paradoxical. How should one introduce the involuntary without doing that voluntarily and thus negating the involuntariness? However, regarded more closely, aesthetic strategies introducing the involuntary are not paradoxical at all. Artists developing such strategies do not voluntarily determine what is declared as ‘involuntary’, but rather restrict voluntariness in such a way that something involuntarily may happen. They voluntarily do something that is supposed to have involuntary effects. So, they do not just voluntarily produce something, but rather allow involuntariness to take place as a complement of voluntary production. Probably the most prominent example for this in modern Western art is Jackson Pollock’s “Drip Painting.” Instead of neatly applying paint by a brush to a canvas positioned on an easel, Pollock poured, dropped, and splattered fluid paint to canvases lying on the ground. 393

Günter Figal

Thus performing voluntary gestures, Pollock initiated effects he had not intended and realized by way of a method. Another prominent artist introducing the involuntary into painting is Gerhard Richter. Richter’s aesthetic strategy is different from Pollock’s, but also establishes an interplay of voluntary practice and involuntary effects. For his abstract paintings, Richter generously applies oil paint on a canvas fixed to a wall of his studio before, using a board’s edge as long as the canvas in a vertical position, partly, scrapes off the paint so that irregular traces of paint are left to the canvas. Because Richter repeats this procedure again and again, a painting regarded as ‘accomplished’ shows a rich and deep texture of many layers partly covering deeper layers and partly allowing them to be visible. Like Pollock’s paintings, Richter’s have not been planned in advance. Instead of being methodically produced, they originate. Nonetheless, however, their becoming is no natural event like a sea storm or the growing of a tree but rather a result of artistic practice as a combination of voluntary and involuntary aspects. The involuntary is even more dominant in the art of Japanese ceramics, which, despite its age-long tradition, is true modern art.5 Already the process of throwing a vessel is as such less controlled than painting. During long-time exercise, master artists would acquire a kind of certainty that makes planning and intentionally controlled work at least partly unnecessary. Working at a potter’s wheel often is more a practice like walking or riding a bicycle than like intentional work. Moreover, it is no work depending solely on the artist’s activity. Rather, a ceramic artist would answer to the movement of the lump of clay on the potter’s wheel and thus would produce a vessel only, as it were, ‘in interaction’ with the wheel and the clay on the wheel. The turning of the potter’s wheel, though initiated by the artist, is thus an involuntary counterpart to the artist’s activity. The firing is even more an ‘interactive’ process, especially with a traditional anagama, a “cave kiln,” which is fueled with firewood producing fly ash that would settle on the vessels and thus form a natural ash glaze. Though ceramic artists working with an anagama are able to control the firing to a certain degree and also to influence the character of the ash glaze by positioning the vessels in the kiln, they will never be able to plan the particular character of the glaze. So, firing an anagama ceramic artists essentially include involuntary processes into their artistic practice. They do so because of aesthetic reasons, and not because they would have no choice. Artistic production thus described is not that much an exception as it might appear. Though production normally is not meant as interplay between voluntary and involuntary aspects that allows something original and individual like an artwork to appear, art may function as a paradigm for understanding production in general. Every production includes aspects that cannot be actively originated and pursued. All woodwork, for instance, is dependent on the growth of trees and on the drying of timber necessary for fabrication. Gardening and agriculture are productive only to a certain degree, and even industrial production would control only a limited section of what necessarily belongs to it. In all these cases, one must allow something to happen in order that something can be done. However, for the most part, these involuntary aspects appear as marginal because production is concentrated on its intentional and effective and thus voluntary aspects. The involuntary also enables or at least improves social interaction and thus actions in the more specific or narrow sense. This may best be illustrated by one of the most basic modes of interaction, namely by dialogue. Persons can only converse with each other, make attempts to come to a common decision, or clarify something whatever, if they are able to listen to each other. Listening, however, is not an action, but rather something involuntary. Though persons can decide on listening to someone else more or less attentively or more or less focused, listening as such is nothing intentional. Persons listening do not aim at something and 394

Involuntariness

try to realize their intention, but rather allow other person’s narratives or statements to be performed freely, and in a way gaining recognition. This would especially be so, if listeners would not just be up to grasping other person’s words, but furthermore be receptive for their articulated feelings, views, and insights. This requires openness just to understand instead of taking what has been said as a confirmation of one’s own presuppositions. However, no one open for understanding will necessarily understand. One may wish to understand, but one will not be able to effectuate understanding, neither one’s own nor another person’s. Whereas persons can be forced into confessing or stating something, they cannot be forced to understand. So, understanding is no possible aim to be realized in voluntary action. It is involuntary, as Plato’s Socrates clearly realized when confining himself to indirectly supporting other person’s understanding in ‘maieutic’ dialogue. The involuntariness of understanding is an especially important example for how the interplay between the involuntary and the voluntary should be conceived. Whereas productive activities, as has been shown, are mostly more or less explicitly complemented by the involuntary, actions in the narrow sense are dependent on it if they are dependent on understanding. The latter, however, is the case with all actions performed in an interpersonal context or even intended as interactions. Such actions will necessarily fail in whatever way if they are not guided by understanding each other at least to a certain degree, but only by presuppositions, strategic assumptions, or egoistic motives. Manipulating others, instead of understanding them and, in open dialogue or debate, find solutions acceptable to all persons involved, is essentially inadequate to persons. In accordance with their personality, persons should, in principle, be related to each other in interpersonal life, and not be regarded as means or ‘objects’ of voluntary actions. Accordingly, it is the involuntariness of understanding that, once recognized, again and again confirms that indirectness is essential to interpersonal life.

Situations Voluntary actions, to sum up, being essentially, and more or less obviously complemented by or even dependent on the involuntary, cannot adequately be conceived if isolated. As a consequence, discussing actions philosophically one should take their interplay with the involuntary into account. Actions thus should be contextualized and described in their contexts formed by different versions of the involuntary. In a certain way, already Aristotle was aware of such a need for contextualizing actions as is mainly documented by his discussion of “mixed actions” – the voluntary action of a ship crew throwing load overboard is not intelligible without the context of the sea storm involuntarily experienced. However, as has been shown, Aristotle regards the involuntary only as limiting voluntary actions, and not as complementing or even enabling it. This one-sidedness, again, allows him conceiving of actions as he does: as aim-oriented and thereby voluntary activities performed in order to realize an aim initially intended. Contextualizing actions does not require abandoning the Aristotelian conception of actions. Complemented by involuntary aspects, actions do not cease to be basically intentional. A ceramic artist, for instance, devoted to the tradition of anagama firing, would intend to make vessels glazed with fly ash. Someone pursuing, for instance, cooperation with someone else would intend an effective agreement concerning this cooperation. However, the intentions just mentioned differ from the canonical Aristotelian model in their partial indeterminacy. They include, as it were, blank spots, aspects that essentially cannot be planned and controlled throughout. As a consequence, persons performing such actions should have sense for the indeterminate particularly relevant for their intentions and activities. They should be 395

Günter Figal

unperturbed enough so as to let things happen instead of making attempts to force what by forcing could only be spoiled. Persons with a sense for such indeterminacy would be able to distinguish what can be voluntarily pursued from what one should allow happening involuntarily. Such person’s practical rationality would thus include in particular a competence of judging which voluntary and involuntary aspects are relevant and how they can be related to each other. What can thus be discerned could no longer be called just ‘an action’. Including actions and more or less decisive involuntary aspects, it is as such neither action nor something happening involuntarily, but rather what could be called the horizon of both, an open field for possibilities and realities at a particular place. The particular state of such a horizon can be called a situation. Situations are nothing to be mastered, since they include aspects that cannot be mastered at all. They only can be understood, and understanding a situation means: realizing a constellation of possibilities and realities one belongs to and is confronted with. Since understanding is involuntary, the clarity of situations cannot be intended and not at all be forced. However, persons can be careful to keep an open-mindedness and a sense of the limits of their actions – not primarily as a sense of restriction, but rather a sense of clearly realizing what under given conditions can be done and what should be allowed just to happen. These considerations directly apply to the aforementioned ideal of free, mature, and responsible persons as it goes along with the Aristotelian conception of action. Contextualizing actions, one has no reason to abandon this ideal, but only to slightly modify it. Contextualized actions, as may be repeated, do not lose the structure Aristotle has so impressively determined. Actions still are to be conceived as aim-oriented activities attempting to realize what has been intended. So, persons who are pursuing something initiated by their decision for a particular aim are responsible for what they are doing or have done. They are – or should be – able to say what they are aiming at, reflect what they are doing, and, if necessary, render an account of what they have effectuated, foreseeable consequences of their actions included. In contrast to responsibility, however, along with contextualizing actions maturity and freedom must be conceived differently. Mature persons then will not only have practical knowledge of such a kind that they will be able to find adequate means for what they are aiming at. They also will be able to discern particular interplays of the voluntary and the involuntary. And, finally, the freedom of persons will not coincide with voluntariness, but rather with the open horizon, the ‘free space’ persons, living their lives, find themselves involuntarily situated in. So, accepting or affirming the involuntary is no resigned fatalism. It is, not least, a recognition of freedom.

Related topics See Chapters 7 (on Reiner), 8 (on Heidegger), 23 (Strawson), and 25 (Drummond).

Abbreviations NE Aristotle, Ethica Nicomachea, recognivit brevique adnotatione critica instruxit I. Bywater, Oxford 1894.

Notes 1 Cf. for instance Plato, Gorgias 466d–468e. 2 For an extensive discussion of the following, cf. Figal (2013), especially pp. 66–74.

396

Involuntariness

397

27 MORAL EXPERIENCE Its existence, describability, and significance Uriah Kriegel

Introduction: three questions In summarizing J.L. Mackie’s (1977) argument for error theory, John McDowell writes: J.L. Mackie insists that ordinary evaluative thought presents itself as a matter of sensitivity to aspects of the world. And this phenomenological thesis seems correct. When one or another variety of philosophical non-cognitivism claims to capture the truth about what the experience of value is like, or (in a familiar surrogate for phenomenology) about what we mean by our evaluative language, the claim is never based on careful attention to the lived character of evaluative thought . . . (McDowell 1985: 110) Jonathan Dancy sounds a similar note: [W]e take moral value to be part of the fabric of the world; taking our experience at face value, we judge it to be the experience of the moral properties of actions and agents in the world. And if we are to work with the presumption that the world is the way our experience represents it to us as being, we should take it in the absence of contrary considerations that actions and agents do have the sorts of moral properties we experience in them. This is an argument about the nature of moral experience, which moves from that nature to the probable nature of the world. (Dancy 1986: 172) Further developing the phenomenological argument for moral cognitivism, Michael Smith writes: . . . we are to argue that our concept of value is the concept of a property that is there to be experienced. The argument for this is to be phenomenological. We are to argue that evaluative experience presents itself to us as the experience of properties genuinely

398

Moral experience

possessed by the objects that confront us. This phenomenological argument is to yield the conclusion that objects seem to have evaluative properties. (Smith 1993: 242) My purpose here is not to evaluate the phenomenological arguments presented by Mackie, McDowell, Dancy, and Smith. My interest is rather in the very notion of phenomenological appeal in moral theory. As the above quotations show, this kind of appeal has a venerable history in analytic moral philosophy. Yet it is only in the past decade or so that moral phenomenology has been explicitly treated (and pursued) as a self-standing area of research. To be sure, it has very much been “thematized” in the phenomenological tradition, but less so in analytic philosophy. As I will understand it here (without prejudging the usefulness of other potential understandings), moral phenomenology is the dedicated study of the experiential dimension of moral mental life, where “experiential” is understood in terms of phenomenal consciousness or what-itis-like-ness (Horgan and Tienson 2005, 2008a; Kriegel 2008). The idea is that while moral philosophy in the analytic tradition has often appealed to certain aspects of our moral life, usually these have concerned the inferential role and representational content of moral mental states; the phenomenal character of moral mental states remains as yet unexplored. Moreover, phenomenological evidence may well bear on some central issues in metaethics and moral psychology, such as cognitivism and noncognitivism about moral judgment (Horgan and Timmons 2006; Kriegel 2012), sentimentalism and rationalism (Gill 2009; Horgan and Timmons 2017), error theory and the objectivistic purport of moral thought and discourse (Loeb 2007; Horgan and Timmons 2008b), and so on. There are, however, three foundational challenges moral phenomenology must overcome before we can take it seriously. In this section, I present the gist of these challenges. In the following ones, I adduce considerations intended to address them. The most basic challenge to moral phenomenology would be the claim that it has no subject matter. The claim can come in two grades. In its strong version, it would be that there are no moral experiences. In its weaker version, it would be that moral experiences do not constitute a natural kind: while there are many individual moral experiences, there is nothing unified about them. (Compare: when philosophers claim that “there is no such thing as emotion,” or that “there are no concepts,” they typically turn out to mean only this: there is no natural similarity, or homogeneity, among the various things designated by “emotion” or “concept.”)1 Even if there is such a thing as (a unified) moral experience, it might be claimed that nothing informative could be said about it. After all, it is a common refrain in discussions of conscious experience that the phenomenal character of an experience, what it is like to have it, is ineffable. It is impossible to make a colorblind person appreciate the phenomenal character of seeing yellow. Such phenomenal character can be named or labeled, but it cannot be described, and therefore cannot be communicated; there is no informative account of it to be had. If so, it is unclear how moral phenomenology could contribute to our theory of moral mental life. Even if it is granted that moral experience exists and is describable, it is unclear what significance it has within our overall moral life. In general, it is widely thought today that conscious experience is but the tip of the mental iceberg: most of what goes on in our mind, determining our behavior and capturing our personality and deepest commitments, goes on below the threshold of conscious experience. If so, understanding the moral dimension

399

Uriah Kriegel

of our mental life requires in the first instance illumination of those more obscured parts, through the patient study of subpersonal cognitive processes, unconscious habits, and so on. Our phenomenological impressions of our moral mental life provide only the most superficial understanding. Moral phenomenology faces a steep challenge, then. To convince us to pursue it, its proponents must provide satisfactory answers to the following three nested questions: a b c

Is there such a thing as moral experience? If there is such a thing as moral experience, can it be informatively described? If there is such a thing as moral experience, and it can be informatively described, is it important for an understanding of moral mental life?

Below, I address all three questions, recommending a positive answer to the first two and sketching a route to a positive answer to the third.

The question of existence: I. Moral experiences Recall that the first, existential challenge to moral phenomenology came in two grades: a strong version claimed that there are no moral experiences, a weaker one that moral experiences do not form a natural kind. In this section, I address the stronger version; in the next, the weaker one. The stronger claim could be undermined by citing moral mental states in which there is something it is like to be. Potential candidates include: 1 2 3 4 5 6 7 8 9

Judging that genocide is wrong. Thinking that I ought to visit my great aunt in hospital. Having the intuition that it is permissible to redirect the trolley. Strongly desiring to meet one’s professional duties in one’s new job. Making the decision to do right by someone. Seeing that what the cat-torturing kids are doing is wrong. Feeling indignant about US police killing another unarmed African-American. Feeling guilty about not helping a blind person cross the street. Feeling feeling deep respect for somebody.

The items on this list can be divided into four groups. First, items 1–3 are cognitive, or intellectual, mental states. As such, many might argue that they have no phenomenal character, and thus do not qualify as experiences in the phenomenal sense. Others, however, will defend so-called cognitive phenomenology (Strawson 1994: ch. 1; Pitt 2004; Chudnoff 2015), and so consider 1–3 genuinely phenomenal.2 Since their contents are clearly moral, involving as they do moral concepts (WRONG, OUGHT, PERMISSIBLE), they would then qualify as proper objects of moral phenomenology. Accordingly, some have developed a phenomenology of moral beliefs (Horgan and Timmons 2007) or of moral intuitions (Bedke 2008). At the same time, many still resist the notion of cognitive phenomenology (Robinson 2006; Carruthers and Veillet 2011), and would presumably reject a phenomenology of moral cognition. Second, items 4 and 5 are “conative” or “motivational” states – states of the will whereby a subject exercises her agency. Traditionally, such states were thought to be best understood in terms of their distinctive functional role in guiding action. More recently, there has been an increasingly lively debate over the existence of a “phenomenology of agency” that outstrips 400

Moral experience

the purely sensory experience of bodily exertion (Bayne 2008; Mylopoulous 2015; but see Ginet 1986 for an early discussion). As in the case of 1–3, there is no question that 4 and 5 are moral mental states; the main bone of the contention is whether they are phenomenal ones. (In the background is also a question about what it takes for a property to be phenomenal – a question I will set aside here, relying entirely on the reader’s intuitive grip on the notion.3) In contrast, item 6 is a perceptual state, so its status as phenomenal is not in question. There are of course cases of unconscious perception (e.g., blindsight), but it is not in question that perception can also occur consciously. Moreover, its content is clearly moral. Traditionally, however, philosophers have been skeptical of the existence of such states as 6, typically on the grounds that moral properties are not sensible: “There is no such thing as a sensation having as its object a quality called moral goodness” (Brentano 1876: 74; see also McBrayer 2010). Presumably, the idea is that strictly speaking what we perceive is just the (nonmoral) supervenience base of moral properties; the moral properties themselves are represented only post-perceptually. For example, although we may describe ourselves as seeing the kindness or generosity shown by one person to another, what is strictly sensible in the exchange are certain “natural,” nonmoral properties of the persons’ behavior, which properties “subvene” kindness or generosity. In recent philosophy, there have been several spirited defenses of genuine moral perception (Harman 1977: ch. 1; Cuneo 2003; McGrath 2004; Audi 2013). Still, this area is too contentious to provide an instance of unquestionably-moral experiences. The dialectical situation here is the opposite of that attending items 1–5: there is no question whether perceptual states can be phenomenal, but there is a question whether they can be moral (whereas with cognitive and conative states there is no question that they can be moral, and the only question is whether they can be phenomenal). From this dialectical perspective, moral emotions – items such as 7–9 – represent the most promising candidate moral experiences. It is widely accepted, both traditionally and in contemporary philosophy of mind, that emotional states can be felt and have a subjective quality, but also that they are often moral. Distinctly moral emotions such as guilt, shame, repentance, indignation, contempt, outrage, resentment, respect, pride, compassion, sympathy, and gratitude are routinely felt and thus occur consciously. This is not yet to say that the phenomenal character of these emotions captures their essence or deep nature; merely that such a phenomenal character exists, and thus constitutes moral phenomenality. It thus seems highly plausible that there exist moral experiences, if only in the form of felt moral emotions, of the sort we are familiar with from first-person acquaintance. In addition, however, there might be moral-perceptual experiences, conscious occurrent moral judgments and intuitions, or experienced moral desires and decisions. It remains that such moral experiences may have nothing in common in virtue of which they are moral experiences. That is, they may not constitute a natural kind. Certainly, this might be the case if some moral experiences are emotional while others are perceptual, cognitive, and/or conative. But even if all moral experiences are emotional experiences, say, it has sometimes been argued that emotions themselves do not form a natural kind (Griffiths 1997); if so, the moral emotions may not either.4 Let us consider this concern next.

The question of existence: II. A phenomenal signature of morality? Suppose a consensus emerges at some point on the full list of moral experience types. Three phenomenological questions would arise immediately. First: •

Is there a phenomenal property φ, such that all moral experiences exhibit φ? 401

Uriah Kriegel

Such a property would represent a phenomenal commonality of moral experiences: something they all share qua moral experiences. A second, complementary question is: •

Is there a phenomenal property Ψ, such that only moral experiences exhibit Ψ?

This would be a phenomenal peculiarity of moral experiences – something distinctive of moral experiences that sets them apart from all other mental states. A third question is: •

Might φ = Ψ?

That is: might there be a phenomenal property which is both common and peculiar to moral experiences, a property that all and only moral experiences exhibit? If there is such a phenomenal property, then for any mental state M, if M exhibits this property, then M is a moral experience, and if M does not exhibit this property, then M is not a moral experience. We may call such a putative property the “phenomenal signature” of morality (or of moral experience).5 If there is a phenomenal signature of morality, then moral experiences constitute a natural kind after all – a kind all of whose members share a “natural” or objective similarity.6 Is there a reason to believe in a phenomenal signature of morality? Persuasive arguments to the contrary have certainly been presented (Sinnott-Armstrong 2008), though capable responses have been offered as well (Glasgow 2013). What I want to argue here is that the question of signature is less crucial to moral phenomenology than might be thought. To see why, consider the fundamental issue surrounding any controversial type of phenomenality. So far, we have treated the fundamental issue as a simple existential question: does the relevant phenomenality exist or not? In reality, however, the question is subtler, inviting a choice between three possible positions. Consider the aforementioned debate over cognitive phenomenology. The debate pits against each other not only (i) proponents of a sui generis cognitive phenomenality characteristic of an experience of thinking and (ii) opponents who claim that there is nothing it is like to think; in addition, there is an intermediate position which (iii) recognizes that there is something it is like to think but attempts to account for it in terms of (a combination of ) already familiar forms of sensory phenomenality (typically: the experience of silent-speech imagery). Position (i) is a form of nonreductivism or primitivism about cognitive phenomenality, Position (ii) is a kind of eliminativism or irrealism about it, while Position (iii) is a sort of reductivist realism. The choice can be appreciated by considering the following inconsistent triad of antecedently attractive propositions: 1 2 3

There is something it is like to engage in conscious cognitive activity (e.g., to think). The phenomenal character of conscious cognitive activity (what it is like to think) is irreducible to the phenomenal character of sensory experiences. The phenomenal character of sensory experiences exhausts all phenomenal character.

To reject (1) in this triad is to embrace what we may call cognitive-phenomenal eliminativism, to reject (2) is to embrace (what we may call) cognitive-phenomenal reductivism, and to reject (3) is to embrace cognitive-phenomenal primitivism. Note, now, that the same three options present themselves in moral phenomenology – regardless of whether there is a phenomenal signature of morality. Consider the following triad:

402

Moral experience

1 2 3

There is something it is like to undergo moral experience. The phenomenal character of moral experience (what it is like to undergo one) is irreducible to the phenomenal character of sensory, cognitive, and motivational experiences. The phenomenal character of sensory, cognitive, and motivational states exhausts all phenomenal character.

Now, in an ambitious cast of mind, we might interpret this triad as concerning a phenomenal signature of morality. It would then amount really to this: 1 2 3

There is a phenomenal signature of moral experience. The phenomenal signature of moral experience (what it is like to have any moral experience) is irreducible to sensory, cognitive, and motivational phenomenal characters. Sensory, cognitive, and motivational phenomenal characters exhaust all phenomenality.

However, even if there is no phenomenal signature of morality, one can intelligibly pose the question of elimination, reduction, or primitivism for individual types of putative moral experience. For example, while an eliminativist might deny that there is anything it is like to experience indignation, and a primitivist might hold that indignation has a sui generis phenomenal character, various reductivists might attempt to reduce the phenomenal character of indignation to a certain combination of negative visceral sensations, conscious thoughts about an injustice, and felt motivation to rectify that injustice. Likewise, we can readily formulate a similar triad for the experience of what Darwall (1977) called “recognitionrespect,” the kind of respect we might feel toward someone not because of her talents and accomplishments, but just because she is a human being: 1 2 3

There is something it is like to feel respect r for a person. The phenomenal character of respecting r (what it is like to feel respect r) is irreducible to (any combinations of ) sensory, cognitive, and motivational phenomenal characters.7 Sensory, cognitive, and motivational phenomenal characters exhaust all phenomenal character.

An eliminativist about the phenomenality of recognition-respect would deny (1), a reductivist would deny (2), while a respect r primitivist would deny (3). The point is that one can debate the correct choice between primitivism, reductivism, and eliminativism about respect r, indignation, guilt, or any other moral experience without first taking a stand on whether there exists a phenomenal signature of morality. Thus, the existence of such a signature is not a precondition for engaging in moral phenomenology. In a way, the issue of phenomenal signature is an issue lying within moral phenomenology, rather than an issue for the very legitimacy or viability of moral phenomenology. I conclude that the existence of a phenomenal signature, which would ratify the status of moral experiences as a natural kind, is immaterial to the viability of moral phenomenology.

The question of characterization: I. Theoretical roles What approach is most suitable for the phenomenological characterization of some moral experience depends on what view one takes of that experience’s phenomenal character. If one is an eliminativist about it, then obviously, one takes it that there is no phenomenality in need

403

Uriah Kriegel

of characterization. If one is a reductivist, however, a very natural approach to phenomenological characterization presents itself: simply list all phenomenal elements the combination or compresence of which constitutes the relevant moral experience’s phenomenal character. Suppose one holds, for example, that the phenomenal character of indignation reduces to the following collection of sensory, cognitive, and motivational components: (i) a sensation of “boiling blood” and flapping nostrils, (ii) the judgment that an injustice has been committed, and (iii) a felt pull to “do something about it.” Then by comprehensively listing those elements, one would be effectively characterizing what it is like to feel indignation. A more delicate question arises within the nonreductivist or primitivist framework. If the phenomenal character of recognition-respect, say, is simple and unanalyzable, how could we offer a substantive and informative description of it? It is for such primitive phenomenal properties, after all, that the claims of ineffability and incommunicability are most plausible. The problem is not special to moral phenomenology – any phenomenological inquiry faces it. But it concerns us specially because of the hope that moral phenomenology will somehow inform moral theory. There are two general approaches to the problem. The first attempts to characterize primitives in some area of inquiry by their theoretical role within the dominant theory in the area. Consider an analogy from mathematics. The notions appearing in the theorems of a given axiomatization of Euclidean Geometry are defined in terms of notions appearing in the system’s axioms.8 But the notions appearing in the axioms are understood in terms of their role within these axioms (Hilbert 1900).9 We can think of the axioms as describing a web of interrelations among nodes, with each node designated by a different primitive notion. While the meaning of theorems’ notions is captured by definitions in terms of axioms’ notions, the meaning of the latter is exhausted by the interrelations specified in the axioms. A technique for regimenting these theoretical roles is the Ramsey sentence (Lewis 1966, 1972). A simple Ramsey sentence for a concept C in a theory T is produced by (i) collecting all of T’s C-employing assertions, (ii) making a long conjunction of these, (iii) replacing occurrences of “C” and cognates with a variable, and (iv) prefacing the conjunction with an existential quantifier. This approach can be applied to a phenomenological theory mentioning the experience of recognition-respect, say, just as much as to a geometric theory employing the notion of a point. The following seem to be phenomenological observations about recognition-respect as we experience it pre-theoretically, culled (essentially verbatim) from the Stanford Encyclopedia entry on respect (Dillon 2014): • • • • • • •

Respect is a particular mode of apprehending the object: the person who respects something pays attention to it and perceives it differently. Respect often means trying to see the object clearly, as it really is in its own right, and not seeing it solely through the filter of one’s own desires and fears or likes and dislikes. Respect is object-generated rather than wholly subject-generated, something that is owed to, called for, deserved, elicited, or claimed by the object. Respect is also an expression of agency: it is deliberate, a matter of directed rather than grabbed attention, of reflective consideration and judgment. Respect involves “a deontic experience” – the experience that one must pay attention and respond appropriately. We respect something not because we want to but because we recognize that we have to respect it. Respect is the recognition of something “as directly determining our will without reference to what is wanted by our inclinations” (Rawls 2000: 153).

404

Moral experience

• •

Respect is thus reason-governed: we cannot respect a particular object for just any old reason or for no reason at all. Respect is universalizing, in the sense that if F is a respect-warranting feature of object O, then respecting O on account of F commits us, other things equal, to respecting other things that also have feature F.

We may add further texture to our notion of respect r experience by consulting some of Kant’s pertinent observations: • •



“Respect is properly the representation of worth that infringes upon my self-love” (Kant 1785/1997: 14). “. . . a human being regarded as a person, that is, as the subject of a morally practical reason, is exalted above any price [and] exacts respect for himself from all other rational beings in the world. . .” (Kant 1797/1996: 186–187). “Respect for the [moral] law, which in its subjective aspect is called moral feeling, is identical with consciousness of one’s duty” (Kant 1797/1996: 210).

A comprehensive phenomenological theory of the experience of respect r would involve a great many observations of this sort, which could then be “Ramsified” to capture the theoretical role of respect r experience in the theory. This would take the form: There is an x, such that x is a particular mode of apprehending the object & the person who xs something pays attention to it and perceives it differently & x often means trying to see the object clearly, as it really is in its own right & x is object-generated rather than wholly subject-generated & x is something that is owed to, called for, deserved, elicited, or claimed by the object & . . . In this way we can provide an informative account of moral experiences’ phenomenal character, even when the latter is primitive and irreducible. We do so by specifying the primitives’ theoretical role as precisely as possible – just as the mathematician does with her own primitives. Various technical issues arise here. First, phenomenological theory is bound to be much less accurate than mathematical theory, with falsehoods sure to creep into our theory of respect r experience; any such error would appear to entail the falsity of the corresponding Ramsey sentence. Fortunately, Lewis (1972: 256) raises this problem and solves it in half a sentence: instead of forming a long conjunction of all observations about respect r experience, we may form an even longer disjunction of conjunctions of most of these observations. With this device in place, the truth of this kind of more flexible Ramsey sentence is guaranteed if a sufficiently large portion of our phenomenological theory is true (see also Jackson 1998: 35). Second, in contrast to mathematics, disagreements plague phenomenological theory, and are specially stubborn – consensus in this area is elusive (Schwitzgebel 2008). This includes the domain of moral phenomenology, where it is impossible to rule out that people’s underlying moral experiences simply differ in phenomenal character (Gill 2008). This problem can be set aside, however, by simply quarantining specific areas of disagreement (Bayne and Spener 2010). In practice, this would mean leaving out of our Ramsey sentence overly contentious phenomenological claims – or at least including those in far fewer disjuncts of the more flexible Ramsey sentence just described. Third, and perhaps underlying the previous two

405

Uriah Kriegel

problems, there seems to be no accepted methodological organon we could use in phenomenological theory-construction. Here I think the proponent of moral phenomenology would be wise to concede that moral phenomenology cannot aspire to the levels of precision, rigor, and consensus that mathematics, and for that matter mainstream analytic philosophy, do. Brutally put, moral phenomenology is an enterprise of the humanities, with their disappointingly approximative and often qualitative (as opposed to quantitative) methods. All the same, if those methods have the potential to deepen our understanding of reality, including moral reality, there is no reason to pretend they are not available to us.10

The question of characterization: II. Contrastive introspective revelation This, then, is the first approach to the characterization of phenomenal primitives: specifying their theoretical role within the relevant phenomenological theory. The second approach is what we might call contrastive introspective revelation. Let me explain this approach in three steps. According to the revelation theory of color, we do not come to understand the nature of colors by appreciating the right philosophical (or other) theory ( Johnston 1992). For example, we cannot grasp the nature of blue by digesting an objectivist theory in terms of reflection and refraction properties, a dispositionalist theory in terms of the tendency to elicit bluish experiences, or any other theory. Rather, we grasp the nature of blue by looking at the sky on a clear day with a properly functioning visual system. When we look at a paradigmatic color in the right conditions, and everything goes well, we are acquainted with the nature of that color. The basic idea in the background – profound for philosophers but peradventure obvious otherwise – is that the intellect (“reason”) is not the only faculty through which we may grasp the deep nature of a property. Sometimes the eyes can disclose the nature of a property better – to provide insight into it in the original sense of the term. And other faculties may do so in other cases. This basic idea applies rather naturally to phenomenal properties. Perhaps the intellect is best positioned to disclose the nature of phenomenal properties that arise from combinations of more basic ones. But when it comes to the most basic, elemental phenomenal properties, it is rather direct introspective encounter that best positions us to grasp the nature of the property. We come to appreciate the nature of the property of being bluish (I use “bluish” technically to refer to the phenomenal property that corresponds to blue), for example, when we introspect a paradigmatically bluish experience and everything goes well. In general, sui generis phenomenal primitives appear to lend themselves most straightforwardly to grasping through introspection. When a phenomenal primitive is introspectively salient, as tends to be the case with bluishness and pain, it is fairly clear how an “introspective revelation” theory works. But asked to directly introspect the nature of a putatively sui generis phenomenal character of respect r, say, most of us are likely to feel stumped: it is not clear “where to look” and “what to do.” For such relatively nuanced and elusive phenomenal properties, it appears crucial that we contemplate a variety of phenomenal contrasts that foreground the specific phenomenal primitive we are interested in and put it in sharper relief. Thus, there are several types of experience that have something in common with, but also something crucially different from, indignation: notably anger and frustration, but also hurt, sorrow, and even surprise (indignation is typically elicited by something that in some sense we do not expect). To appreciate the nature of the phenomenal character of indignation, we must contemplate, side by side as it were, a paradigmatic experience of indignation and a paradigmatic experience 406

Moral experience

of anger, a paradigmatic experience of indignation and a paradigmatic experience of frustration, and so on. Through triangulation from a sufficiently rich set of contrasts, an ever clearer and more acute introspective insight into the distinctive phenomenal character of indignation should arise. It is the moral phenomenologist’s task to provide the kinds of contrast that would bring out subtle moral-phenomenal features. It is worth distinguishing here between a dialectical use of phenomenological contrasts, which serves to argue for the existence of some phenomenal features (see Siegel 2007), and a merely ostensive use, which serves only to focus the mind on these features (Koksvik 2011). My claim is that moral phenomenology can, at the very least, avail itself of phenomenological contrasts in their ostensive capacity. The goal is to focus the mind on any phenomenal primitives moral experience might involve. The exercise is still a theoretical exercise, insofar as we cannot appreciate such contrasts very well when in the clutches of indignation, say. We must do so rather by contemplating remembered or imagined experience pairs (see, e.g., Horgan and Timmons 2017 for a phenomenological contrast between indignation and frustration). We must then hold in our mind, so to speak, all the relevant contrasts at once, and come to “see” (read: grasp) that distinctive quality of indignation experience. The claim that we can do so, at least in some cases, is the core thesis of the contrastive introspective revelation approach to the appreciation of phenomenal primitives. Ultimately, the hope is that between (i) the cases lending themselves to such contrastive introspective revelation and (ii) cases in which a sufficiently articulated phenomenological theory allows us to specify a reasonably textured theoretical role, all phenomenal primitives will be “covered.” Each of these two approaches has its own advantages and disadvantages. Contrastive introspective revelation is in some sense more fundamental, as it acquaints us directly with our subject matter. Indeed, one suspects that the theoretical role approach is at least partially parasitic on it, insofar as the development of phenomenological theory relies in part on introspective encounter and contrast. Characterization by theoretical role also raises the specter of permutability problems of the sort discussed in the literatures on structuralism in the philosophy of science (Demopoulos and Friedman 1985) and on global descriptivism in the philosophy of language (Lewis 1984): the worry that two properties with distinct intrinsic natures may have perfectly symmetric theoretical roles within a theory, allowing for permutation of intrinsic natures without perturbation of theoretical roles (see Newman 1928). The distinctness of their intrinsic natures would then be detectable only in direct introspective grasp. To that extent, contrastive introspective revelation must be the ultimate foundation of moral phenomenology, providing the phenomenologist with original insight into the nature of phenomenal primitives. However, insofar as moral phenomenology is supposed to be a potentially fecund research area, pursued by a community of inquiry, and not just a private exercise, it would also be useful to have accounts of its subject matter statable in public language. The characterization of phenomenal primitives in terms of theoretical role allows for this and is in that respect superior to sheer revelation.

The question of significance: moral experience and moral life As noted in the introduction, it is widely accepted today that conscious processes form the tip of the mental iceberg, and that most mental life – including our deepest behavioral dispositions and the habits most intimately tied to our personality – takes place below the surface of conscious experience. But unconscious processes are not only a crushing majority of mental processes. They are also developmentally and evolutionarily prior to conscious processes. The latter are thus largely derivative phenomena, albeit ones particularly striking from the 407

Uriah Kriegel

first-person perspective. Ontogenetically and phylogenetically, they are Johnny-come-latelys rather than foundational phenomena. If so, their study would seem to be of relatively secondary importance for a deep understanding of our moral mental life. This problem is not special to moral phenomenology, of course; it affects all phenomenology equally. This is not the place to argue for a place for phenomenological inquiry in our attempts at a full understanding of our mental life. Instead, I am going to merely describe the kind of position which, I believe, is the phenomenologist’s best approach to responding to the challenge. As we will see, the issues raised by this approach are far too complex to be handled satisfactorily here; this section should therefore be seen as an exercise in presenting a strategy that needs to be pursued seriously elsewhere. My goal here is thus not to neutralize the challenge, but to describe the response to it that I would recommend to the moral phenomenologist. The general shape of the response is this: while various orders of causal priority flow from unconscious to conscious processes, there is an order of conceptual priority that proceeds in the opposite direction. Below I describe three instances of this general strategy in contemporary philosophy of mind. Goldman (1993) argues that our ascriptions of mental states to others on the basis of their behavior are parasitic on prior introspective ascription of conscious states of the same type to oneself. For example, we ascribe sadness to a friend on the basis of the friend’s behavior, but only against the background of (i) grasping the nature of sadness on the basis of introspective encounter with our own sadness experiences and (ii) awareness of the kinds of behavior we typically engage in when experiencing sadness. This seems to ground our mastery of mental concepts in introspective acquaintance with conscious states falling in these concepts’ extensions. In a similar vein, Searle (1990, 1992: ch. 7) has argued for a “connection principle” between unconscious and conscious intentional states: an unconscious state is intentional only if it is potentially conscious. Thus, an unconscious wish that p is the intentional state it is, and an intentional state at all, only because it is disposed to become a conscious wish that p. For Searle, an unconscious mental state that cannot in principle become conscious is impossible. Although Searle himself does not say so, one might suspect that this condition is built into the very concept of an unconscious mental state. Another view in the same ballpark casts the concept of the mental as a prototype concept, and then claims that all the relevant prototypes are conscious experiences (Horgan and Kriegel 2008; Kriegel 2015: ch. 4). An example of a prototype concept familiar from cognitive psychology is the concept of furniture: for something to qualify as a piece of furniture, it must resemble sufficiently (and relevantly) prototypical pieces of furniture; and the prototypical pieces of furniture are, it appears, tables and chairs (Rosch 1975). The hypothesis under consideration is that, likewise, for something to qualify as a mental state it must resemble sufficiently (and relevantly) prototypical mental states – and that the prototypes here are certain conscious experiences (presumably, the most phenomenally striking ones). In all these accounts, the status of an unconscious mental state as mental depends on its relation to some conscious mental state(s). Whether the relation is one of parasitic-ascribability, potential-becoming, or (relevant) similarity, the core claim is that our grip on unconscious mentality is somehow dependent upon a prior grip on conscious mentality. Assuming that mental states are realized in brain states, but that not all brain states realize mental states (some states of the brain have no mentalistic significance), we might ask what the difference is between the brain states that do realize mental states and those that do not. The above approaches answer in unison: a brain state realizes a mental state when it either (a) realizes

408

Moral experience

a conscious experience, that is, supports the instantiation of phenomenal properties, or (b) realizes a state that bears the right relation to conscious experiences. It is in this sense that conceptual priority may flow from conscious to unconscious mentality even if unconscious mentality clearly enjoys developmental and evolutionary priority. Paul Ricœur (1950: 22) made the same point in his inquiry into the phenomenology of the will: causal processes underlying the operation of the will always proceed from the bottom up, that is, from subpersonal micro-processes to conscious macro-processes; but understanding of their nature tends to proceed from the top down, illuminating those subpersonal micro-processes in terms of their role in underwriting the conscious macro-processes of which we have immediate first-person insight. A parallel point can be made about moral mental life. Surely the causal processes subserving the experience of indignation are subpersonal and unconscious, and perhaps creatures can enter unconscious indignation states at various stages of ontogenetic and phylogenic development at which they are as yet unable to experience indignation. It remains that we naturally understand and classify those unconscious states as indignation, rather than as anger or as frustration, because of what it would be like to experience them if one underwent them consciously. In the background may be two different kinds of curiosity, motivating two different forms of understanding. Of any phenomenon, we may ask “How did it come to be?” but we may also ask “What is it?” The former question courts a causal answer, which tends to proceed from the bottom up, from micro to macro, from part to whole, from process onset to process product. But the second question is completely independent and can easily proceed in the opposite directions. The point is that an answer to the “How did it come to be?” question does not automatically deliver an answer to the “What is it?” question. Granted, sometimes the best answer to “What is it?” is “It is the kind of thing that results from this kind of causal process.” That is, some phenomena owe their identity conditions to their causal origins. But surely many phenomena do not. Thus, when we ask of a mental phenomenon “What is it?” an antecedently promising answer is “It is the kind of thing whose phenomenal character is such-and-such when it occurs consciously.” Even if most indignation phenomena are unconscious and subpersonal, for example, arguably our pretheoretic, original grip on what they are relies on this kind of answer, that is, on our first-person insight into what it is like to experience indignation consciously. Sensitive to this distinction, Brentano (1890/1) distinguished between genetic and descriptive psychology. The former provides causal explanation of the genesis of mental phenomena; the latter merely describes the phenomena. (That is, the former answers questions of the form “How did it come to be?,” the latter questions of the form “What is it?”) Just by drawing this distinction, we can appreciate the manner in which descriptive psychology is prior in the order of understanding to genetic (or “explanatory”) psychology: in an “ideal reconstruction” of science, we would presumably proceed first by describing the phenomena in need of explanation and only then offer an explanation of them.11 Without knowing what “it” is, it is hard to see how we might be able to explain how “it” came to be. And I have suggested that, in the psychological domain (including in moral psychology), the answer to the “What is it?” question often takes the form “It is the kind of thing that has such-andsuch phenomenal character when it occurs consciously.” If so, the very description of the phenomena moral psychology is expected to explain presupposes moral phenomenology. It is worth noting, in this context, that the difference between descriptive and explanatory psychology seems to mirror that between philosophical and scientific understanding: where science answers, in the first instance, causal-explanatory questions of the form “How

409

Uriah Kriegel

did it come to be?,” philosophy tends to focus primarily on questions of essence and identity, questions which often take the form “What (kind of thing) is it?” To that extent, we can see that moral phenomenology’s organizing question places it squarely within the project of moral philosophy. To repeat, this section was only an exercise in describing an approach the phenomenologist might pursue in addressing the challenge of significance. I have not argued for this approach, because the issues it raises are far too vast to admit of serious treatment in the present context.

Conclusion If we want to understand the moral realm, it would be an intellectual missed opportunity to simply ignore our first-person insight into our moral experience. I have belabored the feasibility and relevance of a phenomenology of moral experience mostly because the notion is so foreign to analytic moral philosophy. But it is worth noting that notions in the vicinity are virtually taken for granted in core central-European traditions in moral philosophy (see Brentano 1889; Meinong 1894; Ehrenfels 1897; Scheler 1913; Mandelbaum 1955 inter alia). When reading this material, even with an analytic eye, it is striking just how rich and sophisticated the pursued inquiry into moral reality is. I have argued here that there are no foundational reasons to be skeptical about this kind of inquiry. First of all, it is clear that moral phenomenology has a subject matter: there clearly exist moral experiences, whether or not they share a phenomenal signature of morality. Second, what it is like to undergo a moral experience is describable in a theoretically useful manner, and certainly directly graspable (with the aid of the right contrasts), even when primitive and irreducible. Finally, a descriptive rather than explanatory emphasis in moral-psychological inquiry casts moral phenomenology as potentially centrally relevant to moral theory.12

Related topics See Chapters 19 (Horgan and Nida-Rümelin) and 23 (Strawson).

Notes 1 I will discuss this kind of view of emotions in the second section. For the case of concepts, see Machery (2009). 2 Strawson (1994), for example, argues that there is a phenomenal difference between listening to the news in French as someone who understands French and listening to them as someone who does not understand French. The difference, he claims, is in what he calls “understanding- experience”: the experience of grasping the propositional content conveyed by the relevant sounds. 3 In the literature, a certain movement may be discerned toward explicating phenomenality in terms of the characteristic intellectual puzzlement raised by the phenomenon of consciousness: phenomenality is that which is susceptible to an explanatory gap, or to zombie scenario, or the knowledge-argument reasoning, or the like (see Bayne 2009; Carruthers and Veillet 2011; Horgan 2011; Kriegel 2015: ch. 1). 4 In addition, it might be claimed, post-Freud, that deep understanding of the very nature of emotions is not gained by attending to their conscious manifestations, but on the contrary by studying the behavioral manifestations of the unconscious dispositions they embody. This is to raise again the worry captured in question (c) in the first section. This question will be discussed toward the end of the paper. 5 We can also speak of an “approximate phenomenal signature” or a “phenomenal near-signature” in case there is a phenomenal property that most moral experiences exhibit and almost all of them do.

410

Moral experience

411

Uriah Kriegel Demopoulos, W. and M. Friedman (1985) “Bertrand Russell’s The Analysis of Matter: Its Historical Context and Contemporary Interest,” Philosophy of Science 52, 621–639. Dillon, R.S. (2014) “Respect,” Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/ respect/ Ehrenfels, C.v. (1897) System der Werttheorie, Leipzig: O.R. Reiland. Ellis, J., M.K. Gaillard, and D.V. Nanopoulos (1976) “A Phenomenological Profile of the Higgs Boson,” Nuclear Physics B 106, 292–340. Gill, M.B. (2008) “Variability and Moral Phenomenology,” Phenomenology and the Cognitive Sciences 7, 99–113. Gill, M.B. (2009) “Moral Phenomenology in Hutcheson and Hume,” Journal of the History of Philosophy 47, 569–594. Ginet, C. (1986) “Voluntary Exertion of the Body: A Volitional Account,” Theory and Decision 20, 223–245. Glasgow, J. (2013) “Does Direct Moral Judgment Have a Phenomenal Essence?,” Journal of Moral Philosophy 10, 52–69. Goldman, A.I. (1993) “The Psychology of Folk Psychology,” Behavioral and Brain Sciences 16, 15–28. Griffiths, P. (1997) What Emotions Really Are, Chicago, IL: University of Chicago Press. Harman, G. (1977) The Nature of Morality, Oxford: Oxford University Press. Hilbert, D. (1900) “Über den Zahlbegriff,” Jahrbericht der deutschen Mathematiker-Vereinigung 8, 180–183. Horgan, T. (2011) “From Agentive Phenomenology to Cognitive Phenomenology: A Guide for the Perplexed,” in Bayne and Montague (eds.) (2011). Horgan, T. and Kriegel. U. (2008) “Phenomenal Intentionality Meets the Extended Mind,” Monist 91, 347–373. Horgan, T. and M.C. Timmons (2005) “Moral Phenomenology and Moral Theory,” Philosophical Issues 15, 56–77. Horgan, T. and M.C. Timmons (2006) “Cognitivist Expressivism,” in T. Horgan and M.C. Timmons (eds.), Metaethics after Moore, Oxford: Oxford University Press. Horgan, T. and M.C. Timmons (2007) “Moorean Moral Phenomenology,” in S. Nuccetelli and G. Seay (eds.), Themes from G.E. Moore. Oxford: Oxford University Press. Horgan, T. and M.C. Timmons (2008a) “Prolegomena to a Future Phenomenology of Morals,” Phenomenology and the Cognitive Sciences 7, 115–131. Horgan, T. and M.C. Timmons (2008b) “What Does Moral Phenomenology Tell Us about Moral Objectivity?,” Social Philosophy and Policy 25, 267–300. Horgan, T. and M.C. Timmons (2017) “Sentimentalist Moral-Perceptual Experience and Its Realist Pretentions: A Phenomenological Inquiry,” in K. Stueber and R. Debes (eds.), Moral Sentimentalism. Oxford and New York: Oxford University Press, pp. 66–85. Jackson, F.C. (1998) From Metaphysics to Ethics: A Defence of Conceptual Analysis, Cambridge, MA: Cambridge University Press. Johnston, M. (1992) “How to Speak of the Colors,” Philosophical Studies 68, 221–263. Kant, I. (1785/1997) Groundwork of the Metaphysics of Morals, trans. M. Gregor, Cambridge, MA: Cambridge University Press. Kant, I. (1797/1996) The Metaphysics of Morals, trans. M. Gregor. Cambridge, MA: Cambridge University Press. Koksvik, O.A.K. (2011) Intuition. PhD Dissertation, Australian National University. Kriegel, U. (2008) “Moral Phenomenology: Foundational Issues,” Phenomenology and the Cognitive Sciences 7, 1–19. Kriegel, U. (2012) “Moral Motivation, Moral Phenomenology, and the Alief/Belief Distinction,” Australasian Journal of Philosophy 90, 469–486. Kriegel, U. (2015) The Varieties of Consciousness, Oxford and New York: Oxford University Press. Lewis, D.K. (1966) “An Argument for the Identity Theory,” Journal of Philosophy 63, 17–25. Lewis, D.K. (1972) “Psychophysical and Theoretical Identifications,” Australasian Journal of Philosophy 50, 249–258. Lewis, D.K. (1983) “New Work for a Theory of Universals,” Australasian Journal of Philosophy 61, 343–377. Lewis, D.K. (1984) “Putnam’s Paradox,” Australasian Journal of Philosophy 62, 221–236. Loeb, D. (2007) “The Argument from Moral Experience,” Ethical Theory and Moral Practice 10, 469–484.

412

Moral experience McBrayer, J. (2010) “Moral Perception and the Causal Objection,” Ratio 23, 291–307. McDowell, J. (1985) “Values and Secondary Qualities,” in T. Honderich (ed.), Morality and Objectivity, Boston, MA: Routledge & Kegan Paul. McGrath, S. (2004) “Moral Knowledge,” Philosophical Perspectives 18, 209–228. Machery, E. (2009) Doing Without Concepts, Oxford and New York: Oxford University Press. Mackie, J.L. (1977) Ethics: Inventing Right and Wrong, New York: Penguin. Mandelbaum, M. (1955) The Phenomenology of Moral Experience, Glencoe, IL: The Free Press. Meinong, A. (1894) Psychologisch-ethische Untersuchungen zur Wert-Theorie, Graz: Leuschner & Lubensky. Mylopoulous, M. (2015) “Agentive Awareness Is Not Sensory Awareness,” Philosophical Studies 172, 761–780. Newman, M.H.A. (1928) “Mr. Russell’s Causal Theory of Perception,” Mind 37, 137–148. Pasch, M. (1882) Vorlesungen über neuere Geometrie, Leipzig: Teubner. Peano, G. (1889) Arithmetices principia, nova methodo exposita, Turin: Bocca. Pieri, M. (1900) “Della geometria elementare come sistema ipotetico deduttivo,” Monografia del punto e del moto. Memorie della Reale Accademia delle Scienze di Torino 49, 173–222. Pitt, D. (2004) “The Phenomenology of Cognition; or What Is It Like to Think that P?,” Philosophy and Phenomenological Research 69, 1–36. Rawls, J. (2000) Lectures on the History of Moral Philosophy, Cambridge, MA: Harvard University Press. Ricœur, P. (1950) Le volontaire et l’involontaire. Paris: Aubier. Reprinted as Philosophie de la volonté: Le volontaire et I’involontaire. Paris: Editions Points, 2009. Translated as Freedom and Nature, trans. E. Kohak 1966. Evanston, IL: Northwestern University Press. Robinson, S.W. (2006) “Thoughts Without Distinctive Non-Imagistic Phenomenology,” Philosophy and Phenomenological Research 70, 534–561. Rosch, E.H. (1975) “Cognitive Representations of Semantic Categories,” Journal of Experimental Psychology 104, 192–233. Scheler, M. (1913/1973) Formalism in Ethics and Non-Formal Ethics of Values, trans. M.S. Frings and R.L. Funk. Evanston, IL: Northwestern University Press. Schwitzgebel, E. (2008) “The Unreliability of Naïve Introspection,” Philosophical Review 117/2, 245–273. Searle, J.R. (1990) “Consciousness, Explanatory Inversion and Cognitive Science,” Behavioral and Brain Sciences 13, 585–642. Searle, J.R. (1992) The Rediscovery of Mind, Cambridge, MA: MIT Press. Sider, T. (2011) Writing the Book of the World, Oxford and New York: Oxford University Press. Siegel, S. (2007) “How Can We Discover the Contents of Experience?,” Southern Journal of Philosophy (supplement) 45, 127–142. Sinnott-Armstrong, W. (2008) “Is Moral Phenomenology Unified?,” Phenomenology and the Cognitive Sciences 7, 85–97. Smith, M. (1993) “Objectivity and Moral Realism: On the Significance of the Phenomenology of Moral Experience,” in M. Smith (ed.) (2004) Ethics and the A Priori. Selected Essays on Moral Psychology and Meta-Ethics (Cambridge Studies in Philosophy), Cambridge, MA: Cambridge University Press, pp. 234–258. Strawson, G. (1994) Mental Reality, Cambridge MA: MIT Press.

413

INDEX

Note: Italic page numbers refer to figures and page numbers followed by “n” denote endnotes. act(s): axiological positionality 370, 371; being as consciousness 138; doxic positionality 370–371; engagement, intentionality, and motivational properties 140–142; essentially interrelated moments 137; first-order positions 370, 371; On haunting, projections, and engagement 139–140; hierarchy of 370–371; kinds of 367–368; legal sense 368; l’en-soi, le pour-soi, and the emergence of the value 137–139; l’être parfait 138–140; limits of self-knowledge, freedom, and determinism 143–144; lived intentionality 142; motivational qualities 140; non-positional projective intending 140, 142, 144–145; notion of haunting 139; ontological framework: Sartre’s dualism and his notion of consciousness as néant 135–137; perfect being 138; positional acts 368; positional intending 140, 142; problem of action 142; property of being-to-be-emptied 141; psycho-physical determinants 144; reflective self-ascriptions 143–144; Sartre’s primary interest 144–145; self-determination/self-apprehension 137–138, 144; “structures contained within the very idea of acting” 142–143; theory of 366–367 action: agency, sense of 336–338; being able, sense of 47; bodily expressions 378–379; compossible value-contents 47; conation 44–45, 46; concern-based construals 378; content of efforts 47; desire/wish 379; elements in action/deed 47; ethical personalism 48; feeling- and emotionexperiences 377–379; fiat, notion of 377; first-order/pre-reflective experience 336, 338; goals 45, 46; incomplete actions 378; inference/second-order reflective

consciousness 337; kinematic processes 338–339; moral content 46–47; moral evaluation 49; naïve action explanation 377; ownership, sense of 336, 337; performance dimension 48, 378; presence of situation and object of deed 47; purposes 45–46; states and feelings responding to ends 48; states of sensation and feeling 48; voluntary action 376–377, 379–380; willing 379; willing-to-do 47–48 active intentional experiences 55; desiderative/ willing states 55; emotions 55, 64n9 active life, phenomenological approach to 160–162; analysis of lived experience 162; appearing and being, coincidence of 161; deconstructive method 161; “the human affairs” 160; negativa/negative approach 160; phenomena/appearances 160; practice of discrimination 161–162; prejudices and fallacies 161; two world theory 160 agency: across human activities 163–165; approaches to 102–105; attitudinal theory 209, 216–217; cognitive science (see cognitive science of sense of agency); definition 239; first-person 151; free (see free agency, phenomenology of ); human activities 163–165, 220; illusory experience 340; and intention (see intentional/ intentionality); judicial (see judicial agency); long-term sense 346; moral 49–50; notion of 168; phenomenology of 332; plural 159; political 158, 165–166, 169–171; position in classical phenomenology of 39; practical 317, 318, 325; rational (see rational agency); as selfefficacy 209–211; sense of 336–338; skillful

415

Index coping 221; social context 105; two floors model of 99; views on 96; volitional 88–90 agentive phenomenology 30; agreements, disagreements, and conflicting background assumptions 295–297; Horgan, proposals (see Horgan, T.); intentional and intentionality 297n2; Nida-Rümelin 267–272, 279–281; satisfaction conditions of color experience 298n8 ambiguous body: dual aspect theory 179; genesis of the objective body 177; ideal of objective thought 177–178; naturalism 180; phenomenal body 179; phenomenal powers 179; roles of touching and touched 178–179; theme of impoverishment 178 Ambulo!: action and awareness-of-action 300; from cognitive to conative to volitional phenomenology 305–307; complex structure of action 300–302; conation/conatus 306–307; conative/agentive experiences 306; embodied movement and consciousness 303; “fiat!” character of volition in action 307–308; fusion of sensory and noetic 301; issue of cognitive phenomenology 302; modal model of volition in action 310–311; my lived body vs. my physical body 302–303; noematic structure in volition 308–310; phenomenal/ phenomenological character 305–306; phenomenological and ontological structures 302; phenomenology vs. ontology of my action in walking 303–305; primitive conatus 306, 307; toward a formal phenomenological ontology of action 311–312; volition effects 300–301 Anscombe, E. 190, 191, 192, 202, 240; conceptual impressionism 192; Intention 202; knowing-that (viz. event) and knowing-how (viz. action), distinction between 191 Apel, Karl-Otto 209 Aquinas, T. 88, 214 Arendt, H. 158; acting, notion of 213; agency across human activities 163–165; agency, notion of 168; animal laborans, laboring animal 158, 163–164, 170; case of homo faber 164, 167, 170; citizens, power of 172n1, 173n21; doer and deed 167–169; freedom and power 169–171; Genealogy of Morality 167; The Human Condition 159, 168; labor, work, and action 159, 165; The Life of the Mind 159, 168; metaphor of illumination 165; nonsovereignty of political actor 159, 165–167; The Origins of Totalitarianism 169; people’s inner life 163–164; phenomenological approach to active life 160–162; plural agency 159; plurality 158–159, 162–163; political agency 158, 169–171; prejudice 167; realism, phenomenological kind of 171–172; reality effect of plural action 165; spontaneous

rebellions 171; use of violence in politics 170–171; velvet revolutions of 1989 171; zoon politikon 164–165 Aristotle 88, 250, 291, 388; on action 388–389; concept of habit second nature 11; constraint/ ignorance 389; gestures and facial expressions 390; ideal of free, mature, and responsible personality 389; involuntary 389, 390; mixed actions 390; Nicomachean Ethics (NE) 242, 388; Poetics 236; teleological character of actions 388; theory of action, advantages 388–389; on voluntary and involuntary 389–391 attitude(s) 77n6; affective involvement 212–213; case of Phileas Gage 216; cognitive experiences 71–72; conative 73; doxastic, affective, or conative 70; ethos and habitus, notions of 214; as form of self-efficacy 214–216; independent and self-efficacious 214; intelligible character 214; and knowledge 72; mineness, absence of 215; notion of 212–214; objective facts 216; practices of praising and blaming 213–214; response, contents (Gehalte) of attitudes 71; self-efficacy, conditions for 215; subjective factuality 214–215; virtue of being responses, specification 71–72; volitions 70 Austin, J. 62, 191, 203, 366 authorship, sense of 53, 54 being-in-the-world: agent and world 223; for-the-sake-of-whichs 223–224; human existence 223; mind and world, and context 222–224; skillfull coping 224 belief-desire-intention (BDI) model 365, 366 belief-desire (BD) model 365, 366 bodily intentionality: concrete and abstract movement 181; future arm movement 182; level of personal existence 180–181; personal freedom 181; phenomenon of a phantom limb 180; physical injury 180 Bratman, M. 366 Brentano, F.: absolute accident, concept of 9, 10; abstract phrasings, use of 8–9; actualism 8; actus a voluntate imperatus 7, 9; actus elicitus voluntatis 7, 9; asceticism 11–12; condition for freedom 7; critique of free will; descriptive and genetic psychology 12n2; determinism, demonstration of 9–10; determinism-indeterminism debate 7, 9; extreme and moderate indeterminism 10; fatalism 11; fathers of phenomenological tradition 7; The Foundation and Construction of Ethics 7; indeterminism 8, 13n7; inner perception 9–10; law of causality 10, 13n7; necessitarianism 8, 13n7; The Origin of Our Knowledge of Right and Wrong 7; outer experience 9; Psychology 7; reductio ad absurdum, form of 9; self-improvement, forms

416

Index of 11–12; sense of being free from necessity 9; sophism, motive-cause distinction 10; theory of value and will 7; volitions, account of 7 causal self-referentiality 366 causation: agent causation 36; agentive experience 277; agentive phenomenology 278; appearance of change 274; as-of non-agentive state causation 278; as-of state causation 276–277; construals 275–276; content-based satisfaction 277; Emotive Causation Thesis 316; envisioned experiences 276; evolution and natural selection 278; first putative consequence 275; forms of 213; Intentional Causation Thesis 316; motivation 35; non-agentive state-causal process 276; non-agentive state causation 277; physical state-causal completeness 278–279; psychological state causation (PSC) 267; state causation simpliciter 278 Chisholm, R. 36, 104 choice: emotion-motivated behavior 380; intentional structure 380–381; intentionsin-action 381–382; prior intention-to-do 381–382; voluntary action 380, 381 “Classical Model of Practical Rationality” 365 cognitions: and attitudes, distinction between 69; cognizing/knowledge (Erkennen) 70; objectual having (gegenständliches Haben) of the object 69; states of affairs 70; subject’s cognitive systems 69–70; value-ception (Wertnehmen) 69 cognitive phenomenology vs. sense-feeling phenomenology 353–354 cognitive science of sense of agency: actionrecognition experiment 342; dynamical gestalt model 340; illusory experience of agency 340; intentional binding paradigm 339; intention-free 341–342; motor control processes 342; predictive processing cueintegration theory 341; pre-reflective level 339–340, 341, 342; proprioceptive awareness of one’s movement 340–341; sub-personal motor control processes 340, 341; voluntary and involuntary movement 339 cognitivism 399 compatibilist and incompatibilist elements: desire/pro-attitude 358; emotion of authorship 359–360; in experience of freedom 1 356–357; in experience of freedom 2 357–360; implicit sense of being 358; process of selfdetermination 358; reactive attitudes 359 conative phenomenology 30 consciousness: data of 113; definiens of 134; to experience reality 113; experiencing 113; force of psyche 113; operative influence 112;

sense of content 113; structure of 111; of togethernesses and successivenesses 111 contrastive introspective revelation 406–407; advantages and disadvantages 407; dialectical use 407; direct introspective encounter 406; ostensive use 407; phenomenal contrasts 406–407; revelation theory of color 406 Crowell, S. 98, 104, 248, 254 daimon, Greek notion of 214 Dancy, J. 399 Davidson, D. 190, 239, 240, 241, 248, 365; Actions and Events 192; defining agency 239; dichotomy 80; rational-cum-intentional actions and mere happenings 88; treatment of unintended consequences 240 decision(s): is experienced as caused by something outside of me 36; is experienced as ex nihilo 36; making 368–369; or higher-order attitudes 91; or intention 74 deliberation: correctness 384; determinate from indeterminate intentions 385; habitualities 383–384; intentional structure of choice 383; remarkable antinomies 382; self-responsible truthfulness 383; sense of directedness to 382–383; state of affairs 385; volitional intention 384–385 Descartes, R. 300; formal reality 138 determinism 360n1; Brentano’s 9–10; and empiricism 8–10 D(istal) intention (D-intention) 343–345 doer and deed: concept of performativity 167; elements of action 169; political action 167–168 Dreyfus, H. 219, 237; account of skilled coping 100; authenticity and skillful 228; Being and Time 220, 228; being-in-the-world 222–224; conscious awareness 224–225; deliberative approach to action 220, 225–227; example of instructor pilots 227; expertise and authenticity 220, 227–229; flow and agential perception 225–226; human intelligence and agency 220; instinctive/reflexive behavior 226; lack of conscious awareness 219; maximal grip, notion of 221; mind and world 222–224; motor skills 225; primordial intentionality 226; pursuit of expertise 219, 224–225, 229–232; representations as phenomenological 101; rules 224–225; skill acquisition 224–225; skillful coping 219–222, 229–232; two floors model of agency 99 dualism, beyond: body, being in itself 175–176; conception of kinesthetic sensations 176; consciousness, being for itself 175–176; mental representation and physical movements 176; motor consciousness 175; teleology of bodily action 176–177

417

Index embodiment see The Essential Embodiment Theory The Essential Embodiment Theory: Dynamic Emergence Thesis 316; Emotive Causation Thesis 316; Essential Embodiment Thesis 315; Essentially Embodied Agency Thesis 316; Intentional Causation Thesis 316; Mind-Body Animalism Thesis 316 Essentially Embodied Trying thesis 317 Evans, G. 37 existence: candidate moral experiences 401; cognitive, or intellectual, mental states 400; cognitive-phenomenal eliminativism 402; cognitive-phenomenal primitivism 402; cognitive-phenomenal reductivism 402; conative/motivational states 400–401; moral goodness 401; moral mental states 400; phenomenal signature of morality 401–402; recognition-respect 403; unconscious perception, cases of 401 experiences: essentially actual experiences 122; meaning-conferring experiences 122–123 fatalism: modified fatalism 11; pure/Asiatic fatalism 11 features of action: relevances 128–129; temporality 129–130; typifications 125–128 “fiat!” character of volition in action 307–308 Fichte, J. G. 148, 149, 238 Frankfurt, H. 68, 90, 153, 314, 317, 320; The Importance of What We Care About 314; notion of personal freedom 91; second-order volitions 24n4, 68 free acts: first level of position-taking 60; level of personhood 59; lower-order position-takings 60; positionality 59–60; second-order (or higher) position-takings 60, 61 free agency, phenomenology of: cognitive phenomenology vs. sense-feeling phenomenology 353–354; compatibilist and incompatibilist elements 356–360; experience of freedom 352–353; radical freedom 354–356; ultimate responsibility 354–356; ‘up-to-me-ness’ 354–356 freedom: of agent 154; consciousness of 50; ego/sentiments 50; experience of 50, 352–353; metaphysics of 49–51; moral agency 49–50; and moral responsibility 49; motives 50; ordo amoris 51, 51n4; phenomenology of 43–44 Friston, K. 341 Gassendi, P. 300 Geiger, M. 30, 31, 39, 91; analogical description 30; passion of phenomenology 91 Gibson, J.J. 227 Goldman, A. I. 408

Habermas, J. 209 Hartmann, N. 42, 245; freedom of action, definition 243 Haugeland, J. 248 Hegel, G. W. F. 149, 167, 245; notion of “cunning of reason” 243; notion of experience in the Phenomenology of Spirit 256 Heidegger, M. 102, 106, 149, 219, 220; account of authenticity 228–229; account of freedom qua transcendence 248; action, phenomenological approaches to 97–100, 103; agency, approaches to 102–105; agent’s self-understanding 97–98; authenticity and Das Man 105–106; average everydayness, analysis of 97, 98; awareness of self 102; Being and Time 2, 96, 107, 255–256, 391; being of innerwordly things 391–392; “Bewandtnis” 391–392; capacity for action 96–97; cognition 97; das Man and inauthenticity 228; discussion of ontology 248; essence of truth 253; existential analytic of Dasein 97; first-person descriptions of action 99–100; for-the-sake-of-which/self-understanding 102, 103, 104; Gelassenheit, activity and passivity 106–107; human ‘Dasein’ 391; human resources 106; on involuntariness as complement of action 391–393; nondeliberated action 99; non-reflective 98; normativity 102–105; phenomenology vs. logical analysis 100–102; possibility of actions unconnected to self-understanding 104–105; previous setting free 392; reflective/ deliberative action 98; reflective/deliberative case 98–99; responsibility and action 104; responsibility and normativity 98–99; roles of trying, solicitation and skills 104; setting free 393; social context, individual agency 105; talk of being [Sein]/beyng[Seyn] 107; theories of freedom 105; undifferentiatedness 106; views on agency 96 Heller, A. 235 Henry, M. 112 heterotropic act 62, 64–65n18 Hildebrand, D. von 42, 70, 75; cognitive experiences and attitudes 69; content (Inhalt) 77n8; Freedom of the Will and the Concept of a Person 68; intentional action 67; notion of desire 77n12; perception of actions 67; theory of intentionality 69–72; “to want,” concept of 68, 76; will and intentional agency 68; willing and intentional action 72–75 Hobbes, T. 170, 171 human sphere of power (Machtbereich), Reiner’s 93n7; absolute and profective/ conditional sphere of power 82; creative and deletive sphere of power 82–83; material investigation 83

418

Index Husserl, E.: act, concept of 60, 61; action of willing (Willenshandlung) 16; action-will (Handlungswille) 16, 23, 24n3; act of willing (Willensakt) 16, 305; case of current willing 21; characteristics of voluntary action 21; concept of action 19–23; creative production 21; deliberation of the will (Willenserwägung) 16; fiat, notion of 25n18, 26n26; Fifth Logical Investigation 60, 363; Logical Investigations 2, 84, 110, 304; noesis and noema 113; notes on affectivity 213; notion of action (Handlung) 2; phase of action 22–23; phenomenology of willing 22; position-takings 61; purpose/ resolution 20–21; purpose-will (Vorsatzwille) 15; reading of Pfänder 15–17; resolutionwill (Entschlusswille) 15; self-positing subject/ ego cogito, notion of 199; sense of cogito 305; transcendental ego, notion of 196; volitions, strivings, and intentions 2; willing and wishing, difference between 19; zero point/center-of-orientation of my everyday experience 304 indeterminism 8 instrumental/strategic rationality 364 intention: for action 342; complex and nonunitary phenomenal structure 343; D-intentions 343–344, 345; intention-in-action 343; longterm experience of agency 345; M-intentions 344, 345; P-intention 343, 344, 345 intentional experiences 54; active 55; passive 55; spontaneous 55–56 intentional/intentionality: cognitions and attitudes, distinction between 69–72; Hildebrand’s theory of 69–72; intentional actions 61; intentional acts 60–61; intentional bodily actions 60–61; intentional body movements 315 internalism 38 involuntariness: Aristotle 388–391; artworks 393–394; Martin Heidegger 391–393; productive activities 395; situations 395–396; social interaction 394–395; strategies in arts and understanding 393–395 Jacobi, F. 148 James, W.: Principles of Psychology 307 judicial agency: acting together 151–152; feelings and emotional reactions 151; firstperson agency 151; judge’s everyday life 152; judge’s response 153; legal-juridical system 152; Talmud records 150–151 Kant, I. 88, 238, 318, 355; account of moral action 41; Critique of Pure Reason 79; notion of good will 41; open transcendentalism 242 Korsgaard, C. 98, 153, 254

Levinas, E. 147; ethical subjectivity 147; freedom’s shame 153–156; judicial agency 150–153; subjectivity before freedom 147–150; Totality and Infinity 147 Lewis, D.K. 405 Locke, J. 365 Mackie, J.L. 398, 399 Maimon, S. 148 McDowell, J. 237, 381, 399; Mind and World 180 Merleau-Ponty, M. 79, 149, 175, 220, 231, 304; agency 182–184; ambiguous body 177–180; “this back-and-forth of existence” 231; beyond dualism 175–177; bodily intentionality 180–182; intentionality of consciousness 182; operative intentionality 183–184; organism’s phenomenal body 183; personal acts 231; phantom limb phenomenon 182–183; Phenomenology of Perception 175, 177, 182, 231; The Structure of Behavior 183, 185; will 184–187 Mill, J.S. 41 M(otor) intention (M-intention) 344, 345 moral cognitivism 398–399 moral experience: characterization 403–407; cognitive phenomenology 402; existence 400–403; notion of naturalness 411n6; phenomenal commonality 402; phenomenal peculiarity 402; significance 407–410 moral phenomenology 399 motive and motivation: act of deciding 33; agent causation 36; center of I 36–37; definition of motive 32, 34–35; experiential freedom 36; motivation as noncausal relation 37; (1-P) motivation is not phenomenal causation 36; (1-R) motivation is not real causation 35; nature of 35–37; perceived chill 33; phenomenology of 33–35; recognized fact 33; “relying-on [Stützung-auf ]” 35; remembered facts 34; striving, victory of 33; structure of, argument in 31, 32–33; thought 34; volitional experience 32; and volition, relation between 29 Mylopoulos, M. 339 natural compatibilism 357 Natural Libertarianism: agent-causal theories 326; Classical Libertarianism 326, 330, 332, 333; cognitive penetration 327–328; debunking strategy/error-theory 330; firstorder consciousness 333; free will/deep freedom 317, 318, 325; Hard Determinism 325, 329–332, 333; Hard Incompatibilism 331, 332, 333; human personhood 317, 318; intentional activities 319; Lord-Byronstyle phenomenology 333; mentally caused body-movements 329; natural facts 319; phenomenology of agency 332; practical

419

Index agency 317, 318, 325; psychological freedom 327–329; sheer illusions of deep freedom 328; Soft Determinism 325, 330, 332, 333; transcendental freedom 318, 324; truth of modern science 318 necessitarianism 8 Nietzsche, F. 106, 167, 373; aestheticism 167 non-sense-feeling (NSF) phenomenology 360n3 non-sovereignty of political actor 159; adventure and venture 166; feature of irreversibility 166; foundational metaphysical fallacy 167; human condition of plurality 166; political agency and freedom 165–166 normativity: awareness of self 102; for-thesake-of-which/self-understanding 102, 103, 104; nervous spill case 104; possibility of actions unconnected to self-understanding 104–105; responsibility and action 104; roles of trying, solicitation and skills 104; social context, individual agency 105; theories of freedom 105 Oedipus, case of: “Action and Self hood” 239, 241; conscious deliberation/propositional thought 237; first-person experience of action 236, 237, 238; form of coherence/agreement 237; methodological and metaphysical role of agent 238–239; negativity and conflict 237; problem of unintended consequences 236–237, 238; temporal gap 237–238; thirdperson perspective 238 On haunting, projections, and engagement 139–140; l’être parfait 139; ontological dependence 139–140; positional consciousness/intentionality 139; purity of pure reflection 139 Operari Sequitur Esse: account of intentional action 209; agency as self-efficacy 209–211; attitude, notion of 212–214; attitudinal theory of agency 209, 216–217; minimal self hood and mineness 217; questions of personhood and bodily existence 209; racist motivation 208; self-efficacy, attitude as form of 214–216 O’Shaughnessy, B.: the body-schema 324; causal theory of action 320; deep freedom 322; desires 322–323; dual aspect theory of intentional action 315; The Essential Embodiment Theory of the mind-body relation 321; intentional body movement 320–321; self-consciousness 323; subpersonal/unconscious 321–322; theory of bodily action and mine 321; The Will 319 ought-judgment: and motivation, connection between 37–39; phenomenological argument 37–38; possibility of amoralists 38–39 ownership, sense of 54

Pacherie, E. 343 paradox of plurality 162–163; distinctness 162; equality 162–163; political action, freedom, and power 162 passivity: bodily 83; Gelassenheit 106–107; intentional experiences 55; motivated 87–88; Reinach 53; sensations 59 personhood’s emergence: facing objectivity, first degree 371–372; managing experience, second degree 372; notions of emergence and subject 371; perceptual and emotional illusions 371–372; personal identity, third degree 372–373; personality and character traits 372; second-order position 372; selfconstitutive acts 373 Pfänder, A.: act of choosing 24n2; act of willing 16; actualself-consciousness 18–19; agentive/ conative phenomenology 30; eccentric/ off-center 31–32; ego-center/core and ego-body 31; Husserl’s reading of 15–17; I as phenomenal originator [phänomenaler Urheber] 39; I-center/I-core and I-body 31–32; material and mental realities 30–31; metaphorical 32; motivation, phenomenology of 33–35; motive and motivation 31, 32–33, 35–37; Motives and Motivation 15, 18; oughtjudgment and motivation, connection between 37–39; phenomenal cause, notion of 31; phenomenological description, method of 30–32; position in classical phenomenology of agency 39; retrospective method 18; selfawareness 17; subject of experience 31; subject referent 18; treatment of act of willing 59; volition and motivation, relation between 29; volitive experience, analysis 17–19 phenomenal: agentive-instigation experience 290; body 179, 183; cause 31; character 305–306, 403–404; commonality 402; consciousness 363; content-based satisfaction conditions 290; eliminativism, cognitivephenomenal 402; experiential-acquaintancewise state-causal 290; facts about limits of reliable introspectability 291; I as phenomenal originator [phänomenaler Urheber] 39; manifest facts 291; (1-P) motivation 36; noncontingently implementational satisfaction conditions 290; non-manifest facts 291–295; not as-of one’s action 290; peculiarity 402; phenomenal-cum-intentional dimension of activity 81–82; phenomenally manifest facts 291; phenomenally nonmanifest facts 291–295; powers 179; primitivism, cognitive-phenomenal 402; reductivism, cognitive-phenomenal 402; signature of morality 401–402 phenomenology: accounts of human agency 2; action and phenomenology of

420

Index agency 336–339; agency and intention 342–345; cognitive science of sense of agency 339–342; experiential sense 2; first-person perspective 1; historical designation 1; methodological sense 1; as philosophical method 1–2; sense of agency and social context 345–347 Plato 250 position: eliminativism/irrealism 402; nonreductivism/primitivism 402; positiontaking/stance 77n6; reductivist realism 402 positionality of spontaneous acts: conviction and assertion 58; positionality of experience 57; role of self 57; states and acts, distinguish between 57–58; temporality of experience 57 practical intentionality 365–366 present/proximate intention (P-intention) 343, 344, 345 psyche 110–115; acts of free will 116, 118; being set in motion. 116, 117; and body 119; causality of 112, 113; cause-effect, stimulusresponse 112; comparing and alter ego 113; consciousness 111–113; emotions 114–115; empathy 111, 113–114; and ethical action 115–119; lived experiences 116–117; motivation 116–117; pure ego of consciousness 112; self-mastery/selfpossession 118–119; sensory impressions 111; spirit 116, 119; unity of body, psyche, and spirit 110; values 117–118, 119 radical freedom (RF) 354–356 ratio cognoscendi of freedom 355 rational agency: actuality, subjectivity, and personhood 373–374; degree of personhood’s emergence (see personhood’s emergence); exploring the world, learning from experience, doubting 369–370; hierarchy of acts 370–371; instrumental/strategic rationality 364; intentionality 362–364; kinds of acts 367–368; making decisions 368–369; mental states 374n1; objectivity and corrigibility 364–365; phenomenal consciousness 363; positionality/positiontaking 364; practical intentionality 365–366; state of art 362; theory of acts 366–367; value-based rationality 364 rational-cum-intentional action: ego-centrality (Ich-Zentralität) 90; form of self-determination 90; inner will (inneres Wollen) 88, 89; phenomenon of “letting it happen”/resisting 89 rationality: conceptual and intellectualized rationality 221; instrumental/strategic rationality 364; value-based rationality 364 reasons: (apparently) acts on basis of practical reason 266; agentive phenomenology 266–267; appreciation of practical reason 265–266;

chromatic illumination 265; introspection 266; not as-of one’s (apparent) action 266; not a state-causal process 266; optionality 265; practical reasons (purposes) 265; psychological state causation (PSC) 267; purposiveness 265; the satisfaction conditions 267; self as source 265; state-causal determinism 267; statecausal efficacy of psychological states 267; well-documented scientific evidence 266–267 Reinach, A.: activity 53; agency, sense of 54; The Apriori Foundations of the Civil Law 53, 57; authorship, sense of 53, 54; intentional experience 54; kinds of experience (see individual experiences); ownership, sense of 54; passivity 53; self, role of 54; social acts 53; spontaneity 53; spontaneous acts (spontane Akte) 53; On the Theory of the Negative Judgment 54, 56, 57 Reiner, H. 79, 80, 89; consciousness 81; extended (inner) will 88–90; and Frankfurt 90–91; freedom, activity, and action 81; Freedom, Willing, and Activity 79; higher-order attitudes/decisions 91; inner will, notion of 80–81, 89, 91–92; notion of centrality 94n23; phenomenal-cum-intentional dimension of activity 81–82; Phenomenological Investigations Towards the Problem of Free Will 79; as phenomenologist 81–83; pre-free actions 93n12, 94n20; problem of action 80; rationalcum-intentional action 80; schema of human activity 92; self-determination, idea of 84; solution to free will problem 80; subjective human sphere of power (Machtbereich) 82–83; volitional agency 88–90 relevances: in-order-to motivational 127; interconnection of projects and sub-projects 127; intrinsic and imposed 128–129; social roles 128; topical and motivational 128 resolve 385–386 Ricoeur, P. 189, 193, 194, 198, 242, 252, 307, 409; account of agent’s action 190, 191, 204; action and language 201; active and embodied consciousness 197; act of reading 201–202; Analytic philosophy of action 194; bios and logos 203; continental tradition 193; The Course of Recognition 194–195, 197, 200, 204; criticism of Anscombe and Davidson 190–193; ego of ego cogito 197; Fallible Man 199; Freedom and Nature 194, 197, 198; Freud and Philosophy 195; hermeneutic and reflexive phenomenology of action/agency 198; hermeneutics of action 200–206; intentional exteriorization/ materialization 202; ipseity/ipse-identity 205; The Just 189; Lectures on Ideology and Utopia 200; linguistic phenomenology 195; manner of doing philosophy 189–190; meaningful action 190, 201; Mimesis/configuring activity 201;

421

Index mimesis/prefiguration 201; narrative identity/ self hood 205; noetic-noematic consciousness 199; nonsymbolic action 200; objectifiable action 197; Oneself as Another 189, 191, 194, 197, 199, 200, 204, 205; phenomenology of action 193–199; phenomenology of being able 193–194, 205; philosophy of imagination 206n3; points of convergence 195; reconfigurative emplotment 189; Reflections on the Just 189; role of active reader 197; role of experiencing subject 197; The Rule of Metaphor 195; self-interpretation, or understanding 203; speech act theory 196; symbolic action 201; The Symbolism of Evil 195; symbols of human being 200; teleological explanation 192–193; temporal personal identity and ethical ascription 205; theme of pre-reflexive affirmation 198–199; Time and Narrative 201; “who” of action/ agency 204 Rorty, R. 314; Consequences of Pragmatism 314; Philosophy and the Mirror of Nature 314 Rousseau, J.-J. 170 Russell, B. 230 Ryle, G. 369 Sartre, J.-P. 79, 137; being of consciousness, being-for-itself 135; constitution of motifs 143; contemplative consciousness 137; dualism and his notion of consciousness as néant 135–137; magnum opus 142–143; néant as nothingness 135; not being consciousness 136; notion of action 134; notion of beingfor-consciousness 136; notion of ontological underivability 136; notions of action theory 143; ontological irreducibility 136; power to negate 136; power to nihilate/negate 136; primary interest 144–145; projet originel 214; regions of being: non-conscious being-initself 135; “structures contained within the very idea of acting” 142–143; systematic place of acting 135; transcendence into nothingness 135; unified theory of subjectivity 134 Scheler, M. 41, 42, 49, 372; action theory 44–49; characterization of action as unity 49; doctrine of “Ordo amoris” 43; “duty for duty’s sake,” notion of 41–42; Enlightenment 50; epiphenomenalism 50; ethical personalism 41–43; Formalism 50; Formalism in Ethics 43; Greatest Happiness Principle 42; human freedom 43; lawfulness of life 50; metaphysics of freedom 49–51; moral judgment 51; phenomenology of freedom 43–44; phenomenon of personhood 43; sense of being 43–44; sense of power 44; skepticism 50; value-perception (cognition of value) 41; values themselves 42

Schelling, F.W.J.: account of human freedom 245; interpretation of Oedipus Rex 244 Schmitz, H. 209; acts of bravery and cowardice 210; affective involvement 211; agency, attitudinal theory of 209; attitude 211; causation 210; determinism 210; doctrines of self-consciousness 216; independent initiative 209–211; initiative and choice 211; notion of subjectivity 217n6; subjective factuality 211 Schopenhauer, A. 11, 214 Schütz, A. 121, 123; action (Handeln) 123; behaviorism and dualism 121–122; Cartesian dualism 121; essential features of action (see features of action); experiences 122; inner stream of consciousness 122; “in-order-to” motive of action 123–124, 130, 131; levels of human activity, action, and motives 121–124; memory/anticipating future 123; The Phenomenology of the Social World 121–122; pragma/activities 123; projects of action and rational action 131–132; reciprocity of motives 124; secondary remembrance/ reproduction 122; soul, behaviorism 121; spontaneous Ego-act 123; theory of action 121, 130 Searle, J. 101, 191. 204, 343, 365, 380, 408; bankruptcy of phenomenology 101; conditions of satisfaction 103; mind, analytic philosophy of 101; phenomenological illusion 101 self-efficacy: agency as 209–211; attitude as form of 214–216 self, senses of activity of 56–57 Sellars, W. 292 sense-feeling phenomenology 353–354 sentimentalism and rationalism 399 shame, freedom’s: ethical capacity 155; freedom of agent 154; freedom on ethics 153; free will 153–154; interpersonal responsibility 156; love and charity 155; one’s rolebound responsibilities 155–156; selfdetachment/self-distancing 154; subject’s distinterestedness(désintéressement) 154; Ursubjectivity of subject/agent 153–154 significance: connection principle 407–408; conscious and unconscious processes 407–408, 409; descriptive and explanatory psychology 409–410; genetic and descriptive psychology 409; kinds of curiosity 407–408, 409; moral experience and moral life 407–410; prototypical mental states 408 skillful coping: concepts, representations, and reflectiveness, notions of 222; conceptual and intellectualized rationality 221; and deliberative aspiration, planning, and learning 230; everyday being-in-the-world 221; fluid coping 229–230; human existence

422

Index 222; human intelligence and agency, conceptions of 221; and ongoing pursuit of expertise 219, 229–232; phenomenon of expertise 221; reflective self-awareness 222; training and skill 230 Smith, M. 399 social acts: heterotropic 62, 64–65n18; inner doing of the self 62–63; intentional bodily actions 62; layering of intentional experiences 63; nature of 62; a priori foundations of law 62; second-order position-takings 63; spontaneous 62; spontaneous acts 61–63; zero degree of positionality 63, 65n19 social context, sense of agency and: case of joint actions 345–346; conceptions of autonomy 347; cross-cultural studies 346–347; incentive-sensitization model 346; intention formation and motivation 345; long-term sense of agency 346; role of sub-personal processes 345; WEIRD outliers 346 speech acts 64n15 spontaneous acts: positionality of 57–58; position-takings on position-takings 57–58; as social acts 61–63 spontaneous intentional experiences: authorship of self 56; doing of the self 55; inner doing of subject 56, 59; punctuality of the experience 55; sense of autonomy and self-determination 56; social acts 56 Stein, E. 59, 110; action, definition 118; act of sense-making (Sinngebung) 110; Philosophy of Psychology and the Humanities 110; psyche and ethical action (see psyche); treatment of free acts 59 Steward, H. 326 Strawson, P. F. 359 subjective factuality: affective involvement 212; choosing/self-control 211; mineness 212; notion of 209, 211–212; one’s own initiative 211; phenomenology of pre-reflective selfconsciousness 211–212 subject/subjectivity: acts of free will 148; anti-humanist/anti-subjectivist 147–148; being-in-the-world 149; ego 148–149; ethical subjectivity 148; human existence 150; spontaneity and form of causality 149; Urself hood/Ur-subjectivity 148–149 systematic expectations (SE) 135 Taylor, C. 153, 192 temporality 129–130 Tengelyi, L. 235, 238, 242; “Action and Self hood” 239, 241, 245, 252; from action to agent 239–242; adherence to reality 250; constitution of self hood 249–251; critique of Davidson’s semantic theory 236; on Davidson 239–242; experiential relation 239–242;

first-person involvement 236, 240, 241, 256; forms of determination 243–244, 247; freedom and necessity 242–245, 246; freedom qua transcendence 247; free, ways to be 245–249; human freedom and causal determinacy 235–236; intentional actions and events 240–241; miniature narrative, kind of 240; moral phenomenology 246; narrative identity 249–252, 254; narrative intelligibility 249, 251; natural event and causal effects 241; notion of narratibility 250; Oedipus, case of 236–239; recognition of causality 247–248; recollection and narration 253; reflective self-awareness 250; risking unintended consequences 240, 241, 255; strong indeterminism 245; treatment of human freedom 253; truth, case for 252–257; unintended consequences, meaning of 239, 242–245, 257; validity of theories of narrative identity 250; World and Infinity 235, 239, 242, 243 theoretical roles 403–406; experience of recognition-respect 404–405; nonreductivist/ primitivist framework 404; phenomenal character 403–404; role of respectr experience in theory 405 Thompson, M. 377 typifications 125–128; experiencing object 125; higher level project 126; in-order-to motive of one’s action 125–126, 127; projection of performances/overt actions 125; social relationships 126–127 Uexküll, J. von 183 ultimate responsibility (UR) 354–356 unintended consequences: forms of determination 243–244; freedom and causality 244–245; meaning of 239, 242–245; parallelism of freedom and necessity 243; primordial facts 242; problem of 236–237, 238; risking 240, 241; treatment of 240 ‘up-to-me-ness’ 354–356 value-based rationality 364 variety thesis, four-field schema, Reiner’s 80, 83–88; awareness of power 84–85; bodily motor activity 83; bodily passivity 83; hybrid forms of willed activity 87; intentional actions 84; mental action 86–87; motor activities 87; non-intentional activities 80; non-volitional bodily movements 87; partial acts (Teilakte)/moments 84; passive experiences 87; performance/execution 86; pre-free sphere 87; purely affective and motivated passivity 87–88; states of affairs 85–86; trichotomy 80; volitional stances (Willensstellungnahmen) 85

423

Index volitional experience: act of willing 32; impulse of willing 32; and motivation, relation between 29; voluntary action (Willenshandlung) 32 von Wright, G.H. 204: notion of closed system 203 Waldenfels, B. 245 Wegner, D. 340 will: determinism and empiricism 8–9; emotion and 7; ethics of Bildung 11–12; and intentional agency 68; perceptual and physical powers of body 185; pre-personal bodily existence 185; pre-personal life of consciousness 184–185; self-addressed speech-acts 187; thetic and operative intentionality 185; ultimate fact 185–186; voluntary and rational life 187; voluntary thought and action 186

willing: activities (Tätigkeiten) 72; decision/ intention 74; experiences 72, 73; Husserl’s concept of action 19–23; I-center 33; inert willing, experience of 74; mental action 74; notion of awareness 75–76; projects, awareness of object 33; realization 74–75; state of affairs 72–73; theory of action as theory of willing: Husserl’s reading of Pfänder 15–17; volitive experience, in Pfänder’s analysis 17–19 will-power: essentially embodied trying and natural libertarianism 315–319; O’Shaughnessy and Me 319–324; structuring cause 316; veridical phenomenology of essentially embodied free agency 324–332 Wittgenstein 202, 314; Investigations 314 Wrathall, M. 104, 220, 222, 230, 256

424