Ontology Revisited: Metaphysics in Social and Political Philosophy 2012004328, 9780415574112, 9780203855157

Citation preview

Ontology Revisited Metaphysics in Social and Political Philosophy Ruth Groff

ONTOLOGICAL EXPLORATIONS

Ontology Revisited

Groff’s argument cuts against a familiar anti-metaphysical grain. Social and political philosophy, she maintains, is not as metaphysically neutral as it may seem. Even the most deontological of theories connects up with a set of philosophical commitments regarding what kinds of things exist, as a fundamental ontological matter, and what they are like. These are topics of interest not just to social and political philosophers, but to social scientists and to philosophers of social science as well. “Ruth Groff has broken new ground in demonstrating the connection between social and political thought and the ontology of causal powers. Her account of the structure of Humean thinking about agency is excellent. Especially significant is the role that she assigns to Kantianism in the analysis that she develops. She moves effortlessly between contemporary metaphysics, political theory, critical social theory, and the history of modern philosophy, offering trenchant insights along the way into the work of thinkers ranging from Hume himself to Mill, Adorno, and Martha Nussbaum, and into debates over agent causation and emergence. There is even a discussion, in the final chapter, of Spinoza. This is bigpicture philosophy at its best: rigorous and exacting at the level of detail; original, compelling and systematic in the whole.” – Stephen Mumford, Professor of Metaphysics and Dean of the Faculty of Arts, University of Nottingham Ruth Groff is Assistant Professor of political philosophy at Saint Louis University. She is a member of the Centre for Contemporary Aristotelian Studies in Ethics and Politics, and a former editor of the Journal of Critical Realism. She is the co-editor of Powers and Capacities in Philosophy: The New Aristotelianism.

Ontological Explorations

Other titles in this series: From One ‘Empire’ to the Next Radha D’Souza Science for Humanism The recovery of human agency Charles R. Varela Philosophical Problems of Sustainability Taking sustainability forward with a critical realist approach Jenneth Parker Dialectic and Difference Dialectical critical realism and the grounds of justice Alan Norrie Interdisciplinarity and Climate Change Transforming knowledge and practice for our global future Edited by Roy Bhaskar, Cheryl Frank, Karl Georg Høyer, Petter Naess and Jenneth Parker Conversations about Reflexivity Edited by Margaret S. Archer Relational Sociology A new paradigm for the social sciences Pierpaolo Donati Sociological Realism Edited by Andrea M. Maccarini, Emmanuele Morandi and Riccardo Prandini

The Economics of Science: A Critical Realist Overview Volume 1: Illustrations and philosophical preliminaries David Tyfield The Economics of Science: A Critical Realist Overview Volume 2: Towards a synthesis of political economy and science and technology studies David Tyfield Dynamic Embodiment for Social Theory “I move therefore I am” Brenda Farnell Ontology Revisited Metaphysics in social and political philosophy Ruth Groff

Ontology Revisited Metaphysics in Social and Political Philosophy

Ruth Groff

First edition published 2013 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Simultaneously published in the USA and Canada by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2013 Ruth Groff The right of Ruth Groff to be identified as author of this work has been asserted by her 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 Groff, Ruth, 1963Ontology revisited : metaphysics in social and political philosophy / Ruth Groff. p. cm. — (Ontological explorations) Includes bibliographical references and index. 1. Metaphysics. 2. Ontology. 3. Social sciences—Philosophy. 4. Political science—Philosophy. I. Title. BD111.G73 2013 110–dc23 2012004328 ISBN: 978–0–415–57411–2 (hbk) ISBN: 978–0–203–85515–7 (ebk) Typeset in Times New Roman by RefineCatch Limited, Bungay, Suffolk

For my mom and dad. And for Richard Schuldenfrei, who ignited my love of philosophy.

Contents

Acknowledgments Preface

x xii

1

The Myth of Metaphysical Neutrality

2

Hume: Custom as Metaphysical Necessity

11

3

J. S. Mill: Humeanism and the Perfection of Distinctively Human Capacities

33

Kant and the Frankfurt School: Freedom as Escape from the Transcendental Subject

50

Agents, Powers and Events: Humeanism and the Free Will Debate

73

Metaphysics and the Capabilities Approach: Martha Nussbaum, Political Liberalism and the Ideal of Metaphysical Neutrality

92

4

5

6

7

1

Powers, Ontology and the Appeal to Spinoza

106

Notes Bibliography Index

118 130 134

Acknowledgments

It may well take a village to raise a child, as the saying goes. But it also takes one to write a book. In fact, I am tempted to say that it takes St. Louis, Missouri in particular. I had been thinking about the general ideas in this book as early as 2005, but it was not until I landed at Saint Louis University in the fall of 2008 that I began to think about them in earnest. That I did so is directly related to the fact that that same semester several members of the philosophy department launched a reading group on the topic of dispositional realism and agency. Also decisive is that my own department, in conjunction with the College of Arts & Sciences, saw to it that I got time off in order to write. I am indebted to many friends and colleagues for their time and their intellectual generosity. Here in St. Louis, Bill Charon, Kent Stalley and Scott Ragland had long conversations with me about Kant, Mill and Spinoza, respectively. Jonathan Jacobs responded to endless queries regarding his views. Larry May read and commented on an early draft of Chapter 2, and Elizabeth Forman did the same for Chapter 6. I’d also like to thank the graduate students – especially Corey Katz, Jim McCollum and Josh Anderson – who joined me in an Adorno reading group; Corey and Jim then took part in a seminar that I teach on the Frankfurt School. John Heil replied swiftly and helpfully to questions about Davidson and Kim. Outside of St. Louis, David McNally, Asher Horowitz, and Christian Thorne all helped me to clarify my thinking about Adorno and Kant, and Patchen Markell offered perfectly timed feedback on Chapter 4. Anjan Chakravartty answered every question put to him. Outside of North America, I am extremely grateful to Alexander Bird, Stephen Mumford, Jonathan Lowe and Brian Ellis. Alexander went to heroic lengths in helping me to revise sections of Chapter 5; Stephen Mumford saved the philosophical day more than once, and, along with Jonathan Lowe, read and commented on the chapter in its entirety. Jonathan also generously shared his thoughts with me about emergence in the case of purely physical phenomena. Brian Ellis read the whole book, the later chapters just as soon as they were written, which I confess made me feel a bit like a nineteenth-century novelist. His enthusiasm was enormously heartening, and Chapter 7 was greatly improved by his feedback. Of course, every weakness of the book is attributable only to me. There are two people without whom this book simply would not have come to be. One is Irem Kurtsal Steen; the other is John Greco. Irem and I have talked

Acknowledgments xi philosophy constantly over the past few years, and she is both a brilliant metaphysician and a brilliant interlocutor; her intellectual company has been invaluable. John Greco has been incalculably supportive of me intellectually and professionally, ranging from engaging in early discussions about the conception of the book, to welcoming me into seminars on Hume & Reid and on free will, to providing detailed comments on Chapter 5, to co-editing a volume with me. Think of any example of concrete support that might be offered to a junior colleague, and Greco has done it. I would not say this just to be nice. Finally, I want to say thank you to my wonderful family, and more dear friends than anyone deserves.

Preface

This book grew out of my interest in causal powers in just the way that my argument suggests that it would have: once I was aware of my conviction that causation is not what Hume said it is, and instead involves the expression or display of the real dispositional properties of things, it was natural to think that people, who are manifestly efficacious after all, have such properties too. Moreover, it followed that if there are any sociological entities that cause things, there must be some sense, at least, in which those entities have powers also. Once I was clear about all of this, I couldn’t help but be curious about how the issues were handled in the history of social and political philosophy. Were non-Humean ontological commitments generally presupposed in the modern and contemporary period? Evidently not. At first I thought that the main point would pertain to the conception of freedom: from a dispositional realist perspective, freedom would have to be understood to be the expression of our own distinctively human causal powers, especially that of self-conscious self-determination. And if this were so, if free human agency is a species of what analytic philosophers call substance causation, I thought – and still do think – then the dichotomy between freedom and causal determination inherited from Kant would turn out to make no sense. Nor would Berlin’s preoccupation with negative liberty. Efficacy is efficacy, be it effortless and unimpeded (negative liberty) or disciplined in accordance with some relevant standard of excellence (positive liberty). One might even want to say that the greater the degree of causal determination (by the agent) in relation to any given act, the greater the degree of freedom. One should say that, in fact, at least to begin. Admittedly, Kant too has it that freedom is a kind of necessity: we are subject to moral law, not just legislators of it. But free agency is precisely not, for Kant, a causal phenomenon. That realism about causal powers dissolves the dichotomy between freedom and causation seemed, as I say, to be the crucial point. It soon became apparent to me, however, that there was a different way to tell the story, one in which the metaphysical confusion associated with the standard ways of thinking about freedom was only a chapter. The bigger story – or perhaps “deeper,” or “prior,” is a better way to put it – was a diagnostic one. It was not about powers, but rather about Hume’s rejection of powers, and about the impact on modern and contemporary philosophy of this and other aspects of his ontology. That’s the story that I resolved to tell, and which I have told.

Preface xiii But there was a complication. It is in the nature of the case that one can’t tell such a story without presuming that there is an intimate relationship between social-political thought and metaphysics. Such a presumption has been out of fashion of late, in no small part because it runs counter to the Humean ontology in question. Thus to show how a set of ontological assumptions derived from Hume has issued in confusion at the level of social and political thought, I also needed to challenge what I’ve called the myth of metaphysical neutrality. Happily, the logic holds in both directions: just as I couldn’t tell the story without assuming a connection between metaphysics and social-political thought, the story itself confirms it. The elaboration of a positive, powers-based social-political theory I have set to the side, although I would note that Aristotle, Marx and even Mill have already made significant headway toward such an end, as have a number of contemporary authors working in the spirit of these thinkers. What contemporary proponents of such theories have lacked are the resources with which to defend their implicit metaphysical assumptions in the face of the Humean orthodoxy into which Kant’s dichotomy figures. For now, then, I offer a meta-theoretical critique, coupled with a debunking of the myth of metaphysical neutrality.

1

The Myth of Metaphysical Neutrality

There is by now a well-entrenched habit of thinking, according to which social and political philosophy have nothing to do with metaphysics. The underlying thought is plausible enough: that one can ask certain kinds of questions without having asked, or implicitly answered, other kinds of questions. Even for Aristotle, phronesis was not the same as sophia, politics not a matter of philosophy. Indeed, it would seem to be a settled point, after the rise of modern science, that there are no normative injunctions regarding human arrangements built into the natural world. Or, at least, if there is anything to be learned from Rawls, belief to the contrary would seem to fall outside of the boundaries of acceptable public discourse in a liberal society – a foul play, equally likely to be called out by a postmodern critic of grand narratives as by a post-metaphysical political liberal. For these reasons alone, and no doubt there are others, it would seem to go without saying that whatever it is that we’re doing when we do social and political philosophy, it doesn’t hang on issues of concern to metaphysicians. But – or so I shall argue – social and political philosophy is not metaphysically neutral at all. Received beliefs of the day notwithstanding, even the most deontological of theories connects up, in the end, with an attendant set of basic commitments regarding what kinds of things exist, what they are like and how they are or are not put together. Indeed, just as the once celebrated value-neutrality of liberalism proved to be an unselfconscious affirmation of a specific set of normative principles, the very idea of metaphysical neutrality turns out to be an unselfconscious affirmation of what is called Humeanism (or, less technically, mechanism): the presumptive, anti-essentialist ontology of the contemporary period. It is Humeanism that invites one to imagine that that which exists has no form of its own, and “without a form of its own” is just what the world would have to be like, in order for accounts of social and political phenomena to involve no metaphysical commitments. In this respect, the notion that social and political theory is metaphysically neutral can be seen to be something akin to a false but real appearance, to use the Marxist terminology. It is the meta-theory generated by the dominant philosophical position. That, in the main, modern and contemporary social and political thought rests unreflectively upon a Humean metaphysics is hardly an insight that is original to me, even if I may be the first to put the particular fine point on it that I aim to do.

2

Ontology Revisited

Charles Taylor comes immediately to mind as a contemporary thinker who has long advanced such a view.1 Key members of the Frankfurt School, we might want to say, are others, as is Alasdair MacIntyre – and, of late, a grouping of critical theorists who have taken an interest in the concept of ontology, some via an appeal to Spinoza. Critical realists, too, have made the argument, although primarily in the context of sociological theory rather than that of political philosophy. Still, there is original work to be done in laying bare the infrastructure in question. Such is the present undertaking, the core thesis of which is that Humeanism has shaped and arguably distorted modern and contemporary thinking about agency broadly construed. This thesis is accompanied by two additional claims, one ancillary, the other implied: (a) that recent neo-Aristotelian work in analytic metaphysics allows the effects of Humeanism to come clearly into focus; and (b) that social and political thought is not, in fact, metaphysically neutral. My intention in this introductory chapter is to lay out the competing metaphysical positions at the heart of the analysis. These are Humeanism and its emerging rival, variously labeled dispositional realism, scientific essentialism and/or, less formally, a powers-based ontology. To be clear: by metaphysics I mean something other than moral realism and/or considerations regarding free will (though I shall indeed address the influence of Humeanism on the free will problematic in Chapter 5). Rather, I use the term in its most traditional sense, to refer to a thinker’s basiclevel ontological commitments. The purpose of this chapter, as I say, is to articulate those that constitute Humeanism and dispositional realism respectively. The progression of the argument will then be as follows: in Chapters 2–4, I consider Hume, Mill and – via Kant – Marcuse, Horkheimer, and Adorno, in order to illustrate both the general point that social and political thought is shot through with metaphysics and the specific point that Humeanism in particular has configured social and political thought in the modern and contemporary period. In Chapter 5, I extend the discussion of the dichotomy between spontaneity and determinism in Kant to an analysis of the structure of the free will debate in contemporary analytic philosophy. In Chapter 6, I turn to the so-called capabilities approach associated with Martha Nussbaum (and Amartya Sen). I argue that just as Mill could not square a developmental, capacities-based liberalism with an overt commitment to Humeanism, Nussbaum’s own version of the view requires a powers-based metaphysics. The upshot of these two chapters is that dispositional realism can underwrite our thinking in significant ways, ranging from a re-situation and elaboration of the concept of agent causation to a properly grounded normative appeal to human capabilities. Finally, I conclude in Chapter 7 by considering a recent turn to Spinoza, by continental political theorists, for the purpose of recovering talk of ontology in general and powers in particular. Let me begin with Humeanism, or mechanism, terms that I join Brian Ellis and others in using to refer to a set of interrelated claims endorsed in part or in whole by a range of modern and contemporary thinkers.2 As Taylor has noted with respect to the family of views that he calls naturalism, some of the thinkers who fall under the heading of Humeanism disagree with one another profoundly, over crucial issues – e.g., Hume and Kant. There is, however, a shared core account of

The Myth of Metaphysical Neutrality 3 the nature of things that can be used to define, or at least to pick out instances of, the position in question. Thus, by Humeanism I mean: (i) a rejection of the idea that objects (however they be defined) have causal powers; (ii) an attendant anti-naturalism about causality; and (iii) anti-essentialism. Ellis himself offers a longer, more fine-grained list, characterizing what he calls mechanism as the composite view that: (a) inanimate matter is essentially passive, never intrinsically active; (b) things behave as they are required to by the laws of nature; (c) the dispositional properties of things (including their causal powers) are not real properties, and are never intrinsic to the things that have them; (d) the essential properties of things never include any dispositional ones; (e) causal relations are always between logically independent events; (f) the laws of nature are universal regularities imposed on things whose identities are independent of the laws; and (g) the laws of nature are contingent, not necessary.3 Principle (g) draws a line between Humeans and Kantians that I do not want to draw (though let me state unequivocally that there are points of irreconcilable difference between empiricists and Kantians, and that most of the time these differences are what is most salient); Kantians show up as Humeans, on my pared down model.4 For the purposes of explication, it will make sense to consider the three basic tenets out of order, beginning with anti-essentialism rather than with the rejection of powers. Anti-essentialism is the view that the things that exist do not have essences. Loosely speaking, an essence may be thought of as a property or set of properties, the bearing of which makes something be the kind of thing that it is, rather than something else.5 To say that there are no such things as essences, then, is to say that what things are is not based upon something inherent in them: identity as a this or a that, the anti-essentialist will say, derives instead from an external source, e.g., God, the laws of nature, language. Descartes, for instance, held that the only property that matter has intrinsically is that of extension, presence in space. Belief in the existence of generic, inherently indeterminate “stuff” may not seem unusual to us post-Cartesian moderns, but it’s worth noting that Aristotle found it to be unintelligible, concluding by contrast that matter is always already something, even if what it is not always fully actualized. Social and political theorists often equate anti-essentialism with postmodernity. But it is important to get the timing right. The break with Aristotelian physics was simultaneously a break with Scholasticism. The new, essence-less view of nature associated with the science of mechanics was quickly interpreted and reinforced philosophically – early on by Descartes, later through the rise of British empiricism, which ruled out form or essence on epistemic grounds. Locke, for example, took care to distinguish between the concepts of nominal essence and real essence.6 He defined real essences as the internal structures of things. Nominal essences, by contrast, are the observable features of things. Things are sorted into kinds on the

4

Ontology Revisited

basis of their nominal essences, Locke said, because real essences cannot be observed (or at least couldn’t be at the time of his writing),7 and therefore cannot be known. Eventually the category of real essence dropped out of the (non-Hegelian) modern picture altogether. From Hume onward, the predominant view has been that there are only categories that we impose, onto a substrate of one sort or another. And given that essentialists had thought that things act as they do in virtue of what they are, the concept of causal necessity, too, had to be rejected, asserted dogmatically or somehow re-grounded, once anti-essentialism took hold as the default metaphysics. Anti-essentialism made it into social and political philosophy almost immediately. Leviathan, published in 1651, opens with the brazenly modern assertion that objects only seem to bear intrinsically the qualitative features of them that we perceive. In reality, Hobbes says, what we take to be qualitative properties are merely subjective “fancies” – misinterpretations of the sensations that are produced by the fact of matter impinging upon our bodies. Similarly, universals are nothing other than names that we give to like sensory input. Granted, the argument is that people do, or at least should, naturally seek peace through the institution of a Sovereign. But that this is so isn’t, for Hobbes, a function of a posited human form or essence. Rather, it is simply an analytically derived implication of the reality of instrumental reason. In saying that anti-essentialism is paradigmatically modern, I do not want to be misunderstood. While I take the claim to be uncontroversial, I do not mean to suggest that anti-essentialism itself was uncontroversial. Anti-essentialism did become the default position, but it didn’t become so without opposition. Nor, as I’ve stressed, do I mean to suggest that it was only empiricists who repudiated the classical view of entities as being determined from within, via their form. Kant too assumes an anti-essentialist stance towards the physical world (in that form is thought to be given via the unity of apperception, rather than being an inherent feature of that which provides content to phenomenal experience) — as did Descartes. The point that I want to make is simply that there is nothing new about anti-essentialism. It is a hallmark of modernity, not of post-modernity.8 Modernity also brought with it the rejection of powers. Analytic metaphysicians often talk about powers via the concept of a dispositional property.9 A dispositional property can be thought of as a property of being able to do or act – though it may be that there is no such thing as a generic disposition, that there are only properties of being able to do this or that, to engage in this or that kind of activity. Abilities do not have to be exercised, however. Nor, exercised, do they necessarily issue in those phenomena that (nevertheless) figure in their definition and individuation. Thus a power can be thought of as a not-necessarilymanifest capacity to potentially effect a change of some kind. Humeans deny the existence of powers, regarding belief in them as a kind of pre-scientific animism, a vestige of imagining the world to be suffused with magical, or at the very least teleologically directed, forces. From a Humean perspective, the material world is not potent. It is, on the contrary, inert; it is “the dead world of mechanism,” as Ellis has called it.10

The Myth of Metaphysical Neutrality 5 The disavowal of powers is related to anti-essentialism, if only because the Aristotelian essences that were expelled from the modern conception were, fundamentally, identity-constituting capacities to do certain kinds of things, viz., to engage in the activity, or “work,” that is characteristic of a given kind of thing. From an Aristotelian perspective, it is an essential property of water, for example, that – all things being equal – it can, in its liquid state, dissolve substances such as sugar. In doing away with form and replacing it with mere extension, then, moderns can be seen to have dismissed powers simply as a function of rejecting essentialism. But the embrace of anti-essentialism is only one part of the story.11 There is also the claim – analytically distinguishable from a commitment to anti-essentialism – that powers, just like essences, do not exist. There is simply no such thing as a dispositional property, say many moderns, be such purported properties essential or not. Or, as Hume would have it, the idea of a power or disposition has no meaning. The next move is to make a distinction between those (non-dispositional) properties that are taken to exist objectively, or externally, and those that are conceived as existing only subjectively, in the form of phenomenal experience, e.g., color. Locke refers to the former as “primary” qualities, the latter as “secondary.” After this comes the final, reductive-materialist move – that of concluding that, even amongst only non-dispositional properties, the subjectively constituted ones are ultimately not real. This set of claims, and the relationships between them, are useful for thinking about how Humeans view powers. For example, it is clear that dispositional properties have met with greater animosity from Humeans than have secondary qualities. Secondary qualities may be thought not to exist in any objective sense, but reference to them is treated with a degree of forbearance. Dispositional properties, by contrast, are apt to be renounced as occult. Belief in the reality of redness is charming; belief in the reality of powers is – or was until very recently – thought to be irrational, at odds with science and philosophy alike. As D. H. Mellor puts it, “(d)ispositions are as shameful in many eyes as pregnant spinsters used to be – ideally to be explained away, or entitled by a shotgun wedding to take the name of some decently real categorical property.”12 The third tenet of Humeanism, having to do with what causality is believed to be, is related both to the issue of powers and to questions about metaphysical necessity that arise once there is a ban on essence, or form. When we think of x as being the cause of y, our natural inclination is to think that x has the wherewithal to bring about y, albeit under specified conditions. For example, if we say “The fire is the cause of the toasted marshmallow,” we are likely to regard ourselves as saying that the fire is what made it be that the marshmallow is now so nicely browned; the fire did it. Fire, we are implicitly affirming, has a power, viz., the power to roast marshmallows. Causality itself, we will no doubt want to say, is something like the exercise or display of powers in general. In one way or another it is about activity, about the production of outcomes. But Humeans, because they do not believe in powers, cannot agree to such an account. What sense can be made of causing, then, if it doesn’t involve anything doing anything? Humeans – beginning with Hume himself and continuing to the

6

Ontology Revisited

present – give answers in which causality is equated with the fact of order. Hume’s version of the answer, as we will see in Chapter 2, is that what can be said to exist are impressions, some of which are constantly conjoined in our minds: first one, then another. There is, however, says Hume, no basis for the idea that such impressions are connected ontologically. Hume concludes that our concept of causal force therefore derives from our own expectations regarding such sequences. In reality, to cause something is nothing other than to be the impression that always, in our experience, comes first in our minds, vis-à-vis an impression that always follows, and to therefore be accompanied by a feeling of anticipation. This account of things is precisely what awoke Kant from his dogmatic slumbers, and even Hume saw that the outcome was both unsatisfying and unsettling. Subsequent Humeans, including Kant, have responded by tightening up the order that, from a non-Humean perspective, can be seen as standing in for causation. The strategy introduced by J. S. Mill, for example, is to say that because certain sequences do regularly obtain, we are entitled to infer that they always will. To cause is to be that which invariably comes first. Setting aside Hume’s own worries about the circularity of the regularity assumption (which Mill believed himself to have adequately addressed), if we go on to ask why the world is ordered in the way in which it is, as opposed to any other way, the common answer is the not terribly persuasive one that the order that the world does have is the one that conforms to the laws of nature. I describe the response as not terribly persuasive because, absent a claim that the laws of nature are expressions of God’s will, or some similar underlying anchor – absent such a claim, the laws of nature look to be just re-descriptions of the regularities in question, not explanations thereof. Recognizing this, empiricist Humeans other than Mill have undertaken in various ways to ground laws in something other than the very sequences that they ostensibly govern. Efforts of this type range from the suggestion that the laws of nature are those regularities that figure in the theories that we think are true, to the idea that laws are grounded in regular albeit metaphysically contingent conjunctions of universal properties. It is not too hard to see that these moves at best push the assertion of regularity back a frame. Invariant order is also at the heart of the Kantian approach, although Kant anchors the invariance transcendentally. Kant argued that due to the synthetic a priori operation of reason, we cannot help but experience the phenomenal world as subject to law-like determination – which is why it turns out that we don’t have to worry about it ever being otherwise. In claiming that causality is not simply a matter of regular succession plus the expectation thereof, as Hume would have it, or of pulling oneself up by one’s inductive bootstraps, à la Mill, but rather is a necessary condition of possibility of empirical experience itself – in saying this, Kant is obviously at odds with Humeans who are empiricists. Still, Kant too, when all is said and done, swaps out causality for order. Note that conceiving of causality as a matter of invariant order implies nothing about whether or not free will is consistent with such an order. It simply generates the problematic that frames the debate in analytic philosophy, as we’ll see in Chapter 4.

The Myth of Metaphysical Neutrality 7 Let me turn now to dispositional realism, the neo-Aristotelian position that has begun to take hold in contemporary metaphysics. The challenge to Humeanism as the reigning ontology in analytic philosophy has been several decades in the making. In the late 1960s and early to mid-1970s, there were a handful of philosophers who began to talk seriously about powers and dispositions, e.g., Rom Harre, George Molnar, Roy Bhaskar. The first major book-length arguments for a powers-based ontology were Rom Harre and E. H. Madden’s Causal Powers and Roy Bhaskar’s A Realist Theory of Science, both published in 1975. Bhaskar’s second book, The Possibility of Naturalism, quickly drew attention from nonpositivist social scientists and social theorists, spawning a now well-established school of thought called critical realism. Within analytic philosophy, however, the position began to gain significant traction only in the 1990s. Prominent, explicitly anti-Humean authors comprising the “second wave” of dispositional realism include Nancy Cartwright, Stephen Mumford, Brian Ellis, John Heil, Anjan Chakravartty and, with qualification, Alexander Bird – though this list is not exhaustive. Interest has intensified even further in recent years, marked by increasingly frequent conferences and special issues of academic journals. The analytic literature is now developed enough that it contains lively disagreements amongst dispositional realists over important issues. Nonetheless, once again a core set of commitments can be distilled. Referring back to what I’ve identified as the three main tenets of Humeanism, the powers-based alternative involves: (i) belief in irreducibly dispositional properties; (ii) what I have elsewhere called realism about causality; and (iii) for some, at least, essentialism. Let me consider each in turn. Dispositional realists affirm the existence of powers. As I shall discuss further in Chapter 2, contemporary Humeans are now beginning to claim that they believe in powers too – that it’s just that what dispositional properties are is something nondispositional. (Conversely, there are even some now who take themselves to be dispositional realist critics of Humeanism who define powers in terms of counterfactual stimulus-response sequences.) Let me be clear, then, especially for analytic readers, that by “affirm the existence of powers” I do not mean “affirm the existence of sequences of states of affairs (or counterfactual states of affairs) involving phenomena that are static, passive or inert, and then refer to such sequences as ‘powers.’ ” Such a move, and variants thereof, serve only to obfuscate. It is precisely in virtue of its proponents’ belief in phenomena that cannot be reduced to, or explicated in terms of, phenomena that are non-dynamic, that dispositional realism runs counter to the dominant post-Cartesian view that there is no such thing.13 Once powers are allowed back into the picture, causality can be defined in a fairly straight-forward, intuitively plausible fashion. While there is ample room for disagreement about how to spell it out, the basic idea amongst dispositional realists is that causality has to do with the expression of things’ powers. Given that from this perspective what a thing can do is related necessarily to what it is (even if the having of such a power is not essential to what it is), dispositional realists believe that the relation between a cause and that which it brings about, when it does, is one of metaphysical necessity, rather than one of either empirical

8

Ontology Revisited

regularity or transcendental synthesis.14 This affirmation has notable implications. One is that logically possible worlds, of the sort that abound in the contemporary analytic imagination (logically possible worlds in which sugar does not dissolve in water, for example) – such scenarios, with respect to discussions of metaphysics, show up for dispositional realists as the non-sequiturs that they are. Another implication concerns worries about induction.15 How do I know that sugar will dissolve in water next time, too? Because sugar and water are such that, c.p., sugar is soluble in water. Of course, the skeptic can press to know how it is that she is entitled to believe that the dispositional properties of sugar and of water won’t all at once be different tomorrow, i.e., how she can know that there will be sugar and water tomorrow. The issue requires a fuller discussion than I can provide here, but at a minimum a dispositional realist metaphysics allows the skeptic to rest secure in her kind-based knowledge that, if sugar does exist tomorrow, it will dissolve in water then, just as it does now. Finally, a third implication has to do with what we mean by laws, and whether or not there are such things. At a minimum, dispositional realists agree that if laws exist, they are not fundamental ontologically. Rather, if there are such things as laws, they are a function of the powers that things have, not the other way around. Stephen Mumford, for example, has gone a step further, arguing that belief in laws as phenomena in their own right is in fact superfluous, that there are only powers and the expressions thereof. This brings us to the issue of essentialism.16 For Ellis, whose book Scientific Essentialism was one of the early, systematic challenges to Humeanism in analytic circles, powers are intimately connected to natural kinds, members of which have kind essences as well as particular essences.17 This had been Bhaskar’s view too, 15 years earlier, although he did not explicitly distinguish between the essences of kinds and the essences of particulars. Bhaskar’s 1975 formulation was useful, however, in bringing out the basic idea, which is that the powers that a thing has, it has necessarily, in virtue of being the kind of thing that it is. Bhaskar used unabashedly essentialist language to make the case, referring to “the necessity implicit in the concept of a thing’s real essence.”18 For example: A thing acts, or at least tends to act, the way it is. It should be stressed that the difference between a thing which has the power or tends to behave in a certain way and one which does not is not a difference between what they will do, since it is contingent upon the flux of conditions whether the power is ever manifested or tendency exercised. Rather, it is a difference in what they themselves are; i.e., in their intrinsic natures.19 He concluded that “there is no conflict between explanatory and taxonomic knowledge. Rather, at the limit, they meet in the notion of the real essences of the natural kinds, whose tendencies are described in statements of causal laws.”20 Ellis’ account is more subtle on this point, as he distinguishes between kinds of substances, kinds of processes and kinds of properties, connecting causal powers only indirectly to substance kinds. Still, he called the view scientific essentialism,

The Myth of Metaphysical Neutrality 9 and has defended it against anti-essentialist critics. And Harre and Madden too, affirmed real essences in their work in the 1970s. But not all dispositional realists are essentialists. Chakravartty, for example, suggests that not all natural kinds have essences, and moreover that things can have dispositional properties that are accidental to their kind.21 Mumford has articulated a similar view – though he stipulates that he is simply as-yet un-persuaded, with respect to essentialism – arguing that one may affirm causal powers without being committed to a belief in essences.22 Heil, meanwhile, endorses powers but rejects real (as opposed to nominal) essences.23 All of this said, it will be important for dispositional realists to carefully parse the various notions of essence on hand, including but not limited to an Aristotelian notion of form and the Lockean notion of a real essence. I say this because the objection to essentialism is often that it commits one to positing the existence of an extra, mysterious-seeming “something,” and not just to the belief that things fall into kinds on the basis of what they are inherently or objectively. It is not obvious that the latter idea carries with it a commitment to an ineffable “something more.” Having contrasted the Humean and powers-based paradigms, let me close with a thought about what it might mean to be a dispositional realist with respect to sociological phenomena. Amongst those dispositional realists who are first and foremost analytic metaphysicians (as opposed to those who are critical realists, Marxists, and/or Aristotelians, for example), there is a tendency to fall back onto a reductive view of sociological entities as pluralities of individuals. This may be as much by default as by conviction, a product of the division of labor between metaphysics and social theory, compounded by the continued predominance of Humeanism in analytic philosophy of social science and liberalism in political philosophy.24 Whatever the explanation, there is good reason to take one’s cues from thinkers outside of the specialty of metaphysics in this instance. It may be helpful to come at the question negatively, by clarifying what it would not mean to be a dispositional realist in this domain. Most important, it would not require one to say that sociological phenomena are as they are “by nature,” or that such phenomena are somehow not thoroughly social. Rather, it would be to say (a) that sociological entities exist as bearers of emergent, sociological rather than psychological powers, and (b) that at least some kinds of sociological phenomena have essences, on the basis of which they fall into natural kind analogues. It’s a further question – and an interesting one – which sociological phenomena do have essences, and which do not. I am inclined to think that capitalism does, and patriarchy and racism too. I’m not sure that collective identities do. Note that there is no reason to imagine that social kinds would have to exist a-historically. Kinds of sociological phenomena can and do come into being in the course of history, can and do pass out of being. Finally, to say that there are at least some social kinds, and that it is possible to identify them, is not to suggest that in doing so one would be engaging in social science that is value-neutral. Ellis has claimed that dispositional realism has far-reaching implications for philosophy as a whole. We disagree about what these are, especially in relation to the philosophy of social science (although his views on the matter seem to be

10

Ontology Revisited

evolving), but he is right, in my view, to think that it is not just debates within metaphysics that have been configured, implicitly or explicitly, by mechanism. Accordingly, my aim in the chapters to come is to illustrate the role that the Humean ontology has played in modern and contemporary social and political philosophy. Once we are very clear about the hold that Humeanism has had, we will be in a position to take full advantage of the opening that has been created by dispositional realist work in metaphysics.

2

Hume Custom as Metaphysical Necessity

Humeanism as I have defined it figures in the chapters to come mainly as an ideal type, a heuristic for talking about the deep structure of modern and contemporary debates about agency and the social. But with David Hume himself we have a figure whose thinking so embodies the ideal that it gives rise to the name. In addressing Hume’s views, therefore, I cannot help but flesh out the account of Humeanism given in Chapter 1. Let me emphasize that in so doing my aim is not to advance a new reading of Hume – or, more to the point, to defend an old one, although I shall be forced to do so. Rather, my objective is to illustrate the role that Hume’s own Humeanism plays in his social and political thought. In terms of my argument as a whole, the analysis is intended to do two things. First, I want to introduce something along the lines of a philosophical control group: Hume’s own work I shall take to be Humeanism in its purest form. Second, I want to demonstrate that even Hume can’t talk about actual agents, or the social relations into which we enter, without straying from his stated metaphysical commitments (indeed, it’s not clear that he can even do metaphysics without engaging in talk of powers, at least).

Hume on Causality, Powers and Necessity It has become difficult in recent years to give an account of Hume’s ontology without entering into controversy. As the Humean picture has come under increasing criticism, defenders have begun to suggest either that Hume did not reject realism about powers after all, and/or that, regardless of what Hume himself may have thought, contemporary Humeans are perfectly entitled to talk about powers in realist ways. I shall argue against both of these moves. Hume himself, I believe, presents readers with no significant interpretive challenges with respect to his disavowal of dispositional properties. And the thought that one might agree with Hume and yet help oneself to powers-talk as needed – such a thought, meanwhile, can be only confused or disingenuous. Given the thorniness of the discussion, let me begin by asserting the following, which I trust is indisputable: Hume is adamant that the idea of causation does not mean what we think it does. It doesn’t mean what his philosophical predecessors thought it meant, and it doesn’t mean what non-philosophers naïvely take it to mean, either.

12

Ontology Revisited

The philosophical view that Hume is most keen to reject is that causation which was held by early modern rationalists, which is that causation is a form of logical or conceptual necessity. This position is different from the ordinary, non- or pre-philosophical view, which Hume also rejects. The latter is held by philosophers too (by Aristotelians, paradigmatically), but first and foremost it is what we might call the default, regular-person view, widely if not universally held by anyone who hasn’t, with Hume’s help, had it trained out of them. The regular-person/Aristotelian view, described in Chapter 1, is that causation is about what things can do. Soap plus hot water dissolves the hold of the rice, dried on to the plates from dinner. I recently hung an old temple bell on a nail in my sunroom, where it therefore now hangs. This is the view that figures in dispositional realism and scientific essentialism. Hume rejects the rationalist account on the grounds that it is always possible to imagine a hypothetical scenario in which a putative cause does not usher in the effect with which it is normally associated. Given that such speculative imaginings do not, Hume contends, involve one in any kind of contradiction, then, ipso facto, it cannot be the case that the relationship between cause and effect is one of logical necessity. Hume writes: When I see, for instance, a Billiard-ball moving in a straight line towards another; even suppose motion in the second ball should by accident be suggested to me, as the result of their contact or impulse; may I not conceive, that a hundred different events might as well follow from that cause? May not both these balls remain at absolute rest? May not the first ball remain in a straight line, or leap off from the second in any line or direction? All these suppositions are consistent and conceivable. Why then should we give the preference to one, which is no more consistent or conceivable than the rest? All our reasonings a priori will never be able to show us any foundation for this preference.1 As a corollary of this point, Hume tells us that the ideas that enter into the relation of cause and effect are replications of impressions that are completely separate from each other: “In a word, then,” he writes, “every effect is a distinct event from its cause.”2 The main argument against the Aristotelian view, meanwhile, against the view that causes actually do something (I’ll refer to this as a substance causation view, broadly construed, or otherwise productive causation), is that, try as he might, Hume can identify no impression that would serve as the perceptual basis for it. Hume makes this point adamantly and repeatedly. Here is one instance: “All ideas are deriv’d from, and represent impressions. We never have any impression, that contains any power or efficacy. We never therefore have any idea of [such a] power.”3 Finding no impression that could be the vital original from which the idea of a productive cause might be derived, he concludes that what we take to be such an idea is in fact a misrepresentation of an entirely different kind of experience, namely one of subjective expectation reinforced through habit. I will come

Hume: Custom as Metaphysical Necessity 13 back to the proposed alternative in a moment, but let me say just a little bit more about how the case against powers is put together. In the Enquiry, especially, the hunt for the impression of a causal power is part and parcel of the argument that causation is not a species of logical or conceptual necessity. That is, complementing the claim that causes and effects can be separated in thought with no resultant logical contradiction, and so must be thought to be only contingently related, comes the claim that, what is more, the idea that they are somehow non-contingently related itself has no perceptual basis. But in saying that he can perceive the billiard balls but not any kind of connection between them, Hume is not just weighing in on the question of whether or not causality is a form of logical necessity. The argument is doing double duty. It is introduced to counter the rational necessitation view, but it also supports what I’ll call an antipotency thesis, lodged against those who would contend that a causal connection is when the consequent is actually brought about by the antecedent. Causes can’t be coherently thought to produce their effects, the argument goes, because there is no possible basis in experience for the idea that causes and effects are joined in any way other than that of customary association in the mind. The case against substance or productive causation also draws on a claim to the effect that nothing about the features or identity of a cause tells us which effect is likely to follow it: “The mind can never possibly find the effect in the supposed cause, by the most accurate scrutiny and examination. For the effect is totally different from the cause, and consequently can never be discovered in it.”4 Here, too, Hume is primarily arguing against early modern rationalists. But the claim, if true, implies not just that causal statements are, as Hume would put it, matters of fact rather than of deduction; it implies, too, that whatever it is that makes paper burn, or starch dissolve, it has nothing to do with the properties of fire, or of soap and water – nothing, that is, to do with the nature or operation of the so-called “cause.” I stress this in order to bring out the full strangeness of the view, its distance from the common idea that causes actually cause their effects. Here is Hume again: “When we say, therefore, that one object is connected with another, we mean only that they have acquired a connection in our thought.”5 The centerpiece of the argument, though, as I’ve said, is Hume’s repeated insistence that, as with the idea of a necessary connection, there is no impression to which an idea of a power might correspond. And where there is no impression, there can be no intelligible concept: causality can’t be about the action or influence of things on other things because the very concept of a power, a capacity-to-effect, lacks meaningful content. Having dispensed with both (a) the false belief that what connects causes and effects is logical or conceptual necessity, and (b) the meaningless claim that causes do or are something that brings about effects, Hume offers up what he takes to be the only viable alternative. Causation is when a given impression, in the mind of a subject, is always followed by another. It is when two impressions are constantly conjoined, the mind passing immediately from the one to the other, as Hume likes to put it. The impression that, it turns out, gives phenomenal content to the misconception that causes are tethered in some way to “their” effects is simply our own feeling of expectation, based on what Hume calls custom, or

14

Ontology Revisited

habit. If, in the past, the thought of x has always been followed, in one’s mind, by the thought of y, then one comes to anticipate the thought of y, upon thinking x. This connexion, therefore, which we feel in the mind, this customary transition of the imagination from its one object to its usual attendant, is the sentiment or impression from which we form the idea of power or necessary connection. Nothing farther is in the case.6 And again: Upon the whole, necessity is something, that exists in the mind, not in objects; nor is it possible for us ever to form the most distant idea of it, consider’d as a quality in bodies. Either we have no idea of necessity, or necessity is nothing other than that determination of the thought to pass from causes to effects and from effects to causes, according to their experienc’d union.7 Unfortunately, we then project this expectation onto the external world. We want to say that if x’s and y’s are causally related, then they are somehow actually connected, joined in some way other than that we always pass from the thought of the one to the thought of the other. But any such belief is simply a confused representation of the feeling associated with the presumption that our thoughts will continue to come to us in the same order to which we are accustomed. All that we are properly entitled to say is that the thought of soapy water is always followed by the thought of the rice starch on the dinner plates dissolving. There is nothing more to it. At least, that’s what Hume claims – over and over again, and in no uncertain terms. And he says that it’s alarming. And Kant thought that it was alarming too, for what it’s worth. But of course there is more to it. For one thing, there’s the issue of the “always.” At the point at which he explicitly defines the term “cause,” Hume says: Similar objects are always conjoined with similar. Of this we have experience. Suitably to this experience, therefore, we may define a cause to be an object, followed by another, and where all the objects similar to the first are followed by objects similar to the second. Or in other words where, if the first object had not been, the second never had existed.8 The second formulation is clearly different from the first, in that it includes a counterfactual, the presence of which re-introduces the idea, or hope, that there is, for Hume, a genuinely modal connection between causally related thoughts. Maybe it’s not just that we do, always, pass from the thought of x to the thought of y; perhaps, after all, we must.9 And sure enough, those who want to align themselves with Humeanism, but who are not prepared to endorse a regularity theory of causation, conclude from this passage that a counterfactual theory is what Hume must have meant all along.10 Now, in my view, Hume holds exactly the position that he says he does, and which I have just outlined, viz., that causation is exhausted by the constant

Hume: Custom as Metaphysical Necessity 15 conjunction of ideas in the mind, plus a feeling of expectation. I therefore do not take his second formulation of the definition of cause in the Enquiry to be evidence that he was committed to a counterfactual theory instead, or that he somehow held both a regularity theory and a counterfactual theory. I think it points us elsewhere, if anywhere at all. Certainly it constitutes emphasis. x and y are always conjoined. Beyond this, however, it flags a temptation to account for the “always,” to ground it ontologically. That is, though he doesn’t seem to be particularly vexed by the issue, Hume can’t help but implicitly raise the question of why given thoughts are always followed by given others. I take his honest answer to be “As it happens, reality – in its natural and social dimensions alike – is ordered in a way that is regular, which regularity is mirrored in our impressions.” Unfortunately, such an answer isn’t properly available to Hume, given what he says external objects (i.e., substances) themselves are, viz., impressions. Strictly speaking, we are stuck with the mere fact of ordered perception. Even so, what the counterfactual formulation tells us is that, in Hume’s view, said de facto order is a fixed, definite, determinate one – set in stone as it were, except not. The passage in question is not the only source of ambiguity, however. Another is that Hume at his most emphatic seems to suggest that the reason why there are no impressions of powers is that there are no such things as powers. It is simply an ontological given, happy or no, that the world does not contain them. Now, arguing in this introduces a complication because, in addition to the issue of his phenomenalism, Hume maintains that the very concept of a power is unintelligible. And as Thomas Reid pointed out, if Hume believes the concept to lack meaningful content, then he is in no position to deny the existence of that to which it ostensibly refers: “for how can men have any opinion, true or false, about a thing of which they have no idea?”11 This thought, in turn, may invite one to think that if Hume can’t coherently deny that there are powers, then he ought not to be seen as rejecting them. Even if Reid is right, however, and he probably is, Hume is only barred from explaining why he has been unable to have an impression of a power. It doesn’t give us any reason to think that Hume really does, after all, believe in the existence of powers – or, more to the point, that he would believe himself to have a way to say so, if he did. A more serious problem, though, is that Hume just does, sometimes, talk as though he believes in powers. This has fueled what has come to be known as “the new Hume” debate on the topic, including a volume by that title, edited by Rupert Read and Kenneth Richman.12 What’s new about the new Hume is that he is thought to have either (a) defended the existence of real causal powers (i.e., been a realist about causal powers); (b) believed in them but not to have claimed for the belief the status of knowledge, and/or believed in them but claimed to have known nothing about them beyond their existence (i.e., been a skeptical realist about causal powers); or (c) been entirely agnostic on the question, just as open to the existence of powers as to their non-existence, and holding no other views that would preclude the affirmative option. There is no question but that the old Hume helps himself to realist talk about causal powers. Sometimes, he specifically stipulates what he actually means, so that there can be no confusion – as in the footnote in Section IV of the Enquiry13

16

Ontology Revisited

– but other times he leaves us without an on-the-spot corrective. Thus the question arises: which Hume is the “real” Hume? Does Hume affirm (or at least in principle allow for) the existence of causal powers, or does he not? One way to decide the issue would be to observe that Hume himself tells us that the Enquiry is a reliable source, for anyone who wants to know what he thinks, and to point out that the meaning of the footnote in Section IV is unambiguous: use of the word “powers” in an everyday sense is not meant to be taken as implying a philosophical commitment to the existence of such things. The footnote refers us ahead to the analysis of Section VII, which comes relatively late in the relevant portion of the text and is not subsequently retracted or modified. Anyone with lingering doubts about Hume’s considered position, one might suggest, should simply pay very careful attention to what Hume actually says. When he uses the word “powers,” he explicitly reminds us in Section 4 (lest we be distracted by his rhetoric), he does not actually mean powers. He means regular sequences of mental states. Nor is there any good reason, one might add, to think that Hume would have intended the clarification to hold only in the case of one passage in Section IV of the Enquiry. This seems viable to me as an interpretive strategy, though I don’t mean to suggest that the case hinges on the one footnote alone. Hume directs the reader on to the full argument of Section VII, and with good reason. Unless one is being willfully obtuse, it is just not possible to conclude that the take-away thesis of the Enquiry as a whole is that causality involves the exercise or display of powers or dispositions. Still, what about (c), the idea that Hume is an agnostic about the existence of causal powers, and that it therefore remains an option for him – or for any Humean – to believe in them? The correct response here, I think, is to remind oneself that Hume never equivocates about whether or not the very concept of a power, on any normal usage of the term, has meaningful content. He says that it does not. Thus, as Georges Dicker concludes, citing Kenneth Winkler, Hume can’t be said to be amenable to the existence of powers because in order to coherently assert that powers exist, it would have to be possible to have an impression of a power — and this is precisely what Hume denies.14 The claim that Hume is open to the existence of powers is (therefore) a non-starter.15 Nonetheless, there is an important difference between thinking that Hume meant more or less what he says he meant, and thinking that his treatment of the issue is seamless. It seems to me incontrovertible that, with respect to causation, Hume takes himself to be saying something that has not been said before, and that he believes it to be something that is irreconcilable with our felt convictions. This should rule out any impulse to read Hume as saying nothing out of the ordinary, nothing that runs counter to our pre-philosophical intuitions. But while it rules out a Humeanism that ought not disconcert, it doesn’t yield a position that is without contradictions. Some are of a type that may be perturbing only to logicians and philosophers of language. For example, Hume makes use of the term “power,” as it is commonly understood by English speakers, in order to claim that the concept is unintelligible; similarly, he refers to impressions of powers in order to claim that no such impressions exist. These would seem to be instances of performative

Hume: Custom as Metaphysical Necessity 17 self-contradiction, and they no doubt tempt particularly analytically minded readers to conclude that Hume thinks the opposite of what he says he does. But there are more substantial tensions, too. One I’ve already mentioned: I suspect that Hume believes that the reason why there are no impressions of powers is that there are no such things as powers for there to be impressions of – this whilst maintaining that “things” themselves just are impressions. Another is that Hume, I think, like everyone else, can’t help but believe, when all is said and done, that there is an objective difference in kind, and not just in number of constituent conjunctions, between causal connections and correlations. This leads him to acknowledge our felt sense of the necessity involved in the causal relation, at the same moment that he is arguing that the nature of the relationship is not as it seems, i.e., that it is in fact contingent to the core. Finally, be it in the context of his analysis of causation or in the course of setting out his moral, social and political philosophy, Hume really does avail himself of dispositional language. As I’ve said, I don’t think that we should conclude from his rhetoric that he either does or could consistently believe in powers. I do think, however, that it is philosophically significant that he can’t stay away from such talk. Especially, he can’t avoid saying things that hang on the concept of a generative cause, a doing. That it may not be possible, in the end, to get out from under the idea of substance causation is, in my view, a root source of the contradiction. A different kind of effort to have one’s Humeanism without giving up causal powers depends not upon the invocation of a new Hume, but instead upon a standard move in analytic philosophy called a reductive analysis. According to the rules of a reductive analysis, if one equates x’s with y’s, and if one affirms the existence of y’s, then one is entitled to say that one believes in the existence of x’s. Here, the Humean simply declares that powers “just are” sequences of states of affairs involving nondispositional properties (or any one of the many Humean variations on this theme), and then says that since she believes in such sequences, she believes in causal powers – and is therefore free to help herself to realist talk about powers and/or dispositional properties. Metaphysics is like science, explains the Humean: it turns out that we were just mistaken about what powers are. What they really are, are regularities. The move is an important one, because it allows the Humean to dodge the philosophical consequences of denying the existence of powers, by seeming not to have issued such a denial.16 And to be clear, denying the existence of powers is precisely what she is doing. In identifying powers with regular sequences (or some other Humean equivalent), she is not saying that regularities are actually powerslike; rather, she is saying that powers are really something that is not powers-like. At a minimum, there are two things to be said by way of response. The first is this: even if we grant the Humean the conceit that the reduction of powers to nonpowers is a version of a theoretical advance in the natural sciences, such advances in the natural sciences come at real ontological costs. Sooner or later, one may no longer talk about the existence of phlogiston. In the case of an analytic reduction, one criterion for determining whether or not the original entity has been preserved within the new “theory” surely must be whether or not the replacement entity retains the essential properties of the original. The claim that powers are in reality

18

Ontology Revisited

something that is passive or inert – this claim, it seems to me, demands of its proponent that she have the courage of her metaphysical convictions. Something that is necessarily non-active is not what a power or disposition could be, and still be a power or disposition. One cannot simply stipulate that dynamism is actually non-dynamism: if one has concluded that the world is not dynamic, then one’s view is contrary to the view that it is dynamic, not identical to such a view. Second, the reduction of powers to not-powers depends for whatever plausibility it has in analytic circles on the issues being abstracted away from the substantive metaphysical debates in which they originally arose. In saying such a thing I take my cue from Alasdair MacIntyre, who has shown us all how to employ the history of philosophy in the service of dissolving various contemporary preoccupations. The question of whether or not powers exist, and if so what their relationship might be to law-like regularities, was central to early modern philosophy. As Walter Ott has put it, the period can be schematized just as well in terms of a divide between those who defended an ontology of vis (force) and those who affirmed the new ontology of lex (law) as it is at present in terms of a debate between rationalists and empiricists.17 Ott calls the Aristotelian, vis-based picture a “bottom-up” account of causation, the lex-based picture “top-down.” In the Christianized version of the bottom-up view, God is thought to have endowed substances with causal powers, given by their forms.18 On the top-down view, by contrast, matter has no such intrinsic property. It is inert. Causal interactions are therefore a function of Divine command; matter does what it is required to do by laws that are given by God. What is instantly apparent, even on the basis of this barest of a sketch, is that the powers-based ontology inherited from Aristotle is altogether different from the law-based ontology issued in by the rise of the mechanical sciences. “Is the material world inherently potent or inherently inert?” is a full-blown metaphysical question. Answers to it reflect fundamentally different pictures of the nature of things. When the question is viewed in context, it simply does not plausibly admit of the answer that matter is indeed potent – but that what we mean by potent is impotent.19 I have focused thus far on Hume’s treatment of powers, rather than on his views about necessity. The concepts are related, but they are not the same. (We learn from Kant, for example, that it is possible to believe in a form of synthetic a priori necessity without believing in causal powers.) With respect to necessity, the issue is not one of potency, but contingency – of whether or not the causal relationships between kinds of things could be otherwise than as they are. Hume’s thinking about necessity is ambiguous. Having rejected necessary connections, one would expect him to hold that there is no such thing as a connection that is necessary! But while he says something close to this when he is talking about causation, it’s not the spin he gives in the Treatise, especially, when he is talking at length about necessity as such. Nor is it what he says in the Treatise or the Enquiry when he is talking about necessity in relation to human agency. Instead, and this is significant, he redefines necessity such that it no longer means “necessary” at all. Rather, he claims, it means habitual, customary, in keeping with the present order of things. “I define necessity two ways,” he says; “. . . I place it either in the constant

Hume: Custom as Metaphysical Necessity 19 union and conjunction of like objects, or in the inference of the mind from one to the other.”20 And again in the Enquiry: Our idea, therefore, of necessity and causation arises entirely from the uniformity observable in the operations of nature, where similar objects are constantly conjoined together, and the mind is determined by custom to infer the one from the appearance of the other. These two circumstances form the whole of that necessity, which we ascribe to matter. Beyond the constant conjunction of similar objects, and the consequent inference from one to the other, we have no notion of any necessity or connexion.21 For the purposes of my own argument, the key point here – to which I shall return below – is that the re-assignment of meanings allows Hume to ascribe necessity to various aspects of moral, social and political life, purely on the grounds of their being long-established, and therefore expected. With this in mind, let me move on to Hume’s analysis of what I’ll call the social, beginning with his account of individual and collective entities, followed by his theory of justice.

Persons, Collectivities and Norms Hume’s reflections on personal identity have the same analytic structure as his reflections on causation. As with casual powers and necessary connections, the claim is that there is no impression to be had that could ground the idea of a self. If, in the case of powers and connections, what we’ve done is impute to the world imaginary energetic forces and imaginary bonds sustained by such forces, in the case of the idea of a self what we’ve done is ascribe unity to something that is not actually unified. It’s an understandable mistake, Hume says, but a mistake all the same. And it’s a cause for concern because it leads us to posit the existence of an imaginary entity, viz., the so-called “self.” As he puts it, “the controversy concerning identity is not merely a dispute of words.”22 What constitutes personal identity, then, if there is no such thing as a self? Famously, Hume’s answer is this: “I may venture to affirm of the rest of mankind that they are nothing but a bundle or collection of different perceptions, which succeed each other with an inconceivable rapidity, and are in a perpetual flux and movement.”23 Thus: “[t]he identity, which we ascribe to the mind of man, is only a fictitious one.”24 Moreover, it turns out that the self is not unique in this regard: its identity is “of a like kind with that which we ascribe to vegetables and to animal bodies.”25 In fact, there is no such thing as substance period, for Hume, or forms thereof. In all cases, the identity of objects is a simply a matter of the mind moving customarily from one recalled perception to another. Indeed, the only difference, it would seem, between entities that exist and entities that do not exist is whether or not the habits of mind that comprise them are in turn said to be “of” them.26 The first and most obvious point of interest here is the very fact of Hume’s antiessentialism. As noted in Chapter 1, it is a mistake to think that anti-essentialism

20

Ontology Revisited

marks the rise of post-modernity, or that it is an innovation of later twentiethcentury thought. On the contrary, it is quintessentially modern. And Hume’s antiessentialism has immediate implications for his social and political thought. One has to ask, for example, whether or not a succession of perceptions is robust enough, ontologically, to do the work that Hume wants selves to do. But it’s not just the anti-essentialism that’s of note; the rest of the analysis requires our attention too. While the bundles of perceptions that we call selves are not actually unified, the impressions in question are nevertheless related, says Hume. Specifically, our minds pass regularly from one perception to another, amongst the collection of perceptions the totality of which we confusedly take to be a “self.” As he puts it: “we may observe, that the true idea of the human mind, is to consider it as a system of different perceptions of different existences, which are link’d together by the relation of cause and effect.”27 In terms of my own argument, what is significant here is that we cannot begin to understand Hume’s notion of personhood without understanding what he calls in the Enquiry “Relations of Ideas” (in the Treatise, the “connexion or association of ideas”) – and this because what we imagine to be a self in reality just is a set of habitually associated impressions. In elaborating further what the purported self is and is not, Hume reports that the phenomenon to which it bears the most similarity is a republic, or a commonwealth. “In this respect,” Hume writes, I cannot compare the soul more properly to any thing than to a republic or commonwealth, in which the several members are united by the reciprocal ties of government and subordination, and give rise to other persons, who propagate the same republic in the incessant changes of its parts. And as the same individual republic may not only change its members, but also its laws and constitutions; in like manner the same person may vary his character and disposition, as well as his impressions and ideas, without losing his identity. Whatever changes he endures, his several parts are still connected by the relation of causation.28 Selves and societies are on the same footing, then, metaphysically. Of course, this way of putting it may be misleading, since the point that Hume is trying to make is that there is no such thing as a self. If societies are what the bundles of perceptions that we mistake for selves are most like, then it would seem to follow that societies must not exist either. Such a view is certainly consistent with the common empiricist rejection of sociological holism. Unlike a standard atomist, though, Hume doesn’t argue in a way that privileges individuals, ontologically, vis-à-vis social formations. He doesn’t suggest, for example, that societies reduce to aggregates of individuals because an impression of a body is one that it is possible to have, whilst an impression of a society is not one that it is possible to have. Rather, Hume’s view is that the pseudo-entity “society” and the pseudo-entity “self” both reduce to the relationship between ideas that he calls cause and effect.

Hume: Custom as Metaphysical Necessity 21 Moreover, what is interesting is that in both cases Hume can’t seem to avoid talking about the fictitious entities in question as though they were real. Non-existent selves and non-existent societies have, it would seem, discernible characteristics – stable ones at that – including dispositional properties (though only apparently dispositional, Hume will have to say). In fact, it would seem that non-existent sociological entities even have emergent apparently dispositional properties, i.e., ones that are not held by or reducible to those held by single nonexistent selves. For example, governments, according to Hume, have the property of eliciting the sentiment of allegiance, and are the necessary precondition for the obligation of promises. Society as a whole, meanwhile, has the property of eliciting moral regard for justice. As in the case of selves, an important question will be whether or not bundles of associated ideas can really do such things; for the moment, though, I will note only that Hume is not an atomist in the standard sense of being an ontological individualist. The next question is whether or not “selves” have free will. Hume rejects the Doctrine of Liberty, as he calls it, and affirms the Doctrine of Necessity. In the context of the contemporary debate over free will, Hume is deemed a compatibilist – the classical compatibilist, even – i.e., someone who holds that freedom is consistent with determinism. Chapter 5 is devoted entirely to an analysis of the imprint of mechanism on the free will debate, but the point I’d like to make here is that Hume’s compatibilism hangs upon his peculiar definition of necessity as custom, i.e., as something that is precisely not necessary. Attention to this fact will help to bring into focus just what Hume is and is not saying about agency, when he says that it is marked by necessity. Hume’s argument against the Doctrine of Liberty is this: (1) to say that actions are necessary is to say that they fall into regular patterns; (2) they do fall into regular patterns; (3) therefore they are necessary. In both the Treatise and the Enquiry, he has, from his perspective, already established the first premise by the time he gets to the topic of free will. The weight of the argument therefore falls upon the second premise. Hume advances three main claims in its support. First, he says, if one observes human behavior, one will see that it is law-like. Where it appears not to be, the observer has made an error. Second, although it may seem to oneself that, in any given situation, one need not necessarily have acted as one did, one’s subjective experience of choice is not decisive. Rather, what is decisive is that an outside observer would have passed, in their own thoughts, from what one did first to the thought of what one did second. Even if one were to intentionally try to behave unpredictably (in order to demonstrate that one could, and to thereby prove the Doctrine of Liberty true), one would simply be behaving predictably according to that very objective. Here’s Hume: We may imagine we feel a liberty within ourselves; but a spectator can commonly infer our actions from our motives and character; and even when he cannot, he concludes in general, that he might, were he perfectly acquainted with every circumstance of our situation and temper . . . . Now this is the very essence of necessity, according to the foregoing doctrine.29

22

Ontology Revisited

Finally, Hume claims in the Enquiry that if behavior were not regular, it would have to be governed by chance. But chance, he says, is “universally allowed not to exist.”30 Thus all behavior is necessary. But – and this is the crucial part – it’s not really necessary, not if one means by necessary that something must be a certain way. Hume’s contention is not that at all. On the contrary, his view is simply that all behavior just does, invariably, conform to the expectations of observers – and when it doesn’t, it would have, had the observer been fully informed. This, it must be said, is a remarkably weak form of determinism, if it is one at all. It’s a real question, therefore, just what it is that Hume may be thought to have contributed to the effort to reconcile a belief in free will with a commitment to necessitation with respect to agency. The salient point for the purposes of my own argument, however, is merely that vis-à-vis any claim advanced by Hume asserting the necessity of anything, “necessary” does mean not necessary, but rather “de facto the case” – or, as Hume sometimes puts it, “customary.” Having flagged these basic issues, we can now ask how Hume’s metaphysics matters for the rest of his account of social and political life. Hume himself repeatedly asserts that his views form a coherent whole. It shouldn’t be a surprise, therefore, if they do. The metaphysical picture is one in which the basic ontological units are impressions, impressions that are organized by the imagination into the objects that Aristotelians (and others) mistake for genuine substances. Since most contemporary Humeans understand themselves to be materialists, I shall not pursue the issue of whether or not impressions, in Hume’s view, are backed up in some way by a truly external, external world. I will say again only that, at a minimum, those phenomena that we are entitled to say exist are those that reduce to impressions that it is possible to have. Powers do not meet this criterion, according to Hume. Objects, therefore – or, “objects” – cannot be coherently thought to be imbued with dispositional properties.31 Nor can objects be thought to have essences. Finally, objects, such as they are, are causally connected only in the sense that (a) some of our impressions are regularly conjoined with others, and (b) in such cases, we expect the impression that always comes second to follow the upon the one that always comes first. Hume’s social and political theory is entirely in keeping with this picture. At the most basic level, society just is a set of constant conjunctions, no different in this respect from other pseudo-entities, such as selves, or from other, “real” entities for that matter. Those features of social life that cannot be resolved into relations of ideas (e.g., moral agency) Hume construes as brute, inexplicable, affect-laden facts, which themselves fall into constant conjunctions, “necessary” ones even. In short, Hume’s social ontology is a direct expression of his metaphysics. Politically, meanwhile, it follows from his metaphysics that custom will be taken to be foundational, and that present norms will be viewed as modally rather than instrumentally necessary. And that is just how his thinking does go. Indeed, As MacIntyre suggests, Hume’s objection to those who argue from is to ought is only that they do it badly.32 Even if one wants to say that in the Enquiry he intends only to describe what people do assume to be just (some people, at least; Levellers, for example, who called for a redistribution of land, he believes to be irrational) – even if one maintains that in

Hume: Custom as Metaphysical Necessity 23 the Enquiry there is no suggestion that the is-ness of that which is contributes to its ought-ness, the bottom line even there is that in fact everything that is, ought to be. And in the Treatise, there can be no mistaking the nature of the move. That “is” translates into “ought,” for Hume, can be seen most clearly in the course of his discussion of justice. Hume defines justice as an intersubjective agreement to recognize claims regarding the ownership and control of property – to recognize mine and thine, as Hobbes would say. The spirit of the agreement is as follows: “I observe, that it will be for my interest to leave another in the possession of his goods, provided he will act in the same manner with regard to me.”33 The state (or “government,” in Hume’s terminology) is a way to ensure that such claims are respected. Justice arises in the natural course of things, according to Hume, because (a) society is in people’s self-interest, but (b) social cooperation requires institutionalized property relations as a condition of possibility. Indeed, the benefit of society is so obvious, Hume says, that it makes no sense to imagine any real human beings as ever having been outside of one. Nonetheless, Hume calls justice an artificial virtue. Although it is natural for us to respect property, such respect is an indirect, mediated passion, one that is triggered by society (the very entity that it underwrites), rather than being an a-social sentiment of individuals qua individuals. But Hume doesn’t leave off with the claim that people will invariably seek to stabilize property, and in time come to feel that such an arrangement is morally good. As he puts it: “Tho’ the establishment of the rule, concerning the stability of possession, be not only useful, but even absolutely necessary to human society, it can never serve to any purpose, while it remains in such general terms.”34 The required next step is to determine how “particular goods are to be assign’d to each particular person, while the rest of mankind are excluded from their possession and enjoyment.”35 This step both constitutes property ontologically, and gives content to the theory of justice. Hume’s handling of the issue is also a prime example of how the analysis of Book I of the Treatise is carried over into that of Book III. What, then, is property? Hume’s answer is not what one might expect, no pun intended. “Property,” he says, “does not consist in any of the sensible qualities of the object.”36 Instead, it is an “internal relation” between objects and subjects. More specifically, it’s “some influence, which the external relations of the object have on the mind and actions.”37 What influence? Precisely that of eliciting a feeling of being obliged to leave the object in the hands of its owner, or to return it thereto – which actions, Hume contends, “are properly what we call justice.”38 And the task of assignation follows: if objects become property when and insofar as they are felt to rightfully belong to someone, the question is to whom they so belong.39 Hume tells us that there are five situations in which the associations involved constitute property. The first of these is that of “present possession”: if you already possess something, it’s rightfully yours. This situation is in all respects the same as the second, which Hume calls occupation, except that occupation – along with the remaining three – are thought to hold only once there is a society established, and not just at an imaginary moment of founding. Hume defines possession itself

24

Ontology Revisited

as being “a species of cause and effect,”40 – which is to say that possession, like property more generally, is when there is a constant conjunction between the idea of the owner and the idea of the object in question. But why should such a conjunction carry any normative weight? Hume’s answer is that it is “the effect of custom, that it not only reconciles us to any thing we have long enjoy’d, but even gives us an affection for it, and makes us prefer it to other objects, which may be more valuable, but are less known to us.”41 As a result of this, “ ’Tis evident . . . that men wou’d easily acquiesce in this expedient, that everyone continue to enjoy what he is at present possess’d of.”42 It might appear, then, that property via the natural expedient of present possession – and later via occupation – is grounded in consent. Significant, however, is why we consent. It looks as though Hume’s view is simply that everyone says to themselves “I like my own belongings; therefore I’m in favor of all of us getting to keep what we already have.” But that’s not quite it. The passage that I have just cited is followed by a lengthy footnote, wherein Hume says that “I suspect that these rules are principally fix’d by the imagination.”43 Hume then goes on to remind us that, in his view, “when two objects appear in a close relation to each other, the mind is apt to ascribe to them any additional relation, in order to compleat the union.” At this point he directs us to refer back to Book I, Part 4, Section 5, which is the section entitled “Of the Immateriality of the Soul.” The argument in the rest of the footnote, as in the referenced section, is that the imagination easily fuses together things that are already closely associated. It is this propensity of the imagination that leads us to connect the idea of present possession with the closely related idea of property itself, i.e., with the idea of rightful possession. In short, the underlying reason why we ought to possess what we do presently possess is that “we” are in the habit of thinking that we should. The surprise is not that Hume thought this, but that his argument for it falls out from his account of the relations of ideas – which, given his phenomenalism, just is his metaphysics. The three other situations in which the associations involved constitute property are prescription, accession and succession. Prescription, or long possession, is when property is constituted by the effect of the passage of time on sentiment. In this case, repetition of the constant conjunction of the idea of the owner and the idea of the object is itself repeatedly constantly conjoined with a feeling that the possession is rightful. “[P]roperty being produc’d by time, is not anything real in the objects,” Hume says; rather, the set of reiterated associations comes to be stronger – i.e., more customary – than are those that would constitute a competing claim of ownership.44 Property by accession is defined and legitimated in terms of constant conjunctions too. In this case we rightfully own that which is (a) in close proximity to, (b) lesser than, and (c) associated in our minds with that which we already own. “Thus the fruits of our garden, the offspring of our cattle, and the work of our slaves, are all of them esteem’d our property,” according to Hume, “even before possession.” This passage immediately brings to mind Locke’s claim that he owns the grass that his horse has bit and the turfs his servant has cut.45 In order to lay claim to rightful ownership, however, the Lockean has to

Hume: Custom as Metaphysical Necessity 25 maintain that the labor of one’s servant is in some way one’s own labor. The Humean simply has to affirm that the mind does immediately pass from the thought of one’s slave, whom one owns, to the thought of the fruits of his or her labor. As Hume says, “the ascribing of property to accession is nothing but an effect of the relations of ideas and of the smooth transition of the imagination.”46 Succession, finally, is the requirement that “men’s possessions shou’d pass to those, who are dearest to them.”47 Hume seems to think that the primary cause, as he puts it, for succession constituting property is “the general interest of mankind.” But, he adds, this interest is likely “seconded by the influence of relation, or the association of ideas.”48 When we look closely at Hume’s account of justice, then, what we find is his metaphysics: (i) property; (ii) the very idea of possession; (iii) the self that does the possessing; and (iv) society itself – all four components of the account either are, or rest upon, constant conjunctions of the kind that Hume believes himself to have shown comprise causation and substances alike. And there’s more. The stability of property is a fundamental law of nature, according to Hume.49 Having worked through Hume’s metaphysics, we know that by a fundamental law of nature Hume just means “consistent with present expectations.” But it’s easy enough to forget this (or to have bypassed the metaphysical discussion altogether, on the assumption that metaphysics has no bearing on the social and political issues) – and “fundamental law of nature” is hardly an insignificant designation. Certainly it is far more authoritative than “what we are accustomed to.” This is a point to be taken very seriously. Hume’s reduction of necessity to custom occurs at the level of metaphysics. But at that level, it’s clear that what he means to say is not that custom is actually necessary, under any normal understanding of the term. No. The entire force of the argument is in the opposite direction: what appears to be metaphysical necessity is in reality epistemic habit. By contrast, already in the discussion of free will we see Hume – intentionally or not – exploit the ambiguity of the re-definition. All action is necessitated, he says – when all he really means is regular. Here, at the level of social and political thought, the rhetorical strategy is full-blown. Via the language of natural law, Hume would have us imagine that with respect to property, what we do and think on the basis of custom is genuinely necessary. The sleight of hand is more interesting ideologically than it may appear at first glance. It is neither a simple naturalization nor a standard appeal to natural moral law. Rather, Hume has done Hobbes one better. Hobbes advanced a radical conservativism on the basis of universal self-interest: whatever the existing regime, it is legitimate because any form of order is better than disorder. Hume has produced a similarly generic conservativism, but without making an argument about human needs or goods at all. It’s not that custom itself is commendable, either in general or in any one particular situation; rather, it’s that whatever we have come to expect is, by definition, metaphysically necessary – a fundamental law of nature. At the same time, however, because it implicitly trades on the definition of necessity that Hume rejects, the move ultimately calls Hume out. Does he want to make claims about necessity, and mean it, or does he not? If not,

26

Ontology Revisited

then all he has actually told us about the institution of private property is that we are accustomed to it.

Metaphysics and the Social: Doing without Powers I want to complicate matters now, by suggesting that even Hume is not consistently able to articulate and defend his views on exclusively Humean grounds. One way to advance the thesis that social and political thought has a metaphysical infrastructure is to illustrate a theory’s internal consistency in this regard, as I have just done. The other way is to point to inconsistency, to show that, in point of fact, concepts and arguments at the level of social and political theory cannot be made to sit a-top any and every metaphysics. Of course, if Humeanism fails as a metaphysics, then there will be no successful account of moral, social or political experience that can be made to rest upon it. But I do not need to go that far to challenge the myth of metaphysical neutrality. I need only establish that there may be a mis-match between a thinker’s social and political theory and his or her metaphysical assumptions. I shall consider three issues, here. The first is the fact that realist talk of powers shows up in Hume’s moral, social and political philosophy just as it does elsewhere. Simply as a matter of vocabulary, Hume can’t stay away from words and phrases like “produce,” “exert,” “excite,” “give rise to,” “influence,” “arising from,” “render” – let alone “(to) cause” and “power.” As I said earlier, I don’t think that Hume helps himself to dispositional language merely because he knows that it carries meanings to the reader that regularity-talk does not. I think that it is because it is harder than one might imagine it to be to shed such language. Take a verb such as “to do,” conceived in active terms as something like “to effect” or “to enact.” The principled Humean has two options in the face of such a verb. Either she has to expunge it from her vocabulary, or she has to stipulate that what she means when she uses it is: “things are one way, then they are another way.”50 Let’s assume that she opts for the latter option, stipulating that when she says “to do,” she doesn’t mean “to effect” or “to enact”; she means “things are one way and then another.” Now, this is the sort of claim that rolls off the tongue easily enough, especially the tongues of philosophers of language trained in the Humean tradition. We can even imagine her suggesting that it might be useful to distinguish between “do” and “do*,” where “do*” is the philosophical use of the term. But there are real problems making good on this strategy. At the very least, nobody else thinks that “do” means that. If they did, no one would object to doing the dishes. Nor would the Humean herself be entitled to think of anyone as doing the intellectual work that is philosophy. I say this in a light-hearted way, but it is intended as short-hand for serious considerations, both about the intersubjective constraints to re-definition in general and about the adequacy of the proposed account of “to do.” Moreover, if the proper definition of ostensible action-terms such as “to do” really were “things are one way, then another,” it would seem to follow that a natural language that reliably traced the contours of the world would contain no such terms at all. Thus, to return to my original point, it may not be possible, in the end, to talk as a consistent Humean must. For to do so would seem to require that

Hume: Custom as Metaphysical Necessity 27 one rid one’s language of a whole range of verbs that – notwithstanding Lao Tzu’s dictum that “The way to do is to be” – are normally used to express doing, not being.51 Certainly Hume has not managed the semantics of it. And it’s not just that Hume can’t help but use action verbs in a standard way generally. He also makes arguments that specifically trade on appeals to real powers – for example, the key argument in the Treatise in which he affirms the sentimental basis of morality. Book III opens with the section entitled “Moral Distinctions not Deriv’d from Reason.” Hume starts off the section by telling us that he believes that “our reasoning concerning morals will corroborate whatever has been said concerning the understanding and the passions.”52 Nothing about the prior metaphysical discussion will have to be revised for the purposes of his analysis of what leads us to act. On the contrary, the metaphysics will only be confirmed. In Book II, he had claimed that all action is motivated by sentiment, that “reason is, and ought to be the slave of the passions, and can never pretend to any other office than to serve and obey them.”53 At the start of Book III he refers back to the earlier conclusion, asserting that (1) moral facts compel us to act; (2) reason can’t compel us to act; (3) therefore, morality can’t be based on reason. It seems a simple argument at first, but it bears closer scrutiny.54 Here are several key passages: Morals excite passions, and produce or prevent actions. Reason of itself is utterly impotent in this particular. The rules of morality, therefore, are not conclusions of our reason.55 An active principle can never be founded on an inactive; and if reason be inactive in itself, it must remain so in all its shapes and appearances, whether it exerts itself in natural or moral subjects, whether it considers the powers of external bodies, or the actions of rational beings.56 Reason is perfectly inert, and can never either prevent or produce any action or affection.57 Reason is wholly inactive, and can never be the source of so active a principle as conscience, or a sense of morals.58 Thus upon the whole, ’tis impossible, that the distinction betwixt moral good and evil, can be made by reason; since that distinction has an influence upon our actions, of which reason alone is incapable.59 The question that jumps out is this: how are we to understand the repeated characterization of reason as “impotent,” “inert,” “inactive”? If we have thus far taken Hume at his word, we will have two axiomatic principles that can function as points of reference. First, everything is impotent, inert and inactive. Second, to cause something is nothing more than to always be observed to come first, in a regular sequence. Putting these two principles together, we can infer that what

28

Ontology Revisited

Hume means when he tells us that reason is “utterly impotent in this particular” is that it is never observed to come first, in a constant conjunction with moral action.60 Rigorously formulated, then, the argument of Book III, Part 1, Section 1, is as follows: (1) moral facts compel us to act; (2) reason alone is not observed to be constantly conjoined with moral action; (3) therefore it cannot be that reason alone could be observed to be constantly conjoined with moral action. But now we need to know if the claim “reason is not observed to be constantly conjoined with moral action; therefore it cannot be that it could be so observed” actually captures what Hume is saying in the passages above. It doesn’t. What Hume is saying, over and over again, is not that reason can’t be constantly conjoined with moral action because it isn’t, but rather that reason can’t be constantly conjoined with moral action because reason is such that it is “incapable” in this regard. Specifically, it is inert and inactive, and “an active principle can never be founded on an inactive,” as quoted above. But one is not in a position to say this if one is a Humean. For one thing, reason is far from unique, for Hume, in being inactive. Nothing is active, on a Humean metaphysics. Insofar as Hume wants us to think that reason isn’t up to the job because it’s inactive rather than active, the same will be true of the passions. Second, recall that Hume also maintains, as a separate point, that the regularity of the regular sequences that we observe is in no way determined by the nature(s) of that which comes first. Anything can be founded upon anything. Thus it’s not only that he has no grounds for saying that reason can’t precede moral action because it (like everything else) is inert rather than dynamic; in fact, he has no grounds for pointing to any feature whatsoever, of any postulated antecedent, in an effort to explain why it could or could not figure into a given constant conjunction. All anyone can ever do is name which regular sequences we do or do not observe. Admittedly, it is possible that Hume’s argument really is intended to be only that reason is not observed to be constantly conjoined with moral action, and that therefore it can’t be. But that’s not what he says. And for good reason, since it’s a far weaker argument than the one that he does give. The second and third issues that I want to consider are Hume’s treatment of selves and societies, respectively. Hume has these pseudo-entities do a good bit of work, more than his anti-realism about them permits. With respect to selves, there are four points in relation to which this can be seen. The first, having to do with moral sentiment, is the most general. Recall what Hume says a purported self actually is: “I may venture to affirm of the rest of mankind, that they are nothing but a bundle or collection of different perceptions, which succeed each other with an inconceivable rapidity, and are in a perpetual flux and movement.”61 Given that this is his view, we have to ask whether or not a fluctuating bundle of perceptions is the kind of thing that can be the bearer of moral sentiment. In the case of an immediate, simple, non-moral pleasure, such a pseudo-entity might suffice. With that kind of pleasure, perhaps we can get away with saying that when it seems to oneself as though one is having such a sentiment, really what is happening is that there’s a pleasure sensation mixed in with the bundle of perceptions that just is oneself. But moral sentiment is different. In the non-moral case, what stimulates pleasure in one person may well be different from what does so in another. By contrast, our moral

Hume: Custom as Metaphysical Necessity 29 responses to given stimuli are universal, according to Hume. Once we add in this condition, it becomes far more intuitive to conceive of the having of moral sentiment in just the way that Hume actually talks about it: viz., that it is precisely our “self” that has the feeling, and that it is a fortunate if inexplicable fact of nature that we all feel moral pleasure and moral pain in response to the same things. However, in talking this way – as Hume himself does – we’ve extruded the moral sentiment from the “self” bundle. Having done so, we will now want to know which part of an ever-changing bundle of perceptions that each of us apparently is – which part of such a bundle is robust enough to sustain the purported universality of response? With the concept of character, Hume adds further implicit heft to the selfthat-isn’t. Hume begins by distinguishing between character and acts. Actions are fleeting, superficial. Character is stable, coherent, persistent. For this reason, Hume sees character as the underlying phenomenon in virtue of which individual acts have whatever normative salience they do. Here’s Hume: Actions are by their very nature temporary and perishing; and where they proceed not from some cause in the characters and disposition of the person, who perform’d them, they infix not themselves upon him, and can neither redound to his honor, if good, nor infamy, if evil.62 And again: If any action be either virtuous or vicious, ’tis only as a sign of some quality or character. It must depend upon durable principles of the mind, which extend over the whole conduct, and enter into the personal character . . . . We are never to consider any single action in our enquiries concerning the origin of morals; but only the quality or character from which the action proceeded. These alone are durable enough to affect our sentiments concerning the person.63 The question, then, as before, is whether or not what Hume wants character to do can be done by a bundle of impressions. On the face of it, it would seem that a viable Humean response would be to say that the durability of character doesn’t actually add anything to the picture ontologically. Character is merely the fact that people do, always, behave in the same ways that they have in the past, in given circumstances – i.e., that character is just another term for the “necessitation” of action, where necessitated just means habitual. The problem with this move, however, is that it doesn’t square with Hume’s claim that actions are signs, signs for the dispositions that cause them, as Hume puts it. Regularity of sequence may be able to be substituted for the property of durability, but regularity doesn’t give us the distinction between signifier and signified that is implied by the concept of a sign. To put it differently, on Hume’s account there is nothing for acts to be signs of. Regular sequences of behavior are all there is. What Hume wants character to do, or be, is more than a Humean metaphysics can sustain. A further point is that the claim on behalf of character is

30

Ontology Revisited

problematic for Hume in just same the way that the claim about the inert quality of reason was, i.e., it suggests that there is some property of character in virtue of which it regularly precedes (i.e., is the “cause” of) morally salient sentiments regarding others. But this is not an explanation-form that is properly available to Hume. The self does even more work in relation to pride (and its inverse, humility). Pride, as Hume conceives it, is completely contingent upon this pseudo-entity-forwhich-there-is-no-impression. To begin with, the cause of pride (as well as humility, and other sentiments that are mediated through the self) is what Hume calls a “double relation.”64 The dynamic of the double relation is one in which a single cause – a lovely home, say; Hume often uses this example – is related both to the idea of one’s self and to the impression of pleasure. Without the intermediary of the self (or, “self”), there can be pleasure, but not pride. The self is thus a “cause” of pride, in that the idea of it regularly precedes that of pride.65 Moreover, the self is also the object of pride. As Hume puts it, “(a)ny thing, that gives a pleasant sensation, and is related to self, excites the passion of pride, which is also agreeable, and has self for its object.”66 Pride, we can see, therefore hinges on the idea of a self that is real enough to play a mediating role, both coming and going. Moreover, one of the important sources of pride, as Hume would have it, is reputation. In the case of reputation, what figures in the double relation with pleasure is not an external entity (or, the idea of one), such as a house, but precisely the positive feeling about oneself that is held by another. In the case of pride in relation to one’s home, the double relation between the idea of one’s self and the idea of a house is, for Hume, simply a matter of constant conjunction. The association between one’s self and another’s feelings about oneself, by contrast, is further mediated by what Hume calls sympathy. And sympathy presupposes an even more ontologically robust self than pride alone does. In order to experience, via sympathy, what someone else is feeling, Hume says, our thoughts about the other person’s experience have to undergo a “conversion,” through which they change from ideas into impressions.67 The process is as follows. “’Tis evident,” Hume contends, that the idea, or rather impression of ourselves is always intimately present with us, and that our consciousness gives us so lively a conception of our own person, that ’tis not possible to imagine, that any thing can in this particular go beyond it. Whatever object, therefore, is related to ourselves must be conceived with a like vivacity of conception.68 To put it differently, we begin with our idea of the self, then relate other ideas back to that most vital one. Like charging a car battery with jumper cables, the connection with the idea of self so enlivens what was previously just an idea that it becomes an impression. Here’s Hume again: “The stronger the relation is betwixt ourselves and any object, the more easily does the imagination make the transition, and convey to the related idea the vivacity of conception, with which we always form the idea of our own person.”69 In the case of sympathy, then, we find in Hume’s account not just something that he goes so far as to call

Hume: Custom as Metaphysical Necessity 31 an “impression” of a self (and which certainly seems as though it ought to count as one, as opposed to being merely an idea of a pseudo-self) – in addition, we’ve got a causal connection that (a) is described in terms of a transfer of kinetic energy, and (b) would not work in the argument in which it figures, were it to be formulated in the language of constant conjunction instead. When we turn to what Hume says about society, we find his discussion marked by the same difficulties. Specifically, here too there are questions about (a) whether or not Hume’s stated conception of what societies amount to metaphysically is robust enough to sustain his substantive analysis, and (b) whether or not his claims about what they can do may be reconciled with his account of causation as constant conjunction. As noted earlier, Hume is not an atomist. There are, on his view, things that are attributable to society rather than to individuals. Especially, it is society that elicits from individuals the specifically moral response that constitutes justice as a virtue. It may seem that I’ve got things backwards, in saying this. After all, in his discussion of why we create property laws in the first place, Hume claims not that justice is contingent upon society, but rather precisely the reverse: that society depends for its existence upon justice. But when we look closely at the case that he makes for why justice is not simply functional but is also a normative good, we find that it is a relationship between persons, and not any one or number of individual persons that prompts our moral sentiment. Justice, Hume says, is anomalous as a virtue, in that it often does not elicit moral pleasure in any one instance. It may be just, for example (according to Hume), to give money to a wealthy miser, rather than to give it to a needy child. Hume defends such scenarios by saying that although justice doesn’t necessarily please in any one given case, it pleases as a system. The move from particular to general signals the familiar distinction between act and rule utilitarianism. But what is significant here, and perhaps less often observed, is the ontology that sustains the move. Justice pleases as a system, Hume claims, via the sentiment of sympathy. But with whom would one sympathize, in order to “catch” the impression of moral pleasure in relation to justice? It can’t be any one single individual, since justice is no less likely to please in the particular for someone else than it is for oneself. No, the object of our sympathy (i.e., the initial subject of the impression of moral pleasure, from whom – via the vitality of our idea of the “self” – the sentiment is sympathetically transferred to each of us as individuals) – the entity that generates our sympathy is the collective, sociological subject. Individual subjects feel a positive moral sentiment in relation to the artificial virtue of justice, but they do so via sympathy with an intersubjective “other,” i.e. with society. It is society, therefore, that has the power to elicit moral sentiment in relation to justice. However, such a claim brings with it the same problems as those I just raised in relation to the doings of the self. Namely, it is not clear that eliciting moral sentiment is something that can be properly thought to be the achievement of a bundle of regularities, organized now into a sociological rather than a psychological pseudo-entity. Moreover, the eliciting in question is mediated by sympathy, and sympathy, as we have just seen, is a phenomenon that itself does not translate well into the language of constant conjunction.

32

Ontology Revisited

Let me close with a few general thoughts about Hume’s social and political thought. First, despite the fact that Hume illicitly helps himself to real powers, and to real selves and societies that bear them, on the whole his work is a useful illustration of Humeanism as applied to human agency. The contradictions that I’ve identified emerge, it seems to me, not so much because Hume genuinely wants to say things about human beings that a Humean cannot say, but rather because, as I’ve suggested, it simply may not be possible to talk as a consistent Humean must, and say much of anything at all. Thus Hume, I believe, notwithstanding the problems I’ve noted, can be taken to be a paradigmatic Humean (or mechanist, as I am using the term) not just at the level of metaphysics, but at the level of social and political philosophy. Second, I want to stress that when we do follow through on the logic of Hume’s metaphysics, what we find is that agency, like all other events, must be conceived as “just one little thing and then another,” to borrow Stephen Mumford’s reference to David Lewis.70 To appreciate the force of this point, picture an animation flip-book. The way that a flip-book works is that there is a drawing on every page, each slightly different from the one before. If one bends all of the pages back, and releases the pages one at a time, but quickly, it appears as though the figures in the drawings are in motion. But they aren’t really. Really, there is just one little static image after another. It is exactly the same with agency, for the Humean. Absent the existence of powers, there can be only flip-book-like sequences of static states. This really is what the rejection of powers implies. One can talk about the illusion, or appearance of agency all one likes; but if one wants real animation, and not simulated animation, one cannot have it on a Humean metaphysics. Now, perhaps there is no such thing as real animation, real agency. Perhaps we are, as an ontological matter, literally akin to animated drawings. It’s possible. But we should at least be clear that this is the view that the Humean must defend.

3

J. S. Mill Humeanism and the Perfection of Distinctively Human Capacities

Hume, I have argued, is a relatively consistent Humean at the level of his social and political theory. There are things that he tries to say that he can’t, but the inconsistency mainly belies the limits of Humeanism itself, rather than the presence of any self-consciously held non-Humean views. As an ontological matter, Hume really does regard social and political life to be comprised of contingent patterns of regularity – to be constituted ultimately by custom, to use his term, just as is the order that we experience in nature. With John Stuart Mill the situation is different. In Mill’s case, we find not just a normative injunction to cast off the yoke of custom, as Mill puts it, but a social and political theory that sits quite uneasily, at very best, with the Humean metaphysics that Mill also affirms. As in Chapter 2, with respect to Hume, my aim in advancing such a claim is neither to weigh in on local questions of Mill scholarship nor to bolster the contemporary critique of Humeanism – though I will consider it to be all for the good if either or both of these effects are achieved indirectly. Rather, my objective is to drive home the point that theories at higher orders of complexity presuppose fundamental ontologies with which they are consistent; and, conversely, that adopting a Humean metaphysics issue in constraints, Humean constraints, upon what one may say at the levels of individual and collective agency. As in Chapter 2, I begin with Mill’s fundamental metaphysics, followed by his stated conceptions of the self and society. I then look to see if these ontological commitments permit him to make good on his normative ethical and political claims.

Mill on Causation, Kinds and Material Entities Mill understands himself to be offering a stronger, more robust account of what it is for one thing to cause another than can be found in Hume. For Hume, as we’ve just seen, causality is the fact of given impressions being constantly conjoined, plus – crucially – the subjective feeling of anticipation related to the experience thereof. To be a cause, from this perspective, is simply to be that which always comes first in a regular perceptual sequence – that which, in virtue of its position of antecedent, is itself conjoined with the expectation of the normally associated consequent. For Mill, by contrast, causality is not a matter of subjective expectation at all. Rather, it is the very fact that nature is structured by invariant

34

Ontology Revisited

regularity, or “uniformities,” to use Mill’s preferred term (combined with what Mill calls “permanent causes” or “original natural agents”). Mill takes such uniformities to be laws of nature, and argues that we may generalize from our experience of them that “[I]t is a law, that every event depends on some law,” i.e., “it is a law, that there is a law for everything.”1 This meta-law, as we might think of it, Mill calls the Law of Causation. The Law of Causation expresses the idea that nature is ordered. “To certain facts,” Mill says, “certain facts always do, and, as we believe, will continue to succeed.”2 A cause, then, for Mill, is not sufficiently defined as that which we always experience first, in a given regular sequence. Rather, a cause is that which, given the objective arrangement of things, is always first. As Mill puts it, it is necessary to our using the word cause, that we should believe not only that the antecedent always has been followed by the consequent, but that, as long as the present constitution of things endures, it always will be so. . . . Invariable sequence, therefore, is not synonymous with causation, unless the sequence, besides being invariable, is unconditional.3 Thus – as I claimed Hume wants to say but knows he can’t – for Mill it is because nature is ordered in the way that it is, that we experience it, and will continue to experience it, as we do. The “uniformity of the course of nature,” Mill says, is “the ultimate major premise” of all inductive reasoning.4 It follows from the Law of Causation that every cause (except for the primary, “uncaused” causes) is itself also a consequent, relative to a different unconditionally invariant antecedent. Therefore, Mill concludes, although we tend to think of a cause as being that which most recently preceded an effect, “[t]he real Cause, is the whole of these antecedents; and we have, philosophically speaking, no right to give the name of cause to one of them, exclusively of the others.”5 The cause of any given phenomenon, we might say, is thus causality itself: viz., the fact that all of nature is ordered as it is, rather than in some other way. Mill is not alone in treating the empirical world as a single closed system, comprised of sequences of events that extend, deterministically, both backwards and forwards in time. However, he is notable for having been attentive early on to the difficulty, on this view, of distinguishing between causes and background conditions. If Mill understands himself to be more of a realist than Hume with respect to the uniformities of nature, his realism does not extend to causal powers. When he first introduces the topic in the Logic, he claims that the disputed existence of powers is irrelevant to his considerations therein, and that his position is equally consistent with the affirmation or the rejection of such properties. In a move well-exploited by contemporary analytics, he writes: I make no research into the ultimate or ontological cause of anything . . . to find the true cause, the cause which is not only followed by, but actually produces, the effect. No such necessity exists for the purpose of the present inquiry . . . (Footnote: . . . “We are in no way concerned with the question.”)6

J. S. Mill: Humeanism and Human Capacities 35 It quickly becomes apparent, however, that while he may be in no way concerned with the question, he nonetheless has a considered opinion regarding its answer. In this regard, he calls upon those who think of qualities as powers (viz., powers to produce sensory effects) to show that their opinion is . . . anything in fact but a lingering remnant of the old doctrine of occult causes; the very absurdity which Moliere so happily ridiculed when he made one of his pedantic physicians account for the fact that opium produces sleep by the maxim, Because it has a soporific virtue.7 He also makes reference to “the established doctrine of the inactivity of matter,”8 which doctrine presumably is at odds with belief that matter bears dispositional properties. Finally, he says of what he terms “capacities given to objects of being causes of other effects” that “[t]his capacity is not a real thing existing in the objects; it is but a name for our conviction that they will act in a particular manner when new circumstances arise. . . . [i]t is but the contingent future fact brought back under another name.”9 The exception to this rule is when the purported capacity is nothing other than a present physical state, a “collocation of particles,” as Mill puts it.10 In such cases there is a fact of the matter in the present, but what it is, is a pattern of bits – not a power. While Mill views realism about powers as belief in the occult, he does affirm the existence of objectively given kinds. I want to proceed very carefully here, since this aspect of Mill’s thinking will figure considerably in the analysis to come. The issue is addressed explicitly in Books 1 and 4 of the Logic. Mill gets to what he calls in Book 4 “natural groups” through a series of distinctions related to the nature of classification. In one sense, he says, classification is simply a by-product of the act of naming: Every name which connotes an attribute, divides, by that very fact, all things whatever into two classes, those which have the attribute and those which have it not; those of which the name can be predicated, and those of which it cannot.11 Classification in this loose sense pertains to real and imaginary things alike, and is merely an “incidental effect” of having assigned an object a name.12 By contrast, there is a more rigorous, restrictive sense of classification, called Natural Classification, the aim of which is to capture, through names, the important features of things. Classification of this type applies only to things that actually exist. Natural groups (or kinds) are those into which things fall on the basis of what they are like, rather than as a function of semantics. Mill is no Aristotle. What things are like is such that “there is no property of objects which may not be taken, if we please, as the foundation for a classification or mental grouping of those objects.”13 Still, while there are no real essences, there are properties that are uniformly associated with other properties; it is the task of scientific inquiry to identify such regularities, and Natural Classification is the

36

Ontology Revisited

result. As an empirical matter, there are “groups respecting which a greater number of general propositions can be made, and those propositions more important, than could be made respecting any other groups into which the same things could be distributed.”14 It follows from this that: “[t]he properties . . . according to which objects are classified, should, if possible, be those which are causes of many other properties: or at any rate are sure marks of them”15 — keeping in mind that what Mill means by “causes” is not that the property produces the consequent, but only that it always precedes the consequent. Finally, Mill stipulates that “the reader is by this time familiar with the general truth . . . that there are in nature distinctions . . . not consisting in a given number of definite properties . . . but running through the whole nature, through the attributes generally, of the things so distinguished.”16 Mill refers to these systemic distinctions as distinctions of Kind, and regards them as being of the most scientific value. Given the objectivity of kinds as Mill presents them, one might wonder why he says that there is no property that may not be taken, if we please, as the basis for classification. Are kinds real or not? They certainly seem to be, and if they are, then why the equivocation? I regard the whole of Mill’s thinking as an unsuccessful, if Herculean, attempt to integrate a range of classical insights into a Humean metaphysics. If I am correct, then the equivocation with respect to kinds is par for the course. This said, I think that we can interpret Mill as wanting to say not that kinds are subjective heuristics, but only that they are not based upon essences. The motivation for this is that he does not believe in such things. Nature is ordered, and ordered in such a way that the uniformities that characterize it are clustered into discrete regions of regularity, but regions of regularity are not essences. There are only properties that are always associated with other properties, and objects that do or do not meet the definitions that we give to the terms that we use to pick out such invariant associations. Mill believes himself to be defending real causal relations, ones with genuine modal force. It’s a question, though, whether or not he actually is — or if he even can. The issue is this: if Hume thought that substances are impressions, Mill thinks that they are not even that; instead, they are what he calls permanent possibilities of sensation. In An Examination of Sir William Hamilton’s Philosophy – which, unlike the Logic, is an overtly metaphysical study – Mill announces his own phenomenalism in no uncertain terms: Matter, then, may be defined, a Permanent Possibility of Sensation. If I am asked whether I believe in matter, I ask whether the questioner accepts this definition of it. If he does, I believe in matter: and so do all Berkeleyans. In any other sense than this, I do not.17 The ostensibly realist claim that nature is one way and not another, therefore, turns out to tell us nothing about an external state of affairs that might function as an anchor for our causal beliefs. As with Hume, it amounts only to a claim that certain mental phenomena are marked by regularity (here it is those sensations that we may potentially have, rather than those that we do have).

J. S. Mill: Humeanism and Human Capacities 37 One might be tempted to object that what Mill really means is that the permanent possibilities of sensation are as they are because the material world is arranged as it is. But such a thought is at odds, and quite explicitly so, with what he says. What he says is that matter just is the permanent possibility of sensation.18 The upshot of this is that if, as Mill would have it, causality is the fact that nature is ordered, then causality can be only a dogmatic assertion to the effect that those sensations that we have not yet had, and which we may not ever have, would, if we were to have them, be organized into particular invariant sequences. Why? They just would. Thus the main difference between Hume and Mill, it begins to seem, is simply that Hume was more cautious, preferring to claim only that those sensations (or impressions) that we do have are so configured. Especially, it is important to see that Mill has not, after all, introduced an additional source of necessity into the account; there is only a paradigmatically Humean assertion of order. Already, therefore, even just at the level metaphysics, Mill’s thought contains contradictory elements. In addition to the tension between his phenomenalism and the intended modal status of the Law of Causation, it’s not clear that permanent possibilities of sensation are substantial enough, as it were, to sustain his account of Natural Classification. In order to make sense of the view, one would have to say that specific permanent possibilities of sensation are themselves already linked deterministically to other equally specific permanent possibilities. I don’t want to belabor the point, as nothing in my own argument hangs on whether or not Mill’s phenomenalism can be reconciled with his commitment to kinds, as he defines them. But the approach that he defends strikes me as inelegant, amounting to preloading the much-lauded uniformities of the material world onto the permanent possibilities of sensation. For even if we allow that permanently possible sensations can fall into natural kinds, there is a fundamental sense in which Mill’s attachment to Berkeleyan idealism simply runs counter to his realist proclivities. It’s hard to see why it should be so important to him to insist that there are permanent possibilities of sensation, but nothing that such sensations are sensations of. A related point is his scorn for powers. In defending his phenomenalism against criticism from a Dr. M’Cosh, who, we are given to understand, argued that it is the sun as a real material substance that causes wax to melt, and fire that causes lead to melt, Mill writes: Dr. M’Cosh . . . must have seen that after mentioning the attribute of exiting sensations, it could not be necessary to add that of making something else excite sensations. If Body altogether is only conceived as a power of exciting sensations, the action of one body upon another is simply the modification of one such power, of the sensations excited by another; or, to use a different expression, the joint action of two powers of exciting sensations. It is easy . . . to understand how one group of Possibilities of Sensation can be conceived as destroying or modifying such another group.19 Now, as we have seen with Hume (and with contemporary Humeans), it is always possible to insist, when pressed, that although one has used the language of

38

Ontology Revisited

powers, and has done so in order to convey meanings such as “active,” “energetic,” “forceful,” “capable of doing” – it is always possible to claim that in truth one does not mean to say anything of the kind, but only to assert the fact of a sequence. But there are limits to the generosity that can reasonably be demanded of an interlocutor. Patently, in the passage above Mill is undertaking to rescue his anti-realism about substances by invoking real powers, i.e., precisely those entities which he elsewhere dismisses as occult phantasms.

Social Entities: Individual and Collective Mill defends a conception of the self that is very close to Hume’s, except for the fact that for Hume the basic ontological unit was the impression (it is in the nature of phenomenalism that this is also the primary perceptual-epistemic unit), while for Mill it is the permanently-possible-but-not-necessarily-actual sensation. The self, or Mind, Mill says, “is but a series of feelings, or, as it has been called, a thread of consciousness, however supplemented by believed Possibilities of consciousness which are not, though they might be, realized.”20 And again: “Mind itself [is] merely a possibility of feelings.”21 Mill defends this view from various lines of criticism, e.g., that it implies that other selves don’t exist, or that it implies that God doesn’t exist; he concedes, however, that there is an unsolvable problem with the account, flagged by the fact that it leaves one with no way to explain how it is that one’s memories are experienced as one’s own. Alan Ryan summarizes the situation in his classic book on Mill: [i]n other words, the fatal flaw of the phenomenalist account of personal identity is that in order to construct the series of thoughts which constitute a given person, the principle of selection which we must employ to construct the correct series already involves reference to the person whose thoughts they are. To construct my mind, I must employ only my thoughts, my memories, my expectations.22 Mill, unlike Ryan, does not regard the flaw to be fatal. But it does force him to stipulate that in addition to streams of sensation there exists, with respect to conscious human beings, an inexplicable “original element,” which he tells us he will call the Ego or Self. “As such,” he says, “I ascribe a reality to the Ego – to my own Mind – different from that real existence as a Permanent Possibility, which is the only reality I acknowledge in Matter.”23 One might presume that with the positing of such an element – even if nothing can be said about it other than that it exists – Mill would have understood himself to have given away the argument. But Mill does not seem think that he has. For my own purposes, what matters is only that the self that Mill purports to defend is as thin as it is, and that notwithstanding his recourse to the concept of an Ego, the view that he sees himself as holding with respect to the ontology of the self is generally in keeping with Hume’s. On issue of whether or not the self has free will, Mill’s position is both importantly the same as Hume’s and quite different from Hume’s. Hume, recall,

J. S. Mill: Humeanism and Human Capacities 39 defended the Doctrine of Necessity, but said that it is compatible with freedom because in maintaining that actions are subject to causal necessity, all that one is asserting is that people behave according to habit: if you know what someone is like – and perhaps what people are like generally – you can predict how they will act in given circumstances. The alternative is to imagine that behavior is simply a matter of chance. But, he argued, we can see empirically that this is not so, that people’s behavior is not random. If they were, then punishment would be nonsensical, for people could not be reasonably held accountable for uncaused, arbitrary acts. Recall, too, that the so-called determinism, or necessitation, with which Hume reconciles free will, if he does, lacks any modal purchase. X’s being necessitated means only that, because it is customary, one expects that it will happen. On the face of it, Mill’s position looks to be Hume’s. His initial argumentative strategy replicates Hume’s exactly. “Correctly conceived,” Mill writes, the doctrine called Philosophical Necessity is simply this: that, given the motives which are present to an individual’s mind, and given likewise the character and disposition of the individual, the manner in which he will act might be unerringly inferred: that if we knew the person thoroughly, and knew all the inducements which are acting upon him, we could foretell his conduct with as much certainty as we can predict any physical event.24 But the similarity is relatively superficial. To begin with, in terms of the task at hand, if Mill wants to reconcile free will with determinism, then he, unlike Hume, will have to reconcile free will with real determinism, which he believes himself – correctly or not – to have defended, i.e., with the fact, if it is one, that all events occur within a set of relationships that are not merely customary but unconditionally invariant. The bar with respect to making out a viable case for compatibilism is therefore considerably higher for Mill than it is for Hume. And Mill takes a different approach to the task, too, by relying heavily on the notion of an invariant tendency to be constantly conjoined, rather than on the idea a constant conjunction itself. Mill’s argument, clearest in the Examination, goes as follows: just as the causal connection between the imbibing of poison and death can be disrupted by the presence of an intervening connection between the imbibing of an antidote and not dying, such that the invariant tendency of poison to be followed by death will not be realized, so too the invariant connection between a person’s character and a given action is a tendency that need not necessarily be actualized. Why? One reason is that action, like other events, is multiply determined, and character is not the only causally relevant variable. So while my character alone may be invariably associated with behavior r, if the total causal picture with respect to my action includes an also-relevant casual factor p, such as a threat of punishment, where p is invariably associated with behavior ~r, then the link between my character and behavior r may remain unrealized, and ~r will happen instead. Even more important, character itself is caused. Therefore, Mill contends, it can be changed just as surely as can the action of a poison. When we are young,

40

Ontology Revisited

our character is shaped by others; parents and elders expose us to relevant influences, with the aim of shaping our personalities in different ways. Character can be re-shaped by us as adults, Mill insists. If we wish to change, we need only subject ourselves to causal regularities that will affect us in the intended manner. Punishment is again relevant because punishment, and the threat thereof, makes us revise our choices. Changes in choices, in turn, may lead to changes in what Mill calls “purposes”: habits of willing that, ideally, become detached from the desires or aversions that may have originally motivated them. And then comes the key point: while others can potentially re-shape our character, so too can we. Indeed, it is our ability to do so that constitutes moral freedom, Mill says. Moral freedom is therefore not only consistent with the causal determination of action; it presupposes it. In contrast to Hume’s version of the claim, however, for Mill this is so because freedom is a function of the real necessity of countervailing regularities – of motivational “finks” and “antidotes,” I’ll say, for the analytic metaphysicians in the crowd – not because action were arbitrary rather than habitual not be creditable or blameworthy. Whether Mill’s case for compatibilism is persuasive or not is a different question, to which I shall return in the next section. Finally, Mill differs markedly from Hume at the level of the social. Hume granted the same ontological status to sociological entities as to selves, treating both as collections of regularly associated impressions. Mill, by contrast, is a classical atomist – by which I mean to say that he believes societies to be only numerical aggregates of individuals. His explicit claim to this effect comes in Book 6 of the Logic, wherein he declares that [m]en, however, in a state of society, are still men . . . Men are not, when brought together, converted into another kind of substance, with different properties; as hydrogen and oxygen are different from water, or as hydrogen, oxygen, carbon, and azote, are different from nerves, muscles, and tendons. Human beings in society have no properties but those which are derived from, and may be resolved into, the laws of the nature of individual man.25 But the atomism can also be inferred, for example from the passage in Utilitarianism in which the subject for whom “the general happiness” is good is “the aggregate of all persons.”26 Those who do not reduce sociological entities to aggregates of individuals are holists. To be a holist is to think that wholes, here sociological ones, are entities in their own right. In the history of political thought, Aristotle – and Rousseau following him – are holists who explicitly distinguish between associations and aggregates, or “mere alliances,” as Aristotle puts it in the Politics.27 Aristotle, of course, takes the existence of the association that is the polis for granted; as a modern, Rousseau has to tell a story about how the collectivity that is the Sovereign, or the People, comes to be. The distinction between atomists and holists may also be expressed in terms of the concept of emergence. An emergent phenomenon (property or entity) is one that is not equivalent, ontologically, to the

J. S. Mill: Humeanism and Human Capacities 41 plurality of its parts. Holists believe that the properties borne by sociological wholes are emergent properties of the whole, irreducible to those borne either (a) a-socially by any of the individuals upon whose existence the whole depends, or (b) by the sum total of individuals considered as a numerical plurality. Atomists deny the existence of such properties, and the ostensible wholes that are purported to bear them, maintaining instead that the properties of such entities reduce to those borne by their parts, i.e., by individual persons, and that it is only individuals that may be said to exist. Holists often point to the distinctive powers of sociological wholes in order to establish their ontological standing as emergent entities. Given their disavowal of dispositional properties, this is a move that is unlikely to be made by a Humean, which may go some way toward explaining why atomism seems to be the default Humean ontology. I shall return to the issue of emergence in Chapter 5. Atomism, it should be noted, is strictly speaking an ontological position, not an epistemic, methodological or semantic one. As do non-reductive physicalists in the philosophy of mind, Mill argues that there is a place for talk of macro-level regularities (i.e., for social science versus what he calls the sciences of individual man [psychology and ethology28]), even though it is individuals alone who actually exist. It is important not to be distracted by the allowance when it comes to assessing Mill’s ontology, however. The claim is only that macro-level regularities may be tracked directly, without first being reduced to the individual-level regularities to which they do, in Mill’s view, un-problematically reduce.29 And even this is provisional: in the end, says Mill, the laws identified by social scientists must be shown to derive from the fundamental causal laws of psychology. (By contrast, the laws of psychology do not reduce to the causal laws of physiology.30 Mill is an atomist, but not a physicalist.) Stepping back, then, we can summarize Mill’s metaphysical commitments as follows:

• • • • • • •

what exists are permanent possibilities of sensations, plus the inexplicable given of personal self-awareness, or Ego; the permanent possibilities of sensation are deterministically ordered, forming a fixed-but-not-necessarily-actual total causal structure; causality is the fact thereof; powers are unreal, occult-like pseudo-phenomena; permanent possibilities of sensation fall into natural kinds, which, because essences do not exist, are demarcated by concentrations of regularity; selves, such as they are, operate entirely within the closed, total causal structure that is reality; societies are atomistic aggregates of individuals.

Having established that this is Mill’s position at the level of metaphysics and basic social ontology, I now want to raise the question of how his attachment to a revised but recognizable Humeanism affects what he can say at the level of normative theory. I shall argue that there is a very real sense in which Mill, for

42

Ontology Revisited

reasons of metaphysics, can neither say well nor defend well what he actually believes.

The Clash of Categories in Mill’s Thought The analysis that follows is developed against the backdrop of two established lines of Mill scholarship. The first is exemplified by the work of Alan Ryan. Ryan argues that Mill’s thought is unified by his conception of rationality – that Mill’s views on morality and politics must be seen as being related to his exhaustive discussions of epistemology, method and metaphysics in volumes such as A System of Logic and An Examination of Sir William Hamilton’s Philosophy. Ryan also shows that Mill’s position is weak in places. Ryan is critical, for instance, both of Mill’s philosophy of mind and of his effort to render determinism consistent with free will. The second has to do with Mill’s debt to the Ancients. A good example of work of this type is that of Nadia Urbinati, who argues that Mill must be understood as trying to conceive of a modern Athenian polis. The case that I want to make connects these two broad lines of discussion, the one having to do with Mill’s underlying philosophical apparatus, the other having to do with the persistence of classical ideas in Mill’s thought. In my view, it is not simply that Mill defends various untenable empiricist positions, which he does, nor simply that classical considerations figure significantly in his work, which they do. Rather that what we see in Mill is an unsuccessful if valiant attempt to fit a neo-Aristotelian moral and political agenda into a Humean metaphysical framework. Mill, I think it’s safe to say, believes the following:

• • • • • •

that human beings have an as-yet largely unrealized capacity for a specifically human excellence, or virtue; that the activities that constitute human excellence are sources of the higher pleasures; that the achievement of human excellence hangs on the cultivation of certain character traits; that the ideal character-type is one that involves both a love of inquiry and a well-developed social sentiment; that these traits are socio-historical achievements; and that it therefore makes sense to ask which institutional forms, and what kinds of social relations more broadly, are conducive to bringing such traits into being.

Of course, if one reads Mill with an eye only to the Principle of Utility, as he calls it, then it is possible to miss the fact that Mill’s utilitarianism and liberalism fit into this larger normative scheme. But this is a problem that can be traced back to the reader, and it’s easily rectified. We would run into the same trouble if we were to pay attention only to Utilitarianism and On Liberty, but not to Representative Government (or Principles of Political Economy, for that matter). Or if we were to forget the discussion of ethology in the Logic (or, as G. W. Smith noted, of free

J. S. Mill: Humeanism and Human Capacities 43 will in the Examination), when we turn to the other works.31 Again, in all of these cases we need only attend to the various elements of Mill’s thinking, and then integrate them, in order to gain a proper understanding of the whole. The problem that I am interested in is not so easily resolved. It’s this: the whole of what Mill wants to say is not something that can be said by a Humean. The metaphysics that Mill needs comes from Aristotle, not from Hume – as I believe both Hegel and Marx understood. In order to talk persuasively and coherently about the social and political conditions of possibility for an historically as-yet unachieved aim of human flourishing, and about the complicated relationships between flourishing and happiness and pleasure, one needs, at a bare minimum, a dynamic, open-ended ontology rather than a static, closed one. Moreover, the ontology will need to include a concept of form or essence, which can sustain an end such as eudaimonia. And we might add that in order to talk naturally about how one would grasp such phenomena in thought, one may need an epistemology that involves not just observation and inference, but phronesis – and perhaps nous as well. Mill has recourse only to the categories of an empiricist Humeanism. There is a wonderful phrase of Adorno’s, which I shall address in Chapter 4, the statement that objects don’t go into their concepts without leaving a remainder. Mill’s object, I want to suggest, can barely be made to fit at all. That this is so can be seen in relation to four key issues in Mill’s work: happiness, utility, character and the formation of a social sentiment. Let me begin with happiness. Bentham and James Mill had said that happiness is the feeling of pleasure, and that the maximization of it is morally good. Unlike both Plato and Aristotle, but in keeping with Hobbes, they held that the source of the pleasure that one pursues is of no concern; the injunction is to maximize pleasure tout court. Mill famously objects, arguing that some things – especially, intellectual activity and acting out of concern for others – yield what he calls higher pleasure, while other things – the satisfying of basic physical desires, say – yield lower pleasure. The distinction is both qualitative and quantitative: higher pleasures are more befitting than lower pleasures, and they are also more pleasurable than lower pleasures. But Mill’s argument about what makes the higher pleasures be higher, quantitatively or qualitatively, is notoriously unpersuasive. All Mill can say is that anyone who has sampled both will tell you that the higher pleasures are better. He lets out that there is a “sense of dignity” associated with the exercise of the higher faculties, and that it is “essential” to “the happiness of those in whom it is strong,” such that “nothing which conflicts with it could be, otherwise than momentarily, an object of desire to them.”32 But the argument hinges on the subjective experience: “On a question of which is the best worth having of two pleasures . . . the judgment of those who are qualified by knowledge of both, or, if they differ, that of the majority among them, must be admitted as final.”33 Mill’s defense of the idea that the higher pleasures are more pleasurable than the lower pleasures is immediately recognizable as the second argument of three that Plato has the character of Socrates put forward to Glaucon and Adeimantus in Book 9 of the Republic. In the Republic, however, we are given to understand that it’s inadequate. If the higher pleasures really are more pleasurable, then we

44

Ontology Revisited

are going to need to be able to say why. (And what if it turns out that they’re not, or at least that someone reports that in their own experience they’re not, but we still want to say that they’re better?) Accordingly, Plato doesn’t stop there. He claims further that the more worthy pursuits are extra-pleasurable because they’re good (not, as Mill would have it, good because they’re extra-pleasurable). And Plato has something to say about what being good means, beyond “likely to be pleasurable.” Aristotle deals with the issue differently, by distinguishing between the pursuit of eudaimonia, or flourishing, on the one hand, and the pursuit of pleasure on the other. Flourishing (also translated as happiness) is not unrelated to pleasure, but they are different ends. Pleasure is a sensation. Happiness, or flourishing, by contrast, is the activity of doing well, over the course of a lifetime, the kinds of things that human beings are characteristically suited to do. Flourishing is pleasurable, to be sure, but it does not consist in the pursuit of pleasure. Now Mill, very much like Aristotle, clearly sees a connection between the higher pleasures and the utilization of what he sees as distinctively human faculties – including, above all, our capacities for intellectual and moral excellence, or virtue. After all, the higher pleasures derive from just those activities that involve the faculties in question. But Mill doesn’t have the right metaphysics to make the case. In the end, he has to make do with the idea of pleasure rather than that of eudaimonia not because he actually thinks that pleasure is the Good, but because he doesn’t have an underlying concept of form, or real essence, that could function to ground the notion of eudaimonia. A concept of form in relation to human beings can explain why the higher pleasures are higher. For Plato, the form in question is that of goodness. For Aristotle, it is simply the human form. If, like Aristotle, Mill could talk about the human form, then he could say, as Aristotle does, that happiness, or flourishing, attaches to the full and excellent actualization thereof. Mill could even stipulate, as Plato does, that eudaimonia carries with it not just the best type of pleasure, but the most pleasure. But Mill’s kinds are constituted by regularities. The result is that he can talk at length in chapter 3 of On Liberty about “the qualities which are the distinctive endowment of a human being,”34 and “the danger which threatens human nature,”35 i.e., that people’s characteristically “human capacities are withered and starved,”36 but his antiessentialism precludes him from doing justice to such talk. The effort to re-package eudaimonia as pleasure is heroic in its way, but can’t work. Regularities lack the requisite normative heft, just as they lack sufficient modal traction. There is a similar problem with Mill’s related attempt to redefine the notion of utility. In chapter 1 of On Liberty, Mill writes: “I regard utility as the ultimate appeal on all ethical questions; but it must be utility in the largest sense, grounded on the permanent interests of man as a progressive being.”37 As with the effort to introduce qualitatively different kinds of pleasure into the utilitarian schema, Mill’s impulse here is to try to turn the principle of utility into a counterfactual norm. With respect to pleasure, the distinction was between what people do desire and what they would desire, if they knew better (and therefore should desire); with respect to utility, “interest” is analogously equated not with whatever people may

J. S. Mill: Humeanism and Human Capacities 45 happen at present to regard as advantageous, but rather with what they should regard as advantageous, viz., the “improvement” of humankind. The suggestion is that the improvement of humankind is objectively advantageous, not (only) that qualified desirers would desire it. Consequently, there is even more normative work to be done here by the concept of human nature than in the higher pleasures case. And to accomplish it, Mill now needs not just a concept of form, which would give him a way to differentiate real interests from perceived interests, but the concept of a form that may or may not be yet actualized, historically. Once again, however, Mill doesn’t have the right metaphysics for the job. In order to be able to talk about the permanent interests of man as a progressive being, he needs a truth-maker for “that which does not yet exist, but could – and, more to the point, should.” But unlike Hegel or Marx, both of whom avail themselves of historicized neo-Aristotelian essences in order to say something very much like what Mill means, Mill can talk only about permanently possible sensations that we do not happen to be experiencing at present. Now, in saying this I do not mean to suggest that Mill can’t talk at all about un-actualized potential. There is a sense in which he is doing just that with the very notion of a permanent possibility of sensation. And, as we have seen, he also believes that regularities can be offset by other regularities, such that it is permissible to talk about what might have happened but didn’t. But neither of these points shows that he in a position to talk about such a thing as a developmental kind. Especially, a claim to the effect that we do not experience all possible sensations at all times falls far short of the idea that human beings are such that we have the potential to become other than we presently are. There is some overlap between this issue and the next, which has to do with Mill’s focus on the shaping of character. His interest in character formation is not just a matter of moral philosophy; it’s also a social and political concern. In Considerations on Representative Government we are told that the fomenting of good character is one of the two proper purposes of government (along with efficient administration), and it should not be lost upon anyone that Representative Government is an instance of what in the Logic Mill calls the science of ethology – the aim of which is to identify the causal conditions linked to different personality traits. The specific character type that Mill is interested in nurturing is defined by those traits that he commends across a range of texts, viz., vitality, originality, autonomy, intellectual fearlessness, public mindedness and identification with the well-being of others. There is also an emphasis – for instance, in his discussion of free will in the Examination – upon self-mastery, or self-control. Here again, though, the trouble is that Mill doesn’t have the metaphysical infrastructure to make good on what he clearly believes. There are at least four important impediments. To begin with, he is fervent about the cultivation of good character, but he can barely acknowledge the existence of a self. In assessing personality types – the “energetic” character of On Liberty, for example, or the “active” versus “passive” characters of Representative Government – Mill is attending to what Aristotle called a hexis: an ingrained, habitually established way of being, of holding oneself together as a person. As I’ve said, for Mill the interesting

46

Ontology Revisited

question is the ethological one, viz., what kinds of institutions and norms produce what kinds of psychic musculature? The argument against benevolent despotism (and against the yoke of custom that endangers representative democracy), for example, is formulated in just these terms: “What sort of human beings can be formed under such a regimen?” Mill asks.38 Aristotle asked the ethological question too, as did Plato. But Mill’s anti-essentialism about the self dissolves the substrate of unified subjectivity that such considerations presuppose. Alan Ryan makes this claim at a more basic level, suggesting that Mill’s whole empiricist philosophy of mind breaks down under the weight of the need for something like a transcendental subject – or at least a proper self (as opposed to an inexplicable, dogmatically asserted fact of self-awareness, stipulated as accompanying the flow of sensations). Ryan is right, in my view, but the point can be extended to Mill’s concern with ethology. Surely one can’t really be a character-type if one isn’t really a self. Second, note the depiction of the character-type the being of which Mill regards as a good in and of itself. Mill consistently and pointedly describes the personality in question in kinetic terms: active, energetic, involving mental and moral analogues to “the muscular powers,” as he would have it.39 “Strong impulses,” he says, “are but another name for energy”; they “make the personal impulses vivid and powerful.”40 And again: “Whoever thinks that individuality of desires and impulses should not be encouraged to unfold itself, must maintain [that] . . . a high general average of energy is not desirable.”41 The contrasting personality is described as passive, “inert and torpid, instead of active and energetic,”42 “pinched and hidebound . . . cramped and dwarfed.”43 Representative Government is an exploration of how to develop the psychic forces in question via appropriate institutional arrangements. On the one hand, it’s a colonialist commentary on the need to hone, discipline and direct the uncultivated energies of pre-civilized peoples. On the other hand, it echoes the concerns of On Liberty regarding the political dangers posed by passivity. That the citizenry have a “striving, go-ahead character”44 is both a condition of possibility of representative government, and an ongoing product of it, says Mill. Conversely, [i]nactivity, unaspiringness, absence of desire, are a more fatal hindrance to improvement than any misdirection of energy; and are that through which alone, when existing in the mass, any very formidable misdirection by an energetic few becomes possible.45 For my purposes, what’s significant here is not whether Mill is correct to believe that apathetic, overly compliant citizenries are ripe for tyranny. Rather, it’s the fact that the entire analysis trades on the surface meaning of the dispositional terms. On a Humean ontology, Mill can be saying only that the passive charactertype routinely “does” x, in response to stimulus z, whilst the energetic charactertype routinely “does” y. But that is clearly not what Mill means to say. What Mill means to say, and plainly is saying, is that the passive personality is not energetic enough; that it lacks mental and moral powers; and that this deficiency of musculature matters. One last example: bureaucracy, Mill suggests, has much to

J. S. Mill: Humeanism and Human Capacities 47 recommend it as a political form. But it cannot be recommended. Why? Because compared to others “it is not equally favourable to individual energy of mind.”46 Third, Mill is interested in change, i.e., in the improvement of character. Here he runs into trouble, and he knows that he does, because of his commitment to determinism. The problem, as Mill notes, is that if behavior is “caused” by the presence of antecedent conditions, conditions upon which given consequents necessarily follow (or tend to follow, in the absence of a countervailing necessary tendency), then there would seem to be no room for novelty – in the psychological case, for self-improvement via the exercise of free will vis-à-vis one’s character. As noted in the previous section, Mill’s response is to say that determinism is not so off-putting or counter-intuitive as it may seem: all he is saying is that given what someone is like, he or she will behave in one way rather than another in given circumstances. Of course, as Ryan points out, all Mill has done is push the question back a frame: if one’s behavior in a given context is determined by one’s character, the issue then becomes whether or not one’s character is itself pre-determined by a prior set of antecedent conditions. Mill believes that he is entitled to answer this question with a no. One’s character, Mill wants to say, can be changed; including by oneself. Ryan argues that this simply gives way to a deterministic regress, and I think that he’s right. In the end, if Mill believes that character is caused, then, given his metaphysics, character must be part of the total set of law-like regularities, which regularities are already built into the permanent possibilities of sensation, as we’ve seen. There is simply no point of entry for the “un-caused” exercise of choice. It is worth recalling, in this regard, that Mill does believe that there are various uncaused “primary causes,” which are exempt from the Law of Causation (in that they assume the position of causes but not that of effect). He does not, however, count agents to be among them. The fact that Mill can’t derive novelty from a deterministic system is not just a metaphysical sticking point. It’s a problem that goes right to the heart of his normative claims. This is so because it is precisely the act of choosing – what to believe, what to care about, who to be – that Mill takes to be both the substance of freedom and the crucial catalyst for “improvement.” Indeed, being self-directed is precisely the activity fueled by the energetic personality’s mental and moral powers. And the ambiguity of “being self-directed” is intentional: the powers in question are constitutive of the energetic character-type, or hexis, and they drive the activities undertaken by the fully developed person. That Mill sees things this way comes across not just in chapter 3 of On Liberty, but also in the Logic and, again, in a more political register in Representative Government. In the Logic, for instance, he says: And, indeed, if we examine closely, we shall find that this feeling, of our being able to modify our own character if we wish, is itself the feeling of moral freedom which we are conscious of . . . And hence it is said with truth that none but a person of confirmed virtue is completely free.47 The quotation recalls the passage in Rousseau’s Discourse on the Origin of Inequality, wherein Rousseau tells us that our spirituality lies in our capacity to

48

Ontology Revisited

know ourselves to be capable of choice.48 But Rousseau, and Kant with far greater acumen, sees that if one presumes a mechanistic metaphysics, then, in order to be able to express such a sentiment, one will have to maintain, as Rousseau does, that while “Nature lays her command” on other aspects of the world, our will escapes such determination. Aristotle, by contrast, who is not constrained by Humeanism in the first place, has no comparable difficulty sustaining a concept of what he calls prohairesis, the exercise of choice regarding what kind of person to be, as exhibited through one’s actions. As Richard Sorabji has suggested, Aristotle can solve the so-called “fresh start” problem without saying that choices are uncaused, because Aristotle doesn’t define causality in terms of a rubric of deterministic law.49 Last, with respect to the issue of character Mill is not able to tell us very much at all about why the recognizably Aristotelian excellences that he commends really are excellences. In a sense this is simply to reiterate the earlier point regarding the lack of available grounding for the distinction between higher and lower pleasures. Mill is entitled to talk of observed regularities: some people routinely prefer the pleasures of a rich intellectual and civic life to the pleasures of individual, materially oriented pursuits. But as we have seen, he has no way to anchor the claim that the former are better. Aristotle, by contrast, can connect such virtues to the soul, i.e., to the human form, the full actualization of which he can plausibly presume to be a natural good. Finally, as a matter of meta-theory, Mill can’t talk at all well about the public sentiment, or social solidarity, the feeling of which he deems to be a higher pleasure, and the inculcation of which he believes to be of the utmost consequence. Here are two passages to bring to mind the flavor of his convictions on the topic. The first is from Utilitarianism, wherein he writes: The deeply rooted conception which every individual even now has of himself as a social being, tends to make him feel it one of his natural wants that there should be harmony between his feelings and aims and those of his fellow-creatures.50 And longer, from Representative Government: Still more salutary is the moral part of the instruction afforded by the participation of the private citizen, if even rarely, in public functions. He is called upon . . . to weigh interests not his own; to be guided . . . by another rule than his private partialities; to apply, at every turn, principles and maxims which have for their reason of existence the common good . . . Where this school of public spirit does not exist . . . (t)here is no unselfish sentiment of identification with the public. . . . The man never thinks of any collective interest . . . Were this the universal and only possible state of things, the utmost aspirations of the lawgiver or the moralist could only stretch to make the bulk of the community a flock of sheep innocently nibbling the grass side by side.”51 Mill is talking here about something very much like what Rousseau called a General Will, a shared identification with a whole that constitutes a “We.” The

J. S. Mill: Humeanism and Human Capacities 49 problem with him doing so is that he’s an atomist. He thinks that societies are aggregates: pluralities of parts, not emergent totalities. Now, let me be clear. Mill is entitled to talk about a harmony between the feelings of individuals if what he means by it is a situation in which the feelings of individual subjects are compatible, or even identical. Imagine, for example, as a limit case, an orchestra in which all of the members play the same parts of a piece at the same time, but on different kinds of instruments. Such a production might be said to be harmonious in a certain unusual sense of the term. In a normal orchestra performance, however, the piece as a whole is played by the orchestra as a whole, not simultaneously by each of its members. And in the normal case, the attribution of harmony would refer to precisely this fact, viz., that the parts have come together so as to form an aesthetically integrated whole. Mill is entitled to talk about harmony in the first sense, but not in the second. The problem is the image of sheep nibbling side by side tells us that he is dismissive of the first, that the orientation that he is so intent upon fixing in citizens is akin to that which would be found in a real orchestra. But such an orientation presupposes the existence of a collective subject – a collective subject for whom the collective good is good, and with whom individual subjects can be encouraged to identify, as Mill does encourage individuals to do.52 Let me be clear that in relation to this issue it is Mill’s atomism, not his Humeanism per se, that renders problematic his talk of the social. Atomism is not entailed by Humeanism, although as suggested earlier there are reasons why it appeals to Humeans by default. I shall return to this issue in Chapter 5. There are three main conclusions that can be drawn from the foregoing analysis. The first is that the familiar tension in Mill’s work between his utilitarianism and his perfectionism is epiphenomenal vis-à-vis the underlying contradiction between his Humean metaphysics and his neo-Aristotelian normative sensibilities. The injunction to maximize pleasurable sensations is readily compatible with Mill’s underlying ontology; the injunction to cultivate the as-yet fully realized capacities that are distinctive of human beings is not. As I also suggested, the deeper contradiction is evident even within his metaphysics. A correct diagnosis of the utilitarianism–perfectionism inconsonance is important because it allows us to not just identify the tension, but to explain it. Second, the analysis I’ve developed highlights the cost of assuming that social and political thought does not rest upon underlying philosophical categories – or, alternately, that it may be over-laid upon any categories. To use Mill as a place-holder, if one’s metaphysics is inadequate to sustain one’s position, then one ends up either talking about pleasure rather than flourishing, or advancing claims that one can’t consistently maintain, and therefore can’t defend well. Or both. Finally, it seems to me that if Mill, who was as serious and self-conscious about the challenge of such an undertaking as he was – if Mill can’t get a Humean metaphysics to support a neo-Aristotelian normative agenda, then it can’t be done, and there is reason call it a day. I will come back to this issue in Chapter 6, wherein I shall address it in relation to Martha Nussbaum’s so-called “capabilities approach.”

4

Kant and the Frankfurt School Freedom as Escape from the Transcendental Subject

Kant is a Humean as I am using the term. Despite his radical critique of Hume, he meets the criteria for three reasons: (1) he rejects a powers-based view of causality; (2) he believes in metaphysical necessity, but he denies that it is grounded in the phenomenal world; (3) he takes the identity of phenomenal objects to be fixed not by any intrinsic features of their own, but by the concepts through which they are unified in judgment. In short, if we ask what the phenomenal world is like, according to Kant, the answer is recognizably Humean, i.e., it is comprised of objects that are (1) causally inert; (2) causally related only in virtue of the nature of reason itself; and (3) essence-less.1 Notwithstanding Eric Watkins’ efforts to reconstruct Kant as a dispositional realist, most readers of Kant will take the preceding to be an unexceptional statement of the basic elements of Kant’s position – and it is not my aim to defend it. Rather, my purpose in this chapter is, first, to show how Kant’s account of free will fits into the Humean picture, and, second, to trace the complicated and contradictory way in which it plays out in appropriations of Kant by Herbert Marcuse, Max Horkheimer and Theodor Adorno. The argument of the present chapter I shall then extend into the next: having filled out the Humean framework such that it can be seen to generate a dichotomy between freedom and causation, I will look more closely in Chapter 5 at the construction of the free will debate in contemporary analytic philosophy.

Kant on Causation and Freedom Kant says in the Prolegomena that Hume’s treatment of causality awoke him from his dogmatic slumbers. Hume, as we have seen, had argued that there is no basis for the idea of a necessary connection between causes and effects: no logical entailment, no perception of a tether or bond of any kind. There is only the fact of sequence, accompanied by the subjective feeling of expectation. Mill, to the extent that it was a concern for him, took himself to have reinstated necessity by asserting (arguably on pain of circularity, though he himself did not believe so) that nature just is law governed, and that we are therefore justified in generalizing from our experience of local uniformities to the Law of Causation. Kant, writing roughly a century earlier, takes a different tack. His response is to say that objects of phenomenal experience are subject to causal law, but that this is so precisely

Kant and the Frankfurt School 51 because they are products of the synthetic faculty of reason, which establishes deterministic causal order a priori. In the course of turning natural necessity into transcendental necessity, it also happens that Kant re-introduces genuine materiality into the picture. Let me begin there, in setting out the view. As Hume had had it, to experience an object is to have a collection of impressions that are then unified by the imagination into a “substance.” We then come to believe, mistakenly, in the existence of real substances, non-conceptual objects to which our unified impressions refer. Kant, by contrast, maintains that there must exist something other than thought, something that provides concepts with their material content, as he puts it. But here it gets tricky. The “something other” is not, in itself, an empirical entity. Not unlike Aristotelian prime matter, it is a type of abstraction, a dialectical moment, we might say. In the Aristotelian case, concrete things are thought to be comprised of matter and form alike, never matter alone. The idea of prime matter therefore picks out an ontological fact – i.e., the materiality of substances – that cannot be isolated except analytically. By analogy, Kant insists that objects as we do or could experience them involve sensuous content and a priori synthesis alike, never just un-synthesized sensuous content. The analogy is a limited one, but it may be helpful for understanding the idea of a “material something” that occupies conceptual space but not, on its own, real space. Kant calls empirical objects, i.e., objects as we do or could experience them, “phenomenal objects.” Kantian phenomenal objects are different from Aristotelian substances in that, as we shall see in a moment, they are subject-dependent (transcendentally subject-dependent) in a way that Aristotelian substances are not. But they are also different from Hume’s collections of impressions and Mill’s clusters of possible sensation. Of the three – Hume, Mill and Kant – Kant is the only one who rejects the idea that phenomenal states exhaust the material content of objects. Whatever else remains to be said about them, Kant’s phenomenal objects are genuinely material. But here it gets trickier yet, because what remains to be said about Kantian phenomenal objects is that their materiality is not enough to constitute them as objects. They must also, for example, be located in time and space; be made to cohere as entities; and be constituted not just as substances, but as x’s or y’s. Moreover – and crucial for the present story – they must be subject to causal law. All of these features of real objects are bestowed upon them from without, Kant contends, through the synthetic operation of reason. This ranges from what he calls the pure forms of intuition that confer the fact of spatial and temporal location, to the use of the concepts that, in the act of judgment, give us objects of different kinds. Kantian phenomenal objects are therefore quite curious. On the one hand, unlike substances made out of Humean impressions or neo-Berkeleyan permanent possibilities of sensation, they are genuinely external, material things. On the other hand, their materiality is an insufficient condition of their existence as objects. As a proponent of the view that I’ve just summarized, Kant describes himself as having rejected subjective idealism, and replaced it with what he calls

52

Ontology Revisited

empirical realism paired with what he calls transcendental idealism. Empirical realism captures the idea that objects do not reduce to mental phenomena; transcendental idealism is the claim that objects have nevertheless always already been acted upon by what Kant calls a “transcendental subject” – that which, in relation to each of us, is the locus of the a priori synthesis presupposed by empirical experience. As noted above, one of the features of experience that Kant claims is given by the faculty of reason is the fact that phenomenal objects are related causally. The contention is that there is a “category of the understanding,” as Kant calls it, which has the function of establishing, a priori, the rule that objects as we experience them are subject to causal law, i.e., that they occur in sequences that we cannot but take to be necessary. Causality, then, on this view, is precisely the transcendental fact that we do, and must, experience the world as subject to such necessity. The faculty of reason does not fix which phenomena or kinds of phenomena precede or follow which others. That x’s are followed by y’s rather than by d’s, if they are, is an empirical matter. The transcendental fact is general: it’s that all objects, whatever regularities the world turns out to exhibit, fall under a rubric of invariant causal order. Thus Hume was mistaken, in Kant’s view, in that causality is a category of the understanding, not a customary feeling of anticipation. But in an important respect Hume was right. That is, causality is most decidedly not about the display of powers borne by entities, or any relationship of natural necessity involved therein. Having claimed that the phenomenal world is governed by necessary causal laws, Kant then has to address the question of whether or not human beings are metaphysically free. Can we initiate actions of our own choosing, or is our behavior, like all other events, the end result of a chain of invariant causal sequences? Kant’s response is this: on the one hand, human beings are phenomenal objects, just like other material entities. And insofar as we act in this capacity, as “empirical subjects,” our behavior is indeed subject to deterministic causal law, as is that of all phenomenal objects. Kant calls behavior that is so enacted and so governed “heteronomous.” Heteronomous behavior ranges from reaching for a cup of tea because one enjoys it so, for example, to grabbing one’s precariously positioned cup to keep it from falling off of the edge of the table. Such behavior is necessitated, in Kant’s view, even if it is done willingly. Unlike other phenomenal objects, however, human beings are endowed with the faculty of reason. In the 1st Critique Kant details what this faculty contributes to experience as it pertains to the cognition of objects. But Kant also provides an account of moral experience. When he shifts from the analysis of pure (theoretical) reason to the analysis of pure practical reason, he posits an analogous transcendental figure to that of the transcendental subject. The “transcendental ego,” as he calls it, is the locus of what Kant takes to be an inescapable presupposition of, in this case, action – namely, that we are the originators of what we do, even if we are coerced, i.e., that we have free will. Just as we cannot help but experience the world as causally ordered, we cannot help but experience ourselves as being the source of our own agency. Unless we are mentally ill, we don’t experience ourselves as being possessed by an external agent who has taken control of

Kant and the Frankfurt School 53 us, over-riding our power of intentionality. Instead, we experience ourselves as being at the helm, as it were, even when we are compelled to follow a course of action that someone else has set for us. In a sense, then, Kant’s response to the problem generated by determinism is to say that we are free in that, unlike other phenomenal objects, we cannot help but experience ourselves as free. But there’s more to it than just this, according to Kant. Specifically, it turns out that the sense of ourselves as free that we hold a priori, via the pure practical reason of the transcendental ego – it turns out, says Kant, that freedom in this sense implies a normative structure for action, such that not all intentional behavior can be said to be free after all. Genuinely free actions Kant calls “autonomous.” Most acts are not autonomous. Most acts are heteronomous. Autonomous action is different from heteronomous action in two ways. It is motivated by a unique purpose, and it satisfies strict conditions with respect to content. Let me begin with the latter. Action that is governed by pure practical reason rather than by causal necessity conforms to what he takes to be several versions of the same requirement, that of the categorical imperative: “Act only on that maxim through which you can at the same time will that it should become a universal law.”2 With respect to motivation, meanwhile, autonomous action is defined as being driven by a sense of duty (i.e., merely because it meets the requirements above), not by desire or inclination. The exercise of free will is thus equated with specifically moral action. To be free is to act autonomously as a legislator of universal norms (qua transcendental ego), rather than heteronomously, in pursuit of that which, for whatever reasons, one desires or regards as advantageous (qua empirical ego). The phenomenal world is deterministically ordered, but we rise above it when we act out of duty, in accord with moral law rather than causal law. Of course, we are embodied even when we act autonomously. Nonetheless, autonomous behavior is taken to escape the hold of causal determination, in that it just is, in Kant’s view, the practical expression of reason itself.3 As one might expect, Kant’s dualism puts pressure on him to try to integrate the a-causal freedom, or spontaneity, that is grounded in the transcendental ego (the locus of the synthetic a priori of practical reason), on the one hand, and the determinism that governs the world of experience, grounded in the transcendental subject (the locus of the synthetic a priori of theoretical reason), on the other. This task he undertakes explicitly in the Critique of Judgment, wherein certain forms of aesthetic (or sensible) responsiveness are distinguished from inclinations that are dismissed as merely subjective, and thus deterministically caused. In aesthetic judgment, we might say, we exercise a form of transcendental reason that is not “pure,” not analytically separable from sensation. It is, therefore, related to the empirical aspect of empirical subjects in a different way than is duty as conceptualized in the Groundwork, say. This is a point to which I shall return, in connection with my comments on Marcuse in particular. Kant’s account of free will also creates an expectation that he will spell out the socio-political conditions of possibility of freedom as he has defined it – though it

54

Ontology Revisited

would seem to be in the nature of the case that Kantian freedom is equally possible in all material circumstances. To offer such an account would be to effect a reconciliation of freedom and necessity at a different level of abstraction, i.e. at the level of history.4 Kant scholars have generated an extensive literature on whether or not Kant has even attempted to do this, which I shall not address here but to which I have tried to be sensitive in my summary. For the purposes of my own argument, I need only to have clarified that Kant (a) having advanced a deterministic version of the Humean ontology; and yet (b) wanting to preserve free will; sees, correctly, (c) that he has no choice but to conceive of free will in terms of a-causal spontaneity, as escape from causal determination as such.

Kantian Metaphysics and the Frankfurt School As I’ve said, I want to reflect upon the significance of Kantian metaphysics for social and political thought not by looking at Kant’s own work, but instead by considering the issue in relation to the critical social theory of Herbert Marcuse, Max Horkheimer and Theodor Adorno. One reason for this strategy is that Katrin Flikschuh, in particular, has done an excellent job in showing the connection between political philosophy and metaphysics in relation to Kant himself.5 But another reason is that I am less persuaded than is Flikschuh that Kant’s own treatment of freedom in the political sphere, to use an unsatisfying locution, is consistent with his definition of the concept. Thus, ironically, there may be a clearer example to be had of Kant’s metaphysics giving shape to loosely Kantian social and political thought in the case of thinkers who are engaged with Kant than in the case of Kant himself, who arguably backs away from the political conclusions entailed by his conception of autonomy. One might object that in focusing on Marcuse, Horkheimer and Adorno I am in some sense cheating, because in one way or another all three of these thinkers were themselves concerned to identify the social implications of Kantian metaphysics. My response is two-fold. First, it is true that I am not the first to argue that social and political thought contains ontological presuppositions – and equally true that those who precede me in developing such a line of argument include members of the Frankfurt School. But this fact does not tell against the thesis. Second, my strategy will not be limited to simply illustrating the ways in which these three thinkers self-consciously introduce the categories of Kantian metaphysics into social and political thought, although I will indeed do so. Rather, in each case I will go on to show how the thinker’s residual attachment to a Kantianinflected Humean metaphysics undermines the position that they are undertaking to advance. Let me begin with Marcuse, as the connection to Kant’s metaphysics is perhaps most direct in his work. Marcuse is well known for his Hegelianism. But in his discussion in Eros and Civilization of the psychoanalytic dynamics that he believes would figure in an objective, rational freedom, he lands ultimately upon Kant. The negative thesis of the work is this: while Freud was right to think that

Kant and the Frankfurt School 55 civilization has heretofore been predicated upon the repressive function of what he, Freud, called the reality principle, he (Freud) was wrong to assume that this must be so. The positive suggestion, meanwhile, is that the faculty that Kant called aesthetic judgment can, at this point in history, sustain what Marcuse terms “libidinal morality.” Libidinal morality, in turn, is associated with a nonrepressive form of sublimation.6 Marcuse advances the negative claim via two distinctions. The first is between the concept of repression, or basic repression, and that of surplus repression. The term basic repression refers to the degree of sublimation required to engage in a fully socialized way with the natural world. Surplus repression, in turn, refers to the difference between basic repression and the degree of sublimation required to sustain any given, historically specific, form of alienated labor.7 Intersecting these categories is a second distinction, that between the concept of the reality principle and what Marcuse calls the performance principle. The reality principle is Freud’s term for those internalized norms of self-denial and self-control that constrain the desires that would otherwise be allowed to be gratified in accordance with what Freud called the pleasure principle. The reality principle is completely general. Marcuse’s notion of a performance principle refers by contrast to “the prevailing historical form of the reality principle,” thereby giving content to Freud’s general concept.8 On the basis of these two distinctions, Marcuse articulates the view that human productive capacity has reached a point (precisely as the result of past sublimation) such that the preservation of civilization no longer requires adherence to a repressive reality principle. In a sense, all repression is now surplus-repression, required by the performance principle that is specific to capitalism, but no longer required for civilized human survival as such. The negative thesis is thus that what appeared to be a timeless, necessary antagonism between the reality principle and the pleasure principle is in fact a reconcilable one; the positive step, as I’ve suggested, is to undertake to give theoretical shape to the identified possibility. Already in Freud’s own work, Marcuse believes, there are categories relevant to the task – in particular those of phantasy and eros.9 Phantasy, or imagination, is a form of conscious thought that, unlike the rest of our conscious awareness, is not subject to the reality principle. Rather, it “continues to speak the language of the pleasure principle, of freedom from repression, of uninhibited desire and gratification.”10 Its role is to [preserve] the ‘memory’ of the subhistorical past when the life of the individual was the life of the genus, the image of the immediate unity between the universal and the particular under the rule of the pleasure principle.11 Unlike the other drives that are aligned with the pleasure principle, however, the imagination is cognitive. Imagination has a truth value of its own, which corresponds to an experience of its own – namely, the surmounting of the antagonistic human reality. Imagination

56

Ontology Revisited envisions the reconciliation of the individual with the whole, of desire with realization, of happiness with reason.12

The concept of eros, meanwhile, Marcuse takes to open the way for a “quantitative and qualitative aggrandizement of the meaning of sexuality itself.”13 Freud understood eros to be a “life instinct,” not exclusively or even distinctly sexual.14 But, Marcuse argues, the very idea suggests the potential for a broadened, polymorphous sexuality, released from the hold of genital supremacy, as it’s termed.15 The transformation of sexuality into eros, as Marcuse puts it, would differ from “the release of constrained sexuality” in a context determined by the reality principle. In the latter case, “the libido continues to bear the mark of suppression and manifests itself in the hideous forms so well known in the history of civilization.”16 A polymorphous sexuality, by contrast, would be likely to be selfregulating. Rather than requiring localized expressions of pent-up desire, it would involve a generalized eroticization of non-explicitly sexual activity – especially, Marcuse thinks, the activity of work.17 “The culture-building power of Eros,”18 Marcuse contends, “creates its own cultural order.”19 In particular, it creates – or, more precisely, implies – an order in which labor could take on the quality of “free play,” i.e., in which labor would be an artistic expression of the imagination. All of this said, it turns out, according to Marcuse, that the conceptual resources latent in Freudian theory are to be found in their most rigorous formulation in Kant. As noted previously, in the Critique of Judgment Kant proposes a third transcendental faculty, one that generates a rational type of sensuous pleasure. Marcuse’s claim is that aesthetic judgment as theorized by Kant provides a model for the “new rationality of gratification in which reason and happiness converge”20 that is implicit in the idea of phantasy – a model for the self-regulating “libidinal morality” of eros. In aesthetic judgment, sensuousness combines with cognition in the pleasurable apprehension of form. In Marcuse’s words, “in a free synthesis of its own, it [aesthetic imagination] constitutes beauty”; in so doing, “sensuousness generates universally valid principles for an objective order.”21 Marcuse notes that the order in question is characterized by Kant as one of “purposiveness without purpose” and “lawfulness without law.”22 In the case of aesthetic judgment, the object is experienced “as freely being itself”; “all links between the object and the world of theoretical and practical reason are severed, or rather suspended.”23 Marcuse gives a Heideggerian gloss to this scenario, describing the object as thereby being encountered by the subject as “given,” having been “release[d] . . . into its ‘free’ being.”24 In such a context, “subject and object become free in a new sense,” he says.25 Marcuse takes such a mode of being to be the core of a non-repressive reality principle, a post-capitalist performance principle that would serve to reinforce what Schiller called a “play impulse.” A society in which the play impulse is expressed rationally as a totality of generally eroticized, democratic social relations that conform to laws of beauty would, in Marcuse’s view, be genuinely free. Now, one thing to be said of Marcuse’s account of freedom is that, patently, it is a self-conscious appropriation of Kantian metaphysics at the level of social

Kant and the Frankfurt School 57 and political thought (with the caveat that, in Marcuse’s view, free will appears in the Kantian architectonic not where Kant thinks it does, as pure practical reason exercised in accordance with duty, but rather as the spontaneity associated with aesthetic judgment). As such, it is a relatively straight-forward illustration of the claim that social and political thought has a metaphysical infrastructure. This said, one might reply that while Marcuse’s reliance on Kant may be of interest to a historian of ideas, or to an intellectual biographer, it’s incidental to the account that he offers of freedom: omit the chapter on Kant in Eros and Civilization and there may be a rhetorical loss, but the theory would not be weakened philosophically. There are two responses that one might give to this kind of objection. The first is that it is not at all clear that the metaphysics does no work. True, it may be that Marcuse uses Kant merely to advance by analogy his own independently grounded conception of freedom as creative play in keeping with a non-repressive reality principle. But it may not be. It may be, instead, that Marcuse, if challenged by a determinist, would fall back upon a Kantian argument, in the end, in order to preserve the very idea of free will. Second, and more important I think, claims at the level of social and political theory commit one to basic metaphysical claims whether one likes it or not. If one advances an argument on behalf of freedom construed as a-causal “purposiveness-without-purpose,” for example, one is already implicitly committed to a metaphysics that allows for the existence of agents for whom such spontaneous, indeterministic activity is a real possibility. Ultimately, then, it is the overtly Kantian nature of Marcuse’s account of freedom that is the truth-maker, as the analytic philosophers like to say, for the claim that Marcuse’s account of freedom has Kantian metaphysics built into it. But there is more to my own case for the relationship between metaphysics and social and political thought than that Marcuse has self-consciously transposed Kant’s metaphysics into a socio-political register – which he has done. The further point, which is perhaps less obvious, is this: it’s not clear that a Kantian metaphysics can do the work that Marcuse needs done. I’ll begin with the idea that aesthetic judgment involves a non-coercive (or nonrepressive) relationship between subject and object, a relationship in which the object is “given” to the subject in such a way that it, the object, is “released into its ‘free’ being,” as I’ve quoted Marcuse stating it above. The problem here is that Marcuse’s conception of a non-coercive relationship between subjects and objects presumes objects that have their own forms – forms that are not given by the subject, but rather are met (with pleasure) by the subject. But Kantian phenomenal objects do not have form “on their own,” in the sense that Marcuse requires; they get their form from subjects. I will set aside the question of whether (a) Marcuse has misconstrued Kant’s account of aesthetic judgment or (b) Kant himself has this same problem, in connecting the 3rd Critique back to the 1st. What matters is that if Marcuse is going to define freedom in terms of a response to objectivelygiven form, as he does, then he is already committed to an ontology that admits of objectively-given form, which Kant’s does not. Of course, Marcuse is not forced to abandon his account of freedom simply because it is at odds with Kantian

58

Ontology Revisited

metaphysics. He has only to abandon an appeal to Kantian metaphysics on its behalf. For the purposes of my own argument, however, it is the mere fact of the inconsistency that is telling. One cannot rest one’s social theory a-top any given metaphysics. Marcuse has other form-related problems, too. In “The Concept of Essence,” for example, he runs into trouble by defining a materialist, or dialectical, concept of essence in terms of what he calls real potentiality.26 Here, the issue is a bit different. Here it’s not that he’s talking about objectively-given form, but trying to do so by way of a metaphysics that disallows such a thing; here the problem is that he’s advancing a not-so-robust account of form, and wanting it to do more than it can. That is, if being a real potential is all it takes to be an essence, then – unless there is at all times only one real potential vis-à-vis a given situation – it follows that appeals to essence are not going to help one decide which of the several real potentials that exist in any one historical moment ought to be realized. One will not be able to suggest, as Marcuse does, that the achievement of socialism is preferable to the achievement of fascism because, although the potential exists for both, the former represents the full actualization of humanity’s essence and the latter does not. Finally, Marcuse’s account of freedom – in Eros and Civilization and elsewhere – doesn’t just presuppose objects with their own forms; it also presupposes powers. For one thing, Marcuse’s neo-Freudianism involves him in unabashedly realist talk of drives, impulses and the repressive or non-repressive sublimation thereof. These are all dispositional terms. “Drive” and “impulse” refer here to something like psychic kinetic energy. The terms “repression” and “sublimation,” meanwhile, refer precisely to the constraint and redirection of said psychic forces. If the forces weren’t actually forceful, then the concepts that refer to their constraint and redirection would lose their intended sense. As would any claim to the effect that civilization is the by-product of such constraint – civilization being “fueled,” as it were, by redirected psychic energy. One could still talk about certain states of affairs following regularly upon others. But Freud himself was saying something more than that. Freud was talking about the repression and sublimation of active forces, of drives. Finally, if the concept of repression didn’t refer to the constraint of real powers, then it would be hard to make sense of its normative sense – be the valuation a positive one, as for Freud, or a negative one, as with Marcuse. Similarly, to the extent that Marcuse has taken on Marx’s analysis in Capital, he has committed himself to a theory that, on its own terms, is non-Humean to the core. This is not the place to enter into a discussion of Marx’s metaphysics, but at a minimum it’s clear that Marx draws upon robust notions of powers even in his most basic description of the commodity, considered both as use-value and as exchange-value. Use-value is defined in the opening chapters of Capital in overtly dispositional terms, as that which something has the power to do (in virtue of the kind of thing that it is). Exchange-value, meanwhile, as an indicator or expression of what Marx calls value, turns out to refer to a quantity of abstract labor power, itself nothing other than generic productive capacity – “activity as such,” we

Kant and the Frankfurt School 59 might say. Less technically, even in the Communist Manifesto the language is incontrovertibly dispositional: history is construed as being given shape not by static laws that govern sequences of states of affairs, but rather by seismic-like pressures created by the force of human creative powers coming up against prevailing social relations of production. At the same time, even Marx has his Kantian moments. And it should be noted that Marcuse makes much of Marx’s observation that capitalism, because of its enormous productive and technological capacities, has set the stage for a minimization of the sphere of necessity relative to the sphere of freedom. Marcuse’s attraction to this bit of Kant in Marx is of course consistent with his, Marcuse’s, explicit appropriation of elements of Kant’s metaphysics. Let me be clear. I do not mean to suggest that Marcuse is in the same position as Mill. The point is not that he has committed himself unequivocally to a Kantianstyle Humean metaphysics. The point is that when he does make metaphysical appeals, they are of the wrong type. For the purposes of sustaining his critical social theory, he needs a dispositional realist ontology, not a Humean one. Horkheimer’s relationship to Kant is more attenuated than is Marcuse’s. Marcuse gives us an account of freedom that is explicitly modeled on the spontaneity of Kantian aesthetic judgment. Horkheimer, by contrast, decries talk of metaphysics and is at pains to insist that what he calls materialism involves him in no such thing. Nevertheless, the approach to freedom that he commends amounts to a materialist transposition of the purported autonomy of pure practical reason, and he himself puts it in these terms. First, therefore, let me explain what Horkheimer has done with Kant’s conception of free will, so that the way in which it configures his social theory will be clear. Then, as in the case of Marcuse, I will ask whether or not a Kantian-inflected Humean metaphysics can do the work that Horkheimer’s social theory needs its implicit metaphysics to do. Kant’s claim, recall, was that freedom is the fact that we have the ability to act autonomously, i.e., in a way that is governed by the logic of moral duty rather than by causal necessity. Acts governed by duty, again, are those that are done for no reason other than that they are consistent with the principle of being enjoined categorically as opposed to being a means to a desired end. The seat of freedom is the transcendental ego, which is the locus of pure practical reason. Pure practical reason is connected to the empirical subject via the will. When we act heteronomously, our actions occur within a metaphysically deterministic context, just as other phenomenal events do. But when and insofar as we exercise our will qua autonomous moral agents, the context is rendered non-deterministic. As Rousseau put it, in the passage that I called to mind in Chapter 3, For physics may explain, in some measure, the mechanism of the senses and the formation of ideas; but in the power of willing or rather of choosing, and in the feeling of this power, nothing is to be found but acts which are purely spiritual and wholly inexplicable by the laws of mechanism.27

60

Ontology Revisited

Horkheimer’s reaction to all of this can be reconstructed as follows. He assumes, to begin with, that Kant’s view contains sociological information in encoded form. The encrypted but accurate insight is this: on the one hand, in capitalism people are required to act individually and instrumentally. Moreover, we are forced to pursue our interests in an environment that is outside of our control. (For the majority of us, this is so in an immediate and obvious way, in that we are compelled to sell our labor power to someone other than ourselves, be this an individual or a corporation, who owns both the means of production and the fruits of our and others’ labor power. Ultimately, however, even those who are owners – themselves acting individually and instrumentally – must bow to the Law of Value, to the imperative to realize maximal profit.) On the other hand, society is in reality an interrelated whole, one that in principle could be collectively and self-consciously directed by its members – even if at present it is not. The fact that the relations of production in capitalist society are experienced by its members as a set of invariant and coercive natural economic laws appears in Kant’s work, says Horkheimer, as the heteronomous determinism that is taken to obtain in the phenomenal realm. The additional fact that things could be different, that society is a potentially (though not actually) rational collectivity – this fact shows up in Kantian philosophy as universal moral law. Finally, the conflict in Kant’s work between the two (between heteronomous action and duty) Horkheimer sees as an abstract expression of the fact that, in capitalism, individual behavior at best advances the collective good only accidentally. It’s worth quoting Horkheimer at length here: The basis of the spiritual situation in question is easily recognized upon consideration of the structure of the bourgeois order. The social whole lives through unleashing the possessive instincts of all individuals. . . . Each is left to care for himself as best he can. . . . The categorical imperative holds up ‘universal natural law,’ the law [Lebensgesetz] of human society, as a standard of comparison to this natural law of individuals. This would be meaningless if particular interests and the needs of the general public intersected not just haphazardly but of necessity. That this does not occur, however, is the inadequacy of the bourgeois form: there exists no rational connection between the free competition of individuals as what mediates and the existence of the entire society as what is mediated. The process takes place not under the control of a conscious will but as a natural occurrence. The life of the general public arises blindly, accidentally, and defectively out of the chaotic activity of individuals, industries and states. The irrationality expresses itself in the suffering of the majority of human beings.28 Admittedly, this much is to say that Horkheimer considers Kant’s metaphysics to be displaced social theory, not really metaphysics at all; it is not necessarily to say that Horkheimer’s own social theory is a working up of metaphysics. One might be tempted to reply that which way this call is made depends upon whom one

Kant and the Frankfurt School 61 asks, and leave it at that. But the metaphysics implicit in Horkheimer’s position comes clearly and independently into focus once we fully reconstruct his response to Kant. In addition to claiming that Kantian philosophy contains encrypted sociological information, Horkheimer argues that the Kantian conception of free will carries with it a call for its concrete, historical realization. “In this society of isolated individuals,” he writes, “the categorical imperative . . . runs up against the impossibility of its own meaningful realization. Consequently, it necessarily implies the transformation of this society.”29 Specifically, then – though I have sharpened the resolution so as to clarify the Kantian character of the analysis – the categorical imperative points to the idea that society must be transformed in such a way that it becomes the material expression of pure practical reason – a collective enactment of autonomy – rather than being the material expression of heteronomous necessity, as it is at present. Individual acts of moral duty might be at odds with the heteronomous pressures of capital, but in order to remove such pressures – and the suffering they produce – the totality must be reclaimed and reoriented such that it, rather than a posited transcendental ego, is a source of real universality. Only in such a society, says Horkheimer, could reason ever be genuinely transparent to itself – the goal that Kant had intended to achieve philosophically.30 Only in such a society would there be a genuine coincidence of reason and freedom. As I’ve said, Horkheimer sees himself as having no part in what he thinks of as idealist philosophical discussions about the nature of free agency. So when I observe that he nevertheless offers up an account of achieved human freedom, and that the account in question can be seen to be an expression of Kantian metaphysics in the form of social and political theory, it should be clear that I do not mean to suggest that, à la Marcuse, Horkheimer has explicitly pointed to Kant as having correctly delineated the structure of autonomous action. What I do mean to say is that whether he likes it or not, whether he does so for dialectical purposes or not (i.e., because he regards Kantianism as the moral theory par excellence of the capitalist period), Horkheimer advances a normative social theory the categories of which are conceptualized in Kantian metaphysical terms. Capitalism is construed as the realm of necessity. The potential for democratic socialism is cast as the real referent of the transcendental ego. Freedom is pure practical reason, made manifest. This is a social theory that plainly presupposes a Kantian metaphysics. I do not mean to imply that Horkheimer’s thinking is exclusively Kantian – that it is not also importantly Hegelian, for instance. My claim is only that whatever else one wants to say of Horkheimer’s social theory, Kantian metaphysics is built into it. The next step is to ask, as I did in relation to Marcuse, whether or not Horkheimer’s social theory requires, and implicitly presupposes, something stronger than a Kantian variant of Humeanism in order to do justice to its own terms. Of particular interest is whether or not Horkheimer needs powers. And as in the Marcuse case, the answer has got to be yes. At a minimum, in keeping with Marcuse, Horkheimer analyzes capitalism not just through the lens of Kant’s Critique of

62

Ontology Revisited

Pure Reason, but also through the lens of Capital. We may add to the points raised earlier in this regard a word about the relatively technical concepts of reification and fetishism. Both are central to Marx’s analysis, and both trade negatively on the idea of real powers. To reify is to represent something active, something with powers to do, as something intrinsically inert, something “thing-like.” If nothing at all had the power to do, then there would be nothing to which the concept could properly apply. It does not follow from this that things do have powers, only that to use the concept of reification is implicitly to affirm the idea that they do. To fetishize, meanwhile, is to attribute powers to entities that don’t really have them, or at least not the ones in question. In the case of chapter 1, section 4 of the first volume of Capital (“The Fetishism of Commodities and the Secret Thereof”), for example, the claim is that commodities are fetishized in the sphere of circulation in that exchange-value, which is really a ratio between quantities of self-valorizing abstract human labor power, is thought to be derived from the properties of inanimate objects. The further argument, of course, is that the confusion about the nature of value arises in the context of a social form in which human creative powers have indeed been systematically alienated. But the bottom line is that the powers in question are understood to be real ones. In addition to undergirding his use of Marx, Horkheimer needs powers – and implicitly helps himself to them – in order to make sense of a different claim in “Materialism and Morality,” one that I have not yet mentioned, which is that morality is ultimately a psychic state at the core of which is love. Of love, he writes: “[t]o all, inasmuch as they are, after all, human beings, it wishes the free development of their creative powers.”31 Of interest here is the ascription to human beings of creative powers real enough to be morally salient. And it is worth adding that Horkheimer treats love itself as a power that may or may not be socially cultivated. Thus he needs an ontology that allows for dispositional properties both to be able to talk in Marxist terms about “the senseless crippling of powers,” and the fact that “at no time have all powers been so horribly fettered as in this generation,” but also to be able to talk about the “atrophy of kind human inclinations.”32 Horkheimer himself, I imagine, would say in response – does say, in “Materialism and Metaphysics” – that materialism as he intends it involves no particular metaphysical commitments whatsoever, other than a belief in the existence of physical sense-data.33 “Idealism,” he writes, understands its various systems to be attempts at answering the same eternal questions, the same eternal riddle, and it likes to speak of philosophers conversing with each other across the millennia, because they all deal with the same theme. But the materialist’s views are essentially determined by the tasks to be mastered at the moment.34 The reply to this type of move has to be that one can appreciate Horkheimer’s pragmatist impulses – an echo of Marx’s own naturalist talk of “real premises” in The German Ideology, and of the dissolution of philosophy into praxis in the

Kant and the Frankfurt School 63 Theses on Feuerbach – without conceding the point. Horkheimer may declare ontological commitments to be the stuff of muddle-headed idealists, but, apart from anything else, the minute he differentiates materialism from positivism, he replaces an initial statement to the effect that proponents of the former agree with those of the latter that what exists is “only what is given in sense experience” – he replaces this with the claim that “every existent [must] manifest itself through the senses.”35 Such a claim is a bona fide metaphysical position, involving him in a bona fide metaphysical disagreement with those who really do hold the positivist view. Moreover, I prefaced the point by stipulating “apart from anything else.” As we’ve seen, the “anything else” refers to a host of ontological commitments, ranging from dispositional properties to collective subjects with the capacity to be free. With Adorno, the relationship to Kant is even more complicated. Marcuse, as we’ve seen, models freedom on the reconciliation of reason and sensuousness that he finds in Kant’s notion of aesthetic judgment. Horkheimer, meanwhile, posits a categorical imperative made historically manifest, and interprets the present real lack of freedom through the lens of Kant’s portrait of a blindly deterministic phenomenal realm. Adorno differs from both, in that his first step (or what can be seen to be a first step, in a reconstruction of an extremely complex analysis) is to argue that Kant isn’t saying what he – Kant – thinks he’s saying. Properly understood, Adorno believes, transcendental idealism captures the fact (as Adorno sees it) that the subject–object relation is actually chicken-and-egg-like – except without recourse to a notion of DNA, i.e., to a notion of objectively given form. If we ask which comes first, the answer will always be that each presupposes the other. This fact, which Adorno takes to be the unintended but actual thesis of the Kantian system, can be seen to be the analogue, with respect to Adorno, of Marcuse’s and Horkheimer’s relatively more straight-forward appropriations of Kant’s notions of aesthetic judgment and pure practical reason respectively. At the same time, however, and more so than Marcuse or even Horkheimer, Adorno uses Kant’s views as Kant himself conceives of them to diagnose what is wrong with existing society. We might say that Kantianism taken at face value yields insight into domination, for Adorno, whilst its implicit “truth-value,” to use Adorno’s locution, provides a starting point for thinking about freedom. I’ll first show how each of these engagements with Kant works, then make the case that Adorno’s ability to see his own line of argument through is limited by his own residual attachment, via Kant, to Humeanism. Kantianism figures in Adorno’s treatment of domination at every turn. On Kant’s account of empirical experience, recall, the faculty of pure reason gives form to that which is not reason, thereby producing phenomenal objects of one kind or another. On Kant’s account of moral experience, meanwhile, pure reason just is the categorical imperative. Those acts that conform to the categorical imperative, and are done for no reason other than that they do, are considered both moral and free. But how does this relate to domination? The key issue is the relationship between subject and object, on Kant’s model. For the Kantian, phenomenal objects derive their identity solely from the cognitive activity of

64

Ontology Revisited

subjects. Objects don’t depend on subjects for their materiality, but absent the synthetic operation of transcendental reason and the application of empirical concepts to the products thereof, there would be only inchoate sensuous manifold. Ontologically, then, phenomenal objects stand to subjects much as cookies stand one who has cut them from dough with a cookie cutter. Adorno characterizes this view of material identity as a violence done to the integrity of objects – objects that, as he puts it, “do not go into their concepts without leaving a remainder.”36 Note that the concern is not, or at least not primarily, with the classification of already-constituted objects, relative to nominal kinds derived from human interests: classifying dogs as “pets,” say, or small, extra-personable dogs as “companion breeds,” and others as “working dogs.” Nor is it primarily a concern with using words to refer to material entities. Rather, to extend the cookie-making analogy, the concern is with the idea that the only aspect of a dog that is attributable to it “on its own” is the brute fact that it is not purely ideational. This view, Adorno suggests, runs rough-shod over that which is not the cognizing subject, and does so at the very deepest level. Once such a point has been made, there are two available responses. Either one says (a) that there is no need to worry, because, as luck would have it, transcendental idealism isn’t true; or one says (b) that transcendental idealism is true, and so, like it or not, we can’t just beef up objects because we wish to defend their ontological honor. As I read him, Adorno is actually ambivalent, in the end, about which way to go. And this because he is ambivalent about whether or not phenomenal objects are something other than synthesized bits of otherwise-amorphous sensuous manifold. I shall return to this issue below. For now, I will say that I see Adorno as manifestly rejecting the Kantian view, but also as not being prepared to move beyond it. Note, however, that with respect to identity thinking, as Adorno calls it, it is not true that either there is (a) no need to worry, or (b) need to worry, but nothing that can be done. For if transcendental idealism is false, and phenomenal objects have their own forms, as Aristotelians maintain, then there will indeed be reason for concern if material entities are nevertheless subject to externally imposed ascriptions of identity, on the basis of a Kantian belief to the contrary. Of course, an Aristotelian herself may very well regard nature as nothing other than raw material for human purposes, but the Kantian view, insofar as it can be said to reduce the material world to undifferentiated sensuous content, is even more likely to be bound up with a purely instrumental attitude toward nature. The Kantian, that is, can’t help but take herself to be master over that which is not conceptual, insofar as phenomenal objects are only objects in the first place in virtue of that which is conceptual. If transcendental idealism is true, meanwhile, one might still be concerned by the God-like stance toward phenomenal objects that it implies for subjects. And if one were, then one might argue, for example, that subjects must be mindful of their metaphysical power over objects at the level of constitution, and must take steps to correct for it at the level of behavior. Perhaps even more pressing, if the relationship between Kantian subjects and Kantian phenomenal objects is by default one of domination, we will want to

Kant and the Frankfurt School 65 know what this means in the case of objects that are human beings. How does the idea of the violence of constitutive subjectivity apply when it comes to relationships between embodied subjects? The first thing to see is that, for the Kantian, people are different from other phenomenal objects. On the one hand, insofar as we are empirical entities, we too are constituted as such via the synthetic operation of reason. On the other hand, we are the kind of thing that we are independently of the cognitive doings of other subjects: we fall into the kind “human being” on the basis of being rational subjects ourselves. This is a non-Kantian way of putting it, to be sure, since for Kant we are never simultaneously a phenomenal object and a bearer of something like the human form. Rather, Kant has to say that we are actually two different kinds of entity, (a) phenomenal objects whose identity as such is determined via the application of empirical concepts, and transcendental subjects (or egos), (b) noumenal selves who just are the form of reason itself. But if we hold at abeyance the commitment to dualism – one that even Kant has difficulty maintaining – we will see that on at least a roughly Kantian picture it will turn out to be the case that phenomenal objects except for human beings are given their identities via the application of concepts by a subject, whereas human objects just are subjects. Even if we stipulate that one’s identity as a human being is contingent upon one synthesizing oneself via one’s own transcendental subjectivity, it would seem that the violence done to objects that are subjects cuts less deeply, on the Kantian metaphysics, than does the violence done to other objects. Moreover, in addition to the natural limit to the reach of constitutive subjectivity provided by our having our own form, there is the injunction – which Kant believes we experience as compulsory – to treat others as form, i.e., as noumenal selves, or ends, rather than as phenomenal objects. For Adorno, however, the very next question will be whether or not there are ways in which Kantian universalism is itself connected to domination. Adorno’s considered answer is “Yes and no,” and I shall come back to this point when I set out how a “dialecticized” rendering of Kant informs his discussion of freedom.37 For the moment, though, it’s the “yes” part that’s of interest. The universality of Kantian practical reason is connected to domination in two ways, Adorno suggests. First, the very particularity that Kant views as morally irrelevant, Adorno takes to be worthy of what I will go ahead and call moral recognition. While the Kantian view may, in the manner set out above, involve or at least require the idea that human objects have our own form, it most certainly does not allow us to have empirical attributes that anyone should care about. On the contrary, it treats as in principle morally inconsequential the concrete characteristics that we do have – and this, one might think, accomplishes by omission the violence to identity done to other kinds of objects via the principle of constitutive subjectivity. Second, formalistic reasoning may be just as easy, if not easier, to use as a tool for domination than are more substantively committed forms of thought. Here the logic is basically the opposite of Kant’s: while there may be reason to believe that morality would be tainted were normative judgments connected to particular interests, one might just as well conclude that an indifferent, purely formal reason will be impervious to suffering.38

66

Ontology Revisited

I have been talking thus far about the way in which Adorno’s diagnosis of domination rests upon Kant’s account of the constitution of objects: domination, from this perspective, is a metaphysical disregard for the integrity of that which is other than the knowing subject qua rational being. But other aspects of Kant’s metaphysics figure into the analysis as well. In particular, Adorno regards Kant’s accounts of freedom and causality alike as having profound political ramifications. Indeed, he characterizes Kant’s treatment thereof a “the theoretical source of the false reconcilement [of freedom with repressive force].”39 The claim is that, for Kant, both freedom and causality are defined, in the end, in terms of a concept of an intractable law that must be obeyed without question, if only because there is no question that can be intelligibly put to it. Moral law and causal law both turn out to be transcendental givens: to ask why literally makes no sense. In a move that brings to mind Isaiah Berlin’s objection to so-called positive liberty (to the effect that it is a misnomer), Adorno claims that Kantianism thereby underwrites an ideal of blind obedience. “Wherever it is maintained that the substance of freedom is that you are free when you freely accept what you have to accept anyway,” he writes, “you can be sure that the concept of freedom is being abused and is being twisted into its opposite.”40 Referencing East Germany, which he describes as being in the ideological hold of “the traditional values of the petty bourgeoisie,” he continues: “the protestation of freedom now enters into the service of repression, of actual unfreedom.”41 This said, it’s not what Adorno calls the felt facticity of Kantian moral or causal law that is problematic. Adorno is not denying the need for an element of force, of external determination, in either case. When it comes to moral necessity, for example, Adorno would agree that unless we are deeply damaged ourselves, we will be disposed to feel ourselves compelled to respond to the suffering of others. The point with respect to morality is that this fact is not the same as casting the internalized norms of a given social form – including, for better and/or for worse, its historically achieved interest in universality – as a transcendentally given law. (Indeed, Adorno again and again affirms the brute materiality of actual subjects, as well the alterity of our unconscious drives.42) With respect to causal necessity, meanwhile, the problem lies in how Kant has handled the issue. Adorno is very clear that Kant is a Humean in just the sense, or at least very much like the sense, described at the beginning of this chapter. Formulating the Kantian position in epistemic rather than ontological terms, he says “our thoughts [according to Kant] are incapable of telling us anything about what might be called the internal interconnection linking . . . events that succeed one another in time, aside from the form of our act of subsumption.”43 While Adorno does not manage to propose a clear dispositional realist alternative, he does suggest that a metaphysics in which causality is thought to be an ordering principle provided a priori by reason, rather than an expression of the powers of things – such a metaphysics makes it harder than it would otherwise be to regard ourselves as causally efficacious, even if we are consigned to act, as Marx said, in conditions not of our own choosing. Kantian metaphysics configures Adorno’s account of freedom, too, not only in his diagnosis of domination. However, as I’ve said, here it is not Kant’s position

Kant and the Frankfurt School 67 taken at face value that does the work, but rather a fundamental insight about the relationship between subject and object that Adorno takes to be implicit in the theory, even while it is at the same time at odds with the theory. Adorno gets to the unintended dialectical nub of transcendental idealism through a critical analysis of two Kantian claims: (1) that subject and object may be considered in isolation from one another; (2) that subjectivity is metaphysically prior to objectivity. Let’s look at each in turn. On my reconstruction of his thinking, Adorno argues first that subject and object can’t be neatly separated. Each always already presupposes the other. If we begin with the category of pure subjectivity, we find that it breaks down in a number of significant ways. Perhaps the most basic claim is that subjects are embodied. Kant, as we’ve seen, is at pains to distinguish between people considered as empirical subjects and people considered as transcendental subjects. But in the end we are material entities. It may be different with respect to other types of rational being, but subjects who are people are, in an inescapable, bodily manner, also objects. Moreover, it would seem that pure subjectivity, if there were such a thing, would be self-contained, and as such it’s hard to see how it could have any hold on its contrary, pure sensuousness. To be sure, this is a problem that Kant attempts to resolve through numerous means, ranging from the schematism section of the Critique of Pure Reason to the full account of synthetic judgment in the 3rd Critique. Adorno believes that the efforts fail – or, maybe, that precisely insofar as they succeed, what Kant paradoxically shows is that pure subjectivity invariably draws upon something outside of itself. In the case of the Kantian will, for example, Adorno points out that insofar as it is an active force, it is – and must be – conceived as a type of impulse, somatic in nature. This fact renders it objectlike, both in the sense that it is just a given and in the sense that it is a non-rational, sensuous given. The impulse in question, Adorno thinks, is precisely the unconscious, non-rational backdrop to the conscious Freudian ego. And once we have put it in these terms, we will have to add that the well-formed ego depends upon the regulatory function of a super-ego – which, in turn, Adorno thinks, consists of introjected norms of a society that is itself unfree. Conversely, with respect to the idea of pure object-hood, it follows from the fact that all subjects are objects that some objects just are subjects. Further, many objects that are not themselves subjects are nonetheless concretely produced by subjects. As artifacts, they are objectifications of human intentionality. In addition, Kant would have it that phenomenal objects are contingent upon the synthetic operation of the reason in order to be anything at all, other than sensuous manifold. Finally, if we shift from phenomenal objects to Kant’s notion of an un-knowable and absolutely other transcendental object, we have a concept that Adorno believes expresses, at the level of thought, the real alienation of embodied subjects from their material conditions. But if subject and object presuppose one another, the relationship between them is not symmetrical. On this point, at least, Adorno and Kant agree. Unlike Kant, however, Adorno thinks that it is the materiality rather than consciousness that must be granted analytic priority. Of course, Adorno by no means intends to replace

68

Ontology Revisited

transcendental idealism with any sort of reductive materialism. The principle of the priority of the object, to use the famous phrase, tells us only that while all subjects are embodied, not all objects are subjects. But there is a further thought, too – consistent with the assessment that a violence is done to phenomenal objects on the Kantian account – that, contra Kant, the metaphysical identity of objects may not be given by the cognitive faculties of subjects after all. Hence the other famous phrase: “Objects do not go into their concepts without leaving a remainder.” As I’ve already said, I think that Adorno is conflicted about this further thought, or at least that he is unable to see it through. I’ll return to the point below. This model that Adorno comes to negatively, through his critique of Kant – this model of the relationship between subject and object as chicken-and-egg-like, as I put it above (albeit with priority given to the object) – can itself be seen to structure Adorno’s positive view of freedom, such as it is. How this gets spelled out is that Adorno sees freedom as requiring both a universalistic and a particularistic ethical stance, each correcting for the other. On the one hand, Kantian universalism is associated with a principled erasure of the particular, a particularity that is thereby violated. Referring to “the abstract nature of legal and moral systems,” Adorno says “[t]hey cut away everything specific to living human beings and treat them as if they were merely impersonal parties to contracts.”44 On the other hand, attention to universal norms is a check on the equally significant potential violence of seeing only certain people or groups of people as morally deserving. Thus he writes, [a]nyone who like me has had experience of what the world looks like when this element of formal equality is removed – from the legal system, let us say – in favour of specific substantive values that are asserted in an a priori fashion, he will know from his own experience, or at the very least from his own fear, just how much of humane value resides in this concept of the formal.45 Similarly: while, on the one hand, the idea of individual spontaneity must be seen to be an essential element of a defensible concept of freedom (especially, spontaneity construed as the potential for protest vis-à-vis conditions of suffering), on the other hand, all acts, even free ones, have to be understood as conditioned, in that they are always embedded in and mediated through given socio-historical contexts. One might be tempted to liken Adorno’s position to Marcuse’s. That is, like Marcuse, Adorno is suspicious of pure practical reason as repressive. And Adorno wants a freedom that retains the best, in a sense, of each side of the subject–object relation, as does Marcuse. But Marcuse, I think, sees a way to reconcile the relata, while Adorno does not. For Adorno, there is nothing to do but use each term to counter-act the logic of the other. This difference is reflected in the fact that Marcuse sees subject and object as properly reconciled in Kantian aesthetic judgment, while Adorno sees Kant’s efforts in this regard as aporetic. The difference can also be seen, I think, in the small detail that is the contrast

Kant and the Frankfurt School 69 between Marcuse’s passing affirmation of Schiller and Adorno’s criticism of him as “Kant’s popularizer.”46 Having undertaken to show that Adorno’s thinking is, as Adorno would be the first to admit, entangled with Kantian metaphysics all the way down, I now want to ask if Adorno’s over-all position holds together. Does he have the metaphysics he needs, in order to say what he wants to say, even about Kant? As I have intimated throughout, I think that the answer is no, not quite. The problem is this: unlike Marcuse, Adorno sees that Kantian objects do not have form on their own, and so cannot be left to freely be themselves, to use Marcuse’s locution.47 At the risk of sounding like a broken record, let me reiterate one last time that the Kantian concept of the transcendental object, like the Aristotelian idea of prime matter, refers only to the fact of materiality. It doesn’t refer to any actual entity. On the Aristotelian account, any actual substance is matter that is enformed. And this is so for the Kantian as well. But for the Aristotelian, objects that are not artifacts have whatever form they do independently of subjects – and even artifacts are made from materials that have their own form. For the Kantian, by contrast, the form is provided by the subject. Adorno sees very clearly that this is the Kantian position, and I think that it’s also clear that he objects to it. I don’t mean by this that he thinks that the position expresses true information about an objectionable social order, although he does think this. I mean that he thinks that it is false. Here is a lengthy but crucial passage on the topic, from a lecture given in June of 1959: This absolute absence of determination . . . means that there is no real resistance to this act of subsumption – that is, this synthesizing of my ideas with the aid of the understanding, of the forms of my consciousness – on the part of whatever is synthesized. This is because whatever is being synthesized is so lacking in qualities that it is infinitely malleable by the forms of consciousness. The absence of qualities, the indeterminate nature of the material of knowledge, confers on it a kind of plasticity . . . that allows us to synthesize these givens without their placing any obstacles in our path. This means that our minds can manipulate them at will. They do not need to have any particular taste; it is enough if there is something to bite on. . . . As long as I have something between my teeth, my sovereign will can do with it whatever I please.48 But the question is: what is he in a position to do about it? If the principle of constitutive subjectivity is bad metaphysics, can Adorno say why? It’s hard to see how he can. Or at least, it’s hard to see how he can say anything more than he does say, and what he does say is not enough to make the case. Specifically, Adorno is limited in his ability to carry through a critique of constitutive subjectivity because his own continuing attachment to Kantianism commits him to an ontology that bottoms out in a transcendental object, rather than in substances with intrinsic forms. He can remind us that materiality isn’t the same as conceptuality, and he can enjoin us to play the categories off of one another.

70

Ontology Revisited

But the real problem with Kant is not that he mistakes society for a transcendental subject, though he does do that. The real problem is that all he can contribute to materialism is the claim that experience includes something that isn’t thought. And it’s not at all clear that Adorno can get any further than this. Adorno can’t give us objects that have their own identities – or even their own unity as material entities as such – any more than Kant can. Human beings might be an exception. As I suggested earlier, it would seem that in the case of people, the Kantian – or at any rate a Kantian such as Adorno – will have to permit us an intrinsic form, even for people considered qua phenomenal objects. And if this is so, then maybe, in the case of objects who are people, Adorno can press beyond Kant. Here, it would seem, to do justice to the priority of the object requires only the very concern for particularity that Adorno does profess. But even so it’s not as straight-forward as it might appear. For one thing, in the context of transcendental idealism – or, if we prefer, a Kantian materialism – in such a context even an object’s particular features are brought to it, in an act of metaphysical predication we might say, by the subject. Raw sensuous manifold doesn’t give us accidental properties any more than it gives us essential ones. This fact seriously limits Adorno’s ability to advance the category of “particularity” as an ontological counter to the violence of constitutive subjectivity. Note that celebrating the pursuit of particularistic rather than universal knowledge-claims as an epistemic goal, which is Adorno’s preferred strategy, is a moot point relative to the task at hand. Knowledge of particulars does no ontological work on behalf of a challenge to transcendental idealism. On the other end of the continuum, meanwhile, there is the problem of individuals coming to exist as material entities at all, on the ontological basis of a transcendental object alone (i.e., on the basis of the bare assertion of the existence of “that which isn’t conceptual”). There is simply not enough in the ontology to sustain anything like self-subsistent material entities. Not being able to point to anything other than a transcendental object doesn’t just limit Adorno’s ability to get beyond Kant. It undermines the critique itself. Why? One reason is that without objects the metaphysical integrity of which is violated by the principle of constitutive subjectivity, it is hard to see why there is reason for concern. If there is nothing over which to run rough-shod, the problem would seem to dissolve. We are left with – what? – the complaint that Kant should have stopped dithering around with a dualism that he knew would fail, and should have gone ahead and attributed transcendental subjectivity to empirical subjects. This may be a fair assessment of Kant, but the original point remains: if there is “a something non-conceptual” that we are meant to care about, it will have to be made out of something more independently determinate than a transcendental object. Adorno’s objection to Kant’s account of causality is similarly undermined by his ultimate unwillingness to break with Kant. As noted above, Adorno sees that Kant, following Hume, has not grounded causality in any features of phenomenal objects. “This so-called external nature of causality [in Kant’s thinking] . . . arises from the fact that causality does not reside in things themselves. It is, rather, an

Kant and the Frankfurt School 71 ordering principle according to which the subject combines successive states with each other.”49 But while he accuses Kant of emptying out causation of all causal force – “as he cures it of naturalistic prejudice, it dissolves in his hands”50 – he poses no alternative. As above, one might reply on that Adorno’s behalf that his concern is not to engage in abstruse metaphysical debate, but rather to show that Kant’s Humeanism represents in ideational form the fact that the causal mechanisms of capitalist society do indeed lie beyond the reach of subjects, and do indeed appear to us in the guise of an externally imposed necessity. My response – as above – will be that while Adorno does offer such a diagnosis, it verges upon disingenuous to suggest that he does not think that Kant’s account of causation is false. This said, the point is not so much that Adorno doesn’t offer an alternative; the point is that he can’t. He can’t because he lacks the metaphysics to do so. Just as an inchoate sensuous manifold doesn’t add up to entities, it doesn’t add up to entities that can do things. It’s in the nature of the case that my discussion of Adorno has been extremely complex. Let me take a step back and summarize the argument, just to be sure that it is clear. To begin, I said, we can see that Kant’s stated metaphysics figures into Adorno’s account of existing relations of domination. Moving out a frame, I argued that we can see that the internal relationship between subject and object that Adorno believes Kant’s stated metaphysics implies – we can see that this corrected-for, dialectically self-conscious Kantian metaphysics provides a template for Adorno’s thinking about freedom. Both levels of analysis, I suggested, serve to illustrate the way in which Kantian metaphysical categories play a role in Adorno’s social and political thought. Next I claimed that the corrected, Kantianderived metaphysics (which I believe Adorno accepts, if only resignedly so) cannot sustain either Adorno’s critique of domination or his appeal to particularity on behalf of the priority of the object. This I consider to be a second-order illustration of the significance of metaphysics for social and political thought. By way of conclusion, let me also review the over-all analysis that I have been developing over the course of the last three chapters. I began with Hume, whose positions at the level of social and political thought I take to be more or less consistently Humean. My aim in doing so was two-fold. I wanted to show what a social-political theory without powers or necessity or essences looks like. At the same time, I wanted to flag difficulties that even Hume can be found to have, in virtue of his Humeanism, in talking about agency. In Chapter 3 I went a step further, undertaking to show that in addition to problems that proponents of Humeanism have in accounting for agency as such, there are also specific ways of thinking about human beings that cannot be countenanced from the perspective of a Humean metaphysics. Thus Mill, I argued, who advances the view that we are the bearers of distinctively human capacities, which may or may not be cultivated or expressed but ought to be, is hampered in his ability to properly articulate and adequately defend his neo-Aristotelian sensibilities because they are at odds with the Humean ontology to which he is also committed. Turning to Kant next, I undertook to show how the injection of necessity into the Humean picture creates its own problem for agency, one that has to do with freedom in the context of

72

Ontology Revisited

determinism, rather than activity in the absence of powers. Kant recognized the problem, whereas Mill did not. In his effort to resolve it, to reconcile freedom and determination, Kant issued in the terms of the contemporary debate over free will: on the one side, the absolute determinism of causal law; on the other side, freedom construed as spontaneity, as escape from the rubric of event-causation. I then pointed to Marcuse, Horkheimer and Adorno as thinkers whose social-political thought is intimately connected with the Kantian metaphysics in question. Having filled out the picture so that we can appreciate the lay of the conceptual land, I now want to look more closely at the contemporary analytic debate surrounding free will, mental causation and agency. Specifically, I’ll suggest in Chapter 5 that once we see how the debate is an artifact of Humeanism, the often odd-man-out-position known as agent-causation appears in a new and far more plausible light.

5

Agents, Powers and Events Humeanism and the Free Will Debate

I want to set aside the history of political thought for a moment, and turn to contemporary analytic philosophy. Central to the theorizing of agency in this quarter is the question of free will – whether or not we have such a thing, and, if we do, what it is and how it works. My contention is that here, too, the debate has been shaped by Humean assumptions. If I am right that it has been, it will be a point worth taking seriously, since we have already seen that Humeanism makes it difficult – if not, technically speaking, impossible – to talk about anything actually doing anything, freely or not, let alone about selves, or persons, doing things of their own accord. The argument proceeds as follows. I begin with an overview of the free will problematic, intended to display the Humean structure of the configuration as a whole. I then look more carefully at how the implicit Humean metaphysics undercuts the position known as “agent causation.”

The Free Will Problematic Disagreement about whether or not we are metaphysically free, and, if we think that we are, uncertainty about how to square such a belief with our other beliefs, pre-dates the modern period. But the arguments in contemporary analytic philosophy center on a specific iteration of the problem, one that does not hang on the idea of fate or Divine predestination. On the one hand, it is presumed, we know that events in the physical world are governed by laws. These, everyone agrees, are the kind of laws that have been revealed by our best scientific practice. To deny that the material world is so ordered would be to reject as false the very foundations of our most venerated body of empirical knowledge. On the other hand, it is also taken as a given that we think of ourselves as being self-determining in the sense of getting to decide if, when and how we wish to act. But if the science is true, then it would seem that the action of human beings, no less than any other event, is a function of the same laws that determine everything else – leaving no room for the sui generis activity that we associate with agency. In recent years, the problematic has been amended so as to incorporate the observation that if the laws of nature turn out to be probabilistic, this will not help matters. For while probabilistic laws may loosen the grip of necessitation, they do not appear to do so in a way that would

74

Ontology Revisited

return control to the agent. Instead, the agent’s act looks to be merely a locus of statistical likelihood.1 The problem field defined in these terms yields an agenda-setting set of philosophical questions, and a limited number of available moves that may be made in response to them. The first question is whether or not determinism is true. Is the world such that everything that can and will happen is, at all points in time, already fixed? Determinists say yes; indeterminists say no. Second, suppose either that determinism or that the existence of free will obtains. If either is so, would such a fact be consistent with the other also being so? Compatibilists say yes. Incompatibilists (be they determinists or indeterminists) say no: free will and determinism are contraries. Third, assume that incompatibilism is true, and also that we do have free will. Just how is free will to be conceptualized, from an incompatibilist and interderminist – or, as it is also called, a libertarian – point of view? Especially, if one thinks that action is unconstrained by deterministic laws of nature, what if anything keeps free acts from being purely random? This is an issue that arises most naturally when we want to assign credit or blame for an action. Proponents of agent-causation contend that while actions may not be caused by laws of nature, they are nonetheless caused by agents – for which reason they are not arbitrary happenings. Now, the first thing to notice about the contemporary free will problematic is that it does not, in fact, capture a theoretically naïve dichotomy between subjective experience and something called “science.” Hume, for example, did not think that the deliverances of science implied any sort of deterministic necessitation at all. Or at least, if they did, it was a determinism that amounted to nothing other than a so-called “necessity” of habit. Indeed, it was for just this reason that Kant found the Enquiry to be so disquieting. If “science” is now assumed to secure the thesis of determinism, this fact is an achievement of modern philosophy, not modern physics. Moreover, the position that authorizes said thesis is readily identifiable: it’s a modally enhanced Humeanism, involving unbroken chains of events, governed by invariant causal law.2 Kant and Mill give us its paradigmatic formulations, historically. In Kant’s version, there is a divide between deterministically related phenomenal events on the one side, and autonomous acts of free will on the other. Mill has it that free acts are free despite being fully determined by prior causes – which causes form a totality of causally connected events at all given points in time. Incompatibilists opt implicitly for the Kantian approach. Compatibilists are often empiricists, though in terms of the issue at hand we could just as well say that they confine themselves to the phenomenal side of the Kantian dichotomy. The second thing to notice about the contemporary free will problematic, then, is that it involves precisely this bit of Humean infrastructure. When those involved in the debate take sides on the question of determinism vs. indeterminism, the determinism that is invoked is the nomological necessity of law-governed causal sequences. When parties to the discussion ask if determinism and free will are compatible, it is this same order of “event causation,” as it’s called, with which free will is thought to be reconcilable or not. It is no surprise, then, that the issues that dog incompatibilism should be specifically Kantian. In particular, when the

Agents, Powers and Events 75 libertarian conception of free will seems to its critics (and sometimes to its defenders as well) to have no causal traction, no way of connecting agents to the events that involve them, it’s because the default underlying Kantian-derived notion is precisely one of a-causal spontaneity. I shall turn to the challenge facing libertarians in a moment, but let me first offer a sampling of passages from the literature, to illustrate the point that the terms of the contemporary free will problematic are neither timeless nor pre-theoretical, but rather recognizably Humean, in the ways that I’ve just said. Roderick Chisholm, for example, opens “Human Freedom and the Self” as follows: 1. The metaphysical problem of human freedom might be summarized in the following way: Human beings are responsible agents; but this fact appears to conflict with a deterministic view of human action (the view that every event that is involved in an act is caused by some other event); and it also appears to conflict with an indeterministic view of human action (the view that the act, or some event that is essential to the act, is not caused at all).3 To put it differently, the metaphysical problem of human freedom is that it appears to be inconsistent with the alternatives permitted by Kant. In the context of defending what he calls a “reasonable libertarianism,” David Wiggins writes: If determinism is true, and if every action of every agent in its particular circumstances really is fixed in respect of its occurrence, in respect of its mode of occurrence, and in respect of the characteristics upon which its character as an action supervenes, and fixed in this way by some antecedent physical condition; then obviously actions cannot be torn free from the nexus of physical effects and fully determining causes.4 He then continues in a footnote: It should be added that even if we adopted the mysterious view that mental events do not occur in the physical world, still, if physical determinism were true, bodily movements would fall within its ambit and the autonomy of the mental would be limited to mental events with no proximate physical cause and no practical (actual) outcome.5 And here’s Kai Nielson, in the course of defending a compatibilist position: ‘Freedom’ is to be contrasted with ‘constraint’ or with ‘compulsion’ rather than with ‘determinism,’ for if we try to contrast it with ‘determinism,’ it is far from clear that we get an intelligible contrast. It is not clear what sort of an action we would count as a ‘causeless action.’6 This same flavor of Humeanism finds its way into Peter Van Inwagen’s classic “Consequence Argument” too, notwithstanding his remark that “the reader will

76

Ontology Revisited

note that the horrible little word ‘cause’ does not appear” in his definition of determinism as “the thesis that, given the past and the laws of nature, there is only one possible future.”7 Van Inwagen summarizes the Consequence Argument as follows: If determinism is true, then our acts are the consequence of the laws of nature and events in the remote past. But it is not up to us what went on before we were born, and neither is it up to us what the laws of nature are. Therefore, the consequences of those things (including our present acts) are not up to us.8 Finally, consider just the title of David Lewis’ 1981 defense of compatibilism: “Are We Free to Break the Laws?”9 That Humeanism serves as the default, or presumptive metaphysics of the contemporary free will debate seems to me to be an entirely uncontroversial claim.10 Nor should it come as any great surprise, given that Humeanism is the dominant ontology in Anglo-analytic circles. The important question is whether or not it matters – whether or not the implicit metaphysics affects what can and cannot be said, or said well. I believe that it does. In particular, it makes it much harder than it ought to be to defend the position known as agent causation. This is what I shall now try to show, first by sorting through those concerns that are primarily related to causation, then by reflecting upon issues connected to the nature of agents as entities.

Agent Causation and Causality Agent causation is the view that agents, rather than antecedent events plus the laws of nature, are the causes of what agents do. Well-known analytic agent causal theorists include Roderick Chisholm, Richard Taylor, Timothy O’Connor and Jonathan Lowe. Before I say anything further about the position, however, let me specify more precisely the Humean parameters that define the conceptual space, relative to the issue of causation, in which contemporary agent causation theories must be defended. By my count, there are five key principles. I shall mainly make reference to the first, second and fifth, but the others are worth stating. Agent causal theorists are not required to accept any of these precepts, but neither are they at liberty to simply dismiss or ignore them; as norms they are coercive along the lines of a Durkheimian “social fact.” The first assumption, which is the most basic, is that causation does not involve the expression of powers. This is the crux of the matter. The second assumption is also crucial, however. It is that events, and events only, may be causes. In one sense, this assumption is even more restrictive than the first, since it disallows agents as causes altogether. (Admittedly, this fact alone renders obvious the thesis that Humeanism undercuts the defense of agent causation; it does not, however, render the thesis trivial.11) The third assumption is that causation is transitive: if A causes B; and B causes C; then A causes C. Note that this assumption will hold most intuitively when it is thought that B must follow A, and that C must follow B (in order for A to cause C by implication from

Agents, Powers and Events 77 its causing B).12 The fourth assumption follows from the third. It is that, as Mill observed, there is – or at least would seem to be – no non-arbitrary distinction to be drawn between conditions and causes. Both are events related lawfully to a given outcome; conditions are just causes that are one or more steps removed. The fifth assumption, finally, is that the relationships between events must be either (a) caused via laws (deterministic or probabilistic); or (b) uncaused, and therefore random. Agent causation is related to the conceptual field constituted by these assumptions in two different ways. Historically, the position is part and parcel of the wholesale critique of Humeanism advanced by Thomas Reid. For this reason, the phrase agent causation is often treated as a Reidian term of art. Reid took it to be self-evident that, as he put it, “[e]verything that begins to exist, must have a cause to its existence, which had power to give it existence. And everything that undergoes any change, must have some cause of that change.”13 The power to create and to effect change, Reid called “active power.”14 To be a cause, he thought, is not, as Hume had argued, to be that which comes first in a constantly conjoined pair, but rather to be the bearer of active power. “That which produces a change by the exertion of its power,” he says, “we call the cause of that change; and the change produced, the effect of that change.”15 As a matter of epistemology, Reid held that we acquire the concept of an active power – and, thereby, of causation itself – through proprioception, i.e., via the felt experience of our own bodily efficacy. But it’s the ontology that’s pertinent. Reid’s position was that to be a cause is, in the manner in which I’ve just described, to be an agent – an originator of change. In fact, he uses the terms “cause” and “agent” interchangeably: [t]he name of a cause and of an agent, is properly given to that being only, which, by its active power, produced some change in itself, or in some other being . . . Active power, therefore, is a quality in the cause, which enables it to produce the effect. And the exertion of that active power in producing the effect, is called action, agency, efficiency.16 Reid further held that “[p]ower to produce any effect, implies power not to produce it.”17 What follows from this, as he saw it, is not only that intentional beings may be causes (and necessarily freely acting ones at that), but also that intentional beings are the only real causes, since it is only intentional beings who can refrain from exercising their powers. Especially, neither purely physical phenomena nor laws of nature are the kind of entity that does or could bear agential abilities of the requisite type. The causes of physical processes that are not initiated by persons are, properly speaking, a mystery, Reid claimed – though the suggestion from him was that they are caused by God. But Reid did not win out, historically. Hume did. And the debate over free will has therefore come to be carried out in the Humean terms that I’ve described above. Vis-à-vis the debate so configured, agent causation enters the picture not primarily as part of a comprehensive rejection of Hume, as it was for Reid, but

78

Ontology Revisited

rather as a solution to a specifically Kantian problem faced by the libertarian, viz., that acts of free will appear, in Kant’s version of the Humean model, to be entirely random, and not determined by antecedent events. The “intelligibility problem,” as it’s called, arises in the Humean context because, from a Humean perspective, to be caused just is to follow upon antecedent conditions in a law-like fashion. Not to do so – as acts of free will manifestly don’t, according to the libertarian – is, ipso facto, to be uncaused. The proponent of agent causation responds by saying that free acts are caused by agents, if not by events. It’s helpful to be able to keep separate a full-blown anti-Humeanism and the more limited appropriation of Reid for the purpose of establishing the non-random character of agency as construed by libertarians. Tim O’Connor and Jonathan Lowe stand out as agent causal theorists who mean to reject Humeanism altogether, with Lowe managing to do so most thoroughly, in my view. I shall return to this point in the next section. The question that I want to pose, then, is this: how does Humeanism affect the reception of the idea that agents are the causes of what they do? What objections does the agent causal theorist need to meet that she otherwise wouldn’t? What moves is she pressured to make? Let me begin with the obvious. From a Humean perspective, there are two glaring problems with the agent causal approach, in relation to the issue of causality. The first is that the approach brings with it the idea that to be a cause, at least when it comes to agents, is to be the bearer of a power – an ability to do, and not just to be.18 The invoking of powers is a deal-breaker, in a Humean environment. The proponent of agent causation who is operating within a Humean context will therefore have two options, if she chooses not to dismiss Humeanism outright. One option will be to say that it turns out that there are two different kinds of causality, only one of which involves powers. The other option will be to insist that agents are causes, but to side with Hume and say that causation is not dispositional, not even in the case of agents. This points to a second glaring problem, however, which is that agent causation is a theory according to which it is substances (i.e., agents), and not events, that are causes. Perhaps unsurprisingly, it is the dispute over substances versus events as causal relata that tends to be the focal point for extended argument, with less attention being paid to the deeper question of whether or not causality involves the exercise or display of powers. The response in a Humean context that one would have expected, vis-à-vis the agent causal theorist’s natural appeal to dispositions, is that the position must be false, since it relies so heavily upon a meaningless term. But the complaints tend to be more nebulous than this. Often it is simply taken as selfevident that there is something improbable, even fanciful about the view. Or there may be an ostensible concern over the proliferation of kinds of causality. Two is too many: a theory that calls for an additional kind of causality on top of the real kind is ad hoc, and inelegant. Or if it is allowed that there are two different kinds of causality, the presumption is that agent causation will have to be shown to be consistent with an order-based model of causation, rather than the other way around. None of this requires the Humean to explicitly claim that agents have no powers because nothing has powers, but the cumulative effect is to negate the thesis of agent causation. And what is important is that there is nothing intuitively

Agents, Powers and Events 79 plausible about any of it. If Humeanism were not the default metaphysics, no one would regard the notion of a causal power as being particularly mysterious. Nor would admitting of causal powers in the case of agency require one to defend the existence of a special type of causation, unique to agents. Nor, finally, would there be a need, in turn, to square a curious, alternate type of causation with a designated standard, or “real” causation. With respect to the problem of agents being cast as causal relata, the Humean charge is that the proponent of agent causation has made a category mistake: it is only events that can enter into causal sequences. There is pressure, therefore, to find a way to translate the idea of an agent having caused something into the idea of an event having caused it. A good indicator of the degree of pressure involved is that even Roderick Chisholm, one of the contemporary thinkers most closely associated with the agent causal view, was responsive to it. In “Agents, Causes, and Events: The Problem of Free Will,” published in 1995, he announced that [i]n my earlier writings on this topic, I had contrasted agent causation with event causation and had suggested that ‘causation by agents’ could not be reduced to ‘causation by events.’ I now believe that that suggestion was a mistake. What I had called agent causation is a subspecies of event causation.19 Admittedly, nothing substantive is decided by the fact that an elderly Chisholm saw fit to revise his position after 30 years. I mean only to illustrate the ubiquity of the assumption that if one can deliver libertarian free will in the form of a sequence of events, one is clearly at a philosophical advantage relative to the proponent of causation by agents. Robert Kane, for example, another leading figure in these debates, advertises the merits of his own position – “teleological intelligibility,” as he calls it – in precisely these terms.20 Agent causal theories, he says, commit adherents to the idea of an “extra (or special) factor . . . over and above the normal flow of events” – viz., a “special kind of ‘agent-’ or ‘immanent’ causation that cannot be explained in terms of ordinary modes of causation by events or occurances.”21 His view, by contrast, invokes only “a notion of mental causation,”22 one in which the subjective states of agents are taken to be “events, occurrences, or states of affairs.” There is no “special kind of nonoccurrent causation by an agent, such as AC [agent causation] theories require.”23 Again, my point is that causation by agents rather than by events only requires special pleading in a context defined by Humeanism. As Reid observed, the idea of agent causation would by all experiential measures seem to be primary. That it has come to seem nearly inconceivably strange is a function of the degree to which Humeanism has shaped contemporary philosophical conversation. The background influence of Humeanism can be seen in other ways too. Randolph Clarke, for example, concludes in “Toward a Credible Agent-Causal Account of Free Will” that action is a product of agent causation and indeterministic event causation.24 Agent causation, on his view, tells us who did x; antecedent conditions tell us why x was done rather than z, and why it was done when it was.

80

Ontology Revisited

Clarke treats causation as only one kind of phenomenon, regardless of whether a given cause is an agent or an event, but his model for it is ultimately Humean (minus the stipulation that causal relata must be events), on my definition of Humeanism. Citing Tooley (but also Armstrong), he suggests that [a]n event causes another just in case the relation of causation obtains between them. Two events can be so related only if they possess (or are constituted by) properties that are in turn related under a law of nature. Ultimately, then, causal relations are grounded in laws of nature, which consist of second-order relations among universals.25 It is the same, he says, when the relata are an agent and her action. Finally, even Tim O’Connor, who believes that all causation is powers-based, and that intentional acts are caused by agents alone, nevertheless preserves the category of event causation (albeit now construed in non-Humean terms) for causal relations involving material phenomena that are not agents. I want to be sure that I am not misunderstood here: nothing in what I’ve said shows any of the referenced positions to be false. Nor do I mean to imply that the philosophers to whom I’ve pointed hold the views that they do purely for reasons of sociology. My goal is simply to draw attention to the extent to which the analytic conversation about agency and free will has been, and continues to be, framed by Humean categories and assumptions. The argument over agent causation looks very different from the perspective of one who is prepared to reject Humeanism altogether. From an unconditionally anti-Humean perspective, two crucial points emerge from reflection upon the contemporary free will problematic. The first is that Humeanism manifestly prohibits the thought that agents do things (as well as the thought that agents, i.e., selves, do things). We already saw in Chapter 2 that for Hume himself, agency amounts only to (a non-unified collection of impressions figuring in) a succession of static states of affairs. I likened agency so construed to the movement of images in an animation flip-book. In the principled rejection of the thesis of agent causation, the implications of the view become clear. In short, the consistent Humean should be baffled by agent causation, as nothing remotely resembling an agent bringing something about is allowed by her metaphysics.26 This is a serious and essential shortcoming of the view; that it has instead come to be regarded as a mark of philosophical seriousness of purpose is remarkable. In any case, for the dispositional realist, it is Humeanism that is defeated by committing its proponents to the view that agents (such as they are, for the Humean) can’t do things, rather than agent causation that fails for being the view that we can. The second point is that free agency is far more readily compatible with causation on a powers-based account of the latter. If one is prepared to say that the real question concerning free will is whether or not the powers of human beings can be squared with the powers of other causal bearers, discovered in the course of our best scientific practice, the answer to the question is yes. This, I contend, is the proper form of compatibilism. Such a compatibilism – one re-conceived along

Agents, Powers and Events 81 dispositional realist lines – would have the following general features: it would be assumed that there is only one type of causality – viz., some version of “the display of a power or powers” – regardless of the type of causal bearer. Of any putative cause, the question would be whether or not it is the kind of thing that could be a bearer of the power(s) ascribed to it. A further assumption would be that different kinds of causal bearers have different kinds of dispositional properties, in virtue of which they can do different kinds of things. The aim of science, one who held such a view would think, is precisely to identify what kinds of powers are potentially expressed by what kind of thing, in what circumstances. As it happens, this is a philosophy of science already in existence. In analytic circles it is referred to as scientific essentialism; amongst social theorists it is the position known as critical realism. Building on the assumption that different kinds of causal bearers are endowed with different kinds of powers, or dispositional properties, the question with respect to human beings, for what I’ll call the Aristotelian compatibilist, would be “Are our actions at least sometimes freely willed and undertaken, and if so, what are the powers that humans would have to have in order for such a thing to be possible?” Aristotle himself thought that agency of this type involved a mix of rational, appetitive and what he called vegetative powers, but developing a positive account is a topic for another day. For now, it will suffice to say that the reconfigured compatibilism that I am describing gives us a way to square the genuine control over action that libertarians see as central to agency with a perfectly viable contemporary philosophy of natural science. Along the way, libertarian-style agency loses its purported mysteriousness without ceasing to be distinctive. Finally, notice that on this model there are no laws that agents must break in order to be free. There are simply many other causal bearers in existence besides individual human beings – ranging from electrons to oxygen molecules to the international banking system. The fact that we have the extraordinary powers that we do does not render us omnipotent. One might object that Aristotelianism compatibilism, as I’m calling it, presupposes not only powers and the thesis that agents can be causes, but also the existence of genuine agents who can be the bearers of the unique powers that constitute free will. And if reductive physicalism is true, then we could have powers and “agent” causation, but no real agents, only physical aggregates. This line of thinking brings me to the second major way that Humeanism undermines talk of agency: not via the metaphysics of causation, but via its anti-realism about the self and its tendency towards reductionism. It is to this second consideration that I now turn.

Agent Causation and Agents Physicalism Agent causation is the idea that agents can be causes. But it carries along with it a further idea, which is that it’s agents who are causes, when they are, and not

82

Ontology Revisited

somehow-animate aggregates of physical parts (à la zombies). This aspect of agent causation – the presumption of the existence of individuals, or selves, and not just bodies – is also made harder to defend in a Humean environment than it might otherwise be. We should not be surprised by this, given Hume’s own views concerning selves and material substances alike. To be sure, Kant’s critique of Hume should give one who is attracted to Hume’s account pause: it is not at all clear that Hume can explain how the discrete phenomenal bits that he takes to be basic cohere into a material substance that is a sentient being. Nevertheless, at least for those Humeans who are empiricists, anyway, it’s the posited existence of agents as ontologically irreducible entities that is taken to be philosophically problematic, not their own inability to account for them. Resistance to the metaphysical integrity of agents plays out in debates over what is called “mental causation.” Mental causation refers to the fact that beliefs, perceptions and/or other sorts of subjective states of consciousness seem, phenomenologically, to be causes of action. The question in the analytic literature is whether or not this experience is illusory: maybe what we take to be mental events are really just physical events. In that case, what we normally think of as mental causation would actually be physical causation. One of the best-known proponents of the view that mental causes are not causes in their own right is Jaegwon Kim. Now, the agent causal theorist might be inclined to agree with Kim on the grounds that she rejects event causation generally. But Kim’s analysis is relevant because he doesn’t just challenge the idea that mental events can be causes in their own right. Rather, his primary aim is to defend physicalism. If his argument is sound, then agent causal theorists will be left with agents who are bodies, at best, not persons. Kim’s argument is directed toward a position called non-reductive physicalism. Physicalism is the view that only physical phenomena exist, properly speaking. By “properly speaking” I mean to signal the convention in analytic philosophy, mentioned in Chapter 2, according to which if one thinks that x’s are actually y’s, and one believes in the existence of y’s, then one is permitted to say that one believes in the existence of x’s. Such a move can function to obfuscate when the question at hand is whether or not x’s and y’s both exist (here mental phenomena and physical phenomena, respectively) – or if, instead, it’s only y’s that exist. Physicalists believe that only y’s exist: if there is any sense in which mental phenomena exist, it is only because, and insofar as, they reduce to physical phenomena. Social and political theorists call such an ontology materialism – more precisely, crude materialism. Non-reductive physicalists hold that a commitment to physicalism can somehow be rendered consistent with a belief in the existence of mental phenomena that do not reduce to physical phenomena. Nonreductive physicalism is therefore not consistent with reductive physicalism; this is so notwithstanding the fact that non-reductive physicalists believe themselves to be physicalists. At first blush, one might take non-reductive physicalism to be the analogue in analytic philosophy of the materialism of a sophisticated Marxism or critical theory. But I think that this is not quite right, as the non-reductive physicalist

Agents, Powers and Events 83 wants to preserve the mental without rejecting a crude, Cartesian-derived materialism, whereas the Western Marxist or critical theorist has rejected – rather than attempted to augment – such a materialism. In the analytic context, the motivation for trying to make a non-reductive physicalism work is to be able to say that reasons are causes of action. To be able to say it and mean it, that is. Acting for reasons is an important idea for compatibilists, and not just for libertarians, since – separate from the mind/body issue – intentionality would seem to be what distinguishes voluntary action from involuntary action, and compatibilists often define free will as being able to engage in voluntary action, be one’s volitions determined or not. Kim has multiple targets. One is a version of non-reductive physicalism that he attributes to Donald Davidson; another is a version that he associates with Hilary Putnam, connected to the issue of multiple realizability, as it’s called; yet another is a version that hangs on the concept of emergence.27 I’m concerned with the first and the third. I should note, too, that I am concerned with Kim’s criticism of Davidson only because it has been influential. For the purposes of my own argument, nothing hangs on whether or not Kim has read Davidson correctly, or whether or not his assessment of Davidson’s position is sound.28 My claim is only that his criticism trades on Humean assumptions. Remove them, and it is far less compelling. And let me be clear: my goal in advancing such a claim is not to defend a marriage of crude materialism and a causal role for mental phenomena. Rather, my aim is to demonstrate that in showing such a union to be impossible, if he has, Kim has not thereby ruled out a neo-Aristotelian alternative to physicalist reductions of the mental to the physical – or, ultimately, of agents to their bodies. Let me set out Kim’s first and third targets – viz., Davidson’s defense of what he calls “the anomalism of the mental” and the issue of “downward causation” implied by emergent phenomena – then consider the criticism that he levels at both. Davidson’s argument is curious, in that he simply assumes that mental events do, indeed, cause physical events, and vice versa. He adds to this assumption two more: that “where there is causality, there must be a law: events related as cause and effect fall under strict deterministic laws”;29 and that there are no such laws governing mental events. Davidson calls these three assumptions the Principle of Causal Interaction; the Principle of the Nomological Character of Causality; and the Anomalism of the Mental, respectively. The task for someone who wants to defend mental causation, he says, is to show how it can be that these three assumptions are consistent with one another, since the anomalism of the mental would seem to conflict with the nomological character of causality, rendering reasons non-causal by implication. Davidson’s solution is to say that while there is a real sense in which mental phenomena are unlike physical phenomena (or, phenomena under a physical description), there is an equally real sense in which mental events are identical to physical events. How so? One way to respond would be to emphasize Davidson’s debt to Spinoza, who thought that that which exists most fundamentally is material and mental alike. If one were to answer via the letter of the text, however, one

84

Ontology Revisited

would say that mental events, for Davidson, admit of being described in physical terms. And when, or insofar as, they do come under such a description, they are thereby eligible to be governed by the strict causal laws that hold between physical phenomena. Reasons may be causes notwithstanding the Anomalism of the Mental, then, says Davidson, because they are such that they may be cast as physical events, without thereby being having been reduced to physical events.30 Kim is also concerned with the concept of emergence. Emergence is perhaps a more straight-forward way to defend non-reductive physicalism. Kim defines the idea as a combination of the following “doctrines,” as he puts it: 1 2 3

[Ontological physicalism] All that exists in the spacetime world are the basic particles recognized in physics and their aggregates. . . . [Property emergence] When aggregates of material particles attain an appropriate level of structural complexity (‘relatednesss’), genuinely novel properties emerge to characterize these structured systems. . . . [The irreducibility of the emergents] Emergent properties are irreducible to, and unpredictable from, the lower-level phenomena from which they emerge.31

It’s worth noting that there are proponents of emergence, especially amongst those who do work outside of analytic philosophy, who would reject Kim’s doctrine #1 (i.e., atomistic physicalism) – and who might have reservations about #3 as well, in that it puts an epistemic spin on the position, which might be thought to be a distraction. In the history of social and political thought, in particular, emergence tends to be associated with a non-atomistic fundamental ontology, and as often as not with a non-physicalist one. This said, Kim is considering emergence as a strategy for physicalists; in such a context it may be reasonable to assume a commitment to physicalism as partially constitutive of the position. In any case, all emergentists will agree that emergent properties presuppose, but are not equivalent to, an underlying base. And the idea is that once they’ve emerged, they can be causally potent. Kim thinks that non-reductive physicalism is unstable, to use his adjective. In particular, he claims that neither Davidsonian nor emergence-based versions of the position can sustain a suitably defined causal role for the mental. The argument is as follows: assuming that all (sufficiently caused) physical events have a sufficient physical cause (which principle, called the “causal closure of the physical,” Kim presumes that all physicalists accept), then any physical event that has a mental cause is going to end up having to have two causes, the mental one and a physical one. Moreover, it looks as though the mental cause isn’t going to get to do any of the real work. On the Davidsonian model, as Kim interprets it, the problem jumps right out, once one is looking for it. If it is precisely in virtue of their identity with physical events that mental events are efficacious, what exactly would their mentality be contributing to the causal relationship? Things are similar with the emergence model, though not as pronounced.

Agents, Powers and Events 85 There, it seems that the material base of a given emergent mental phenomenon will be sufficient, on its own, to bring about whatever it is that is ostensibly caused by the emergent phenomenon. If the effect in question is a new physical state, it looks as though the “original” physical state should have been able to bring it about directly. If the effect is a new emergent phenomenon, it looks as though the prior physical state is still the sufficient cause: it will just have brought about the effect indirectly, via causing the new physical state that undergirds the new emergent phenomenon. Kim sees no viable resolution to the problem. In order to preserve the mental, he infers, one will either have to (a) maintain that actions really do have two, independent, equally efficacious kinds of causes, one mental and one physical; or (b) reject the principle of the causal closure of the physical, and say that actions that are caused by mental events are not, in fact, sufficiently caused by physical events. Option (a) would seem to issue in too many separately acting causes per given effect. (Analytics call this the problem of over-determination.) Option (b), meanwhile, amounts to abandoning physicalism. The only alternative to physicalism, Kim reminds us, is Cartesian interactionism – the view that there are both purely physical phenomena and purely mental phenomena, and that a single causal chain can consist of both. Physicalism is the better bet, Kim thinks. And if one is indeed going to opt for physicalism over Cartesian dualism, as physicalists manifestly do, then – unless Kim has made a mistake – one will have to be a reductive, rather than a non-reductive physicalist. As I’ve said, I’m interested in Kim’s argument not because I want to defend non-reductive physicalism, but because Kim’s argument lends credence to the idea that individuals are equivalent to bodies, which in turn denies agent causal theorists philosophical access to entities that are recognizable as agents. My contention is that Humeanism makes all of this easier than it would otherwise be. It’s not that Kim has made a mistake. Rather, it’s that his position is as seemingly decisive as it is only because Aristotelianism is off the table. With respect to the charge of over-determination, the troublesome and purportedly epiphenomenal “extra” causes are conceived in Cartesian terms as free-standing, purely mental phenomena. Conversely, the physical domain the causal closure of which is potentially called into question is taken to consist exclusively of phenomena that are not just physical, but purely physical. In keeping with this dichotomy, Kim explicitly says that reductive physicalism and Cartesian dualism exhaust the available alternatives to non-reductive physicalism (itself conceived as a tethering of the purely mental to the purely physical).32 From an Aristotelian perspective, Kim’s argument packs significantly less of a punch. The Aristotelian does not imagine agents to be purely material entities bearing purely mental properties. Instead, she regards agents as unified, intentional beings. While we are alive, she will say, our bodies are en-formed through and through by the “rational principle,” as Aristotle calls it, in virtue of which our activity is distinctively human.33 This is a way of thinking about mind and body, subject and object, that is deeply unfamiliar to post-Cartesian moderns. The idea is that our powers are the powers of an integrated whole. When I taste, or think, or

86

Ontology Revisited

experience emotions, or move about – in all cases I do these things in a specifically human way, viz., as a sentient, fully embodied agent, endowed with certain extraordinary capacities not had, or had to the same degree, by other entities. Although agency involves the enactment of consciousness, acting is neither a purely physical event, nor a purely physical event combined with a purely mental event. Kim’s objections to non-reductive physicalism don’t apply to such a view. First of all, the Aristotelian doesn’t invite worries about epiphenomenalism (i.e., by claiming that immaterial mental events cause things). She’ll say that agents cause things, even when we act intentionally, on the basis of reasons. Nor will she maintain that there are sufficient physical causes for actions or the results of actions, thereby rendering mental causes superfluous. Not unless by “sufficient physical cause” one just means irreducibly conscious agents. And irreducibly conscious or not, agents don’t cause their mentality any more than water causes its power to dissolve salt. The relationship is categorical, not causal. Agents are conscious, intentional beings; that’s just what it is to be an agent.34 One might respond that Aristotelians can’t have a non-reductive materialism without violating the principle of the causal closure of the physical any more than the Humean physicalist can. But for the Aristotelian, unlike for the post-Cartesian Humean, nothing of consequence follows from doing so. The principle should be rejected regardless, she will say, inasmuch as it is tied to a flawed metaphysics.35 The fate of physics does not hang in the balance, as no one is suggesting that immaterial entities or processes must be incorporated into natural science. Second, it is simply incorrect to say that the only alternative to the causal closure of the physical is Cartesian dualism. After all, non-reductive physicalism just is a form of Cartesian dualism, and reductive physicalism amounts to a Cartesian dualism minus the mental. An analytic philosopher might counter that property dualism (i.e., dualism with respect to physical properties and mental properties) is different from substance dualism (dualism vis-à-vis physical substances and mental substances), and that only the latter counts as Cartesian dualism. The non-analytic will reply that an ontology that consists of purely physical substances and properties, on the one hand, plus purely mental properties on the other, is sufficiently dichotomous to fit the bill. In any case, variations of the Cartesian picture will only appear to exhaust the metaphysical options if Aristotelianism has been misconstrued as itself being a version of Humean non-reductive physicalism, or if Aristotelianism has been overlooked altogether. In reality, Aristotelianism is a perfectly viable nonreductive position for the materialist who rejects the causal closure of the physical to adopt. Emergence I want to conclude by saying just a bit more about the concept of emergence in particular. Just as agent causal theorists are often reluctant to reject eventcausation as the default template for causal relations, Humeanism makes its way into accounts of emergence even amongst those who defend it. Here it’s not the

Agents, Powers and Events 87 account of causation that comes into play, but an underlying atomism.36 The issue is this: what gets to count as an emergent phenomenon, and on what grounds? For classical social and political theorists, the relevant question is whether a whole is something different from a plurality of parts. As noted in Chapter 3, Aristotle refers dismissively to “mere alliances,” which he distinguishes from the “association” that is the polis. Unlike a mere alliance, an association is thought to be an entity in its own right, comprised of but not reducible to individuals. As Aristotle would have it, social life is replete with such entities, ranging from the character friendship to the component relationships of the household, to the household itself – not to mention the final or best association, the polis. And it is this same conception of irreducibly relational entities that Mill rejects, when he says that people, unlike hydrogen and oxygen, do not form the sociological analogue of a water molecule when they are combined. On a concept of emergence such as this, what differentiates an emergent entity from a plurality of parts is the arrangement of the parts. Unlike a mere plurality, the parts of a genuine whole are organized in such a way that the relational entity exhibits new properties, properties not held by the parts singly or by some number of the parts that is greater than one.37 Pluralities, by contrast, have only additive properties, e.g., 5 lbs of sugar weighs more than a cup of sugar. Emergent entities have parts, as does everything except whatever the most fundamental particles turn out to be, but they nevertheless exist irreducibly and without caveat. If one were to make a list of all of the things that exist, the parts of an emergent whole would be on the list, and the whole would be on the list too. As discussed in Chapter 3, to be a holist on this way of thinking is to think that there are indeed such things as relational entities, bearing non-additive properties. Atomists (often called methodological individualists, though the underlying issue is ontological, not methodological) – atomists such as Mill accept neither. Given, says the atomist, that it is only parts that exist (i.e., individual persons), ostensibly relational features of putatively emergent wholes (i.e., sociological properties, borne by sociological entities) are really psychological properties, borne by numbers of individuals. The situation is different in analytic metaphysics. In the analytic context, one is allowed what at least appears to be a third option. The option is this: one is permitted to claim that one believes in the existence of wholes that are not mere pluralities, but to add that they “are nothing over and above their parts, duly arranged,” or “arranged x-wise.” By convention, the “duly arranged” stipulation does just enough metaphysical work to authorize holist-sounding talk vis-à-vis relational phenomena, but not so much that one is thereby committed, ontologically, to the real existence of anything other than parts.38 Thus the ostensible commitment to the existence of the whole is immediately qualified by the countervailing claim to the effect that while the whole does indeed appear to us to exist, it doesn’t exist at the level of fundamental metaphysics. Other versions of the caveat include the suggestion that there are two senses of the word “exist,” one of which doesn’t carry any ontological weight, and/or that the whole is counted as existing as genuine object “only by courtesy.”39 At the level of properties rather than entities, such wholes – wholes

88

Ontology Revisited

that are said to exist, but that are also said to be nothing other than parts – are sometimes called “structural.” An emergent property, on this approach, would have to be something different than a structural property. And the same goes for entities. An emergent entity would have to be something other than a (mere) whole, since wholes, from this perspective, reduce to their parts. An emergent entity would have to be a whole that consists of parts, but that somehow isn’t constituted by any “mere” arrangement thereof. To be sure, not all analytics opt for this third approach. Some are straight-forward atomists, and deny the existence of emergent wholes altogether. Others are simply holists. E. J. Lowe, for example, thinks that phenomena may be emergent, despite not being remarkable relative to normal wholes.40 Some, finally, are holists but want to reserve the designation “emergent” for some special class of otherwise-normal wholes. My contention was that Humeanism seeps back into the discussion of emergence, even when the concept is defended by proponents of agent causation. It does so via the tacit atomism of this apparent third option. A good example of what I have in mind, because it involves thinkers who in the main are committed to a non-Humean program, is the position advanced by Tim O’Connor and Jonathan Jacobs in their piece “Emergent Individuals.”41 O’Connor and Jacobs define emergence via the concept of a non-structural property. As they put it: I am indeed a biological organism, but some of my mental states are instantiations of simple, or non-structural properties. A property is ‘non-structural’ if and only if its instantiation does not even partly consist in the instantiation of a plurality of more basic properties by the entity or its parts. . . . Emergent features are as basic as electric charge now appears to be, just more restricted in the circumstances of their manifestation.42 Emergent states, according to O’Connor and Jacobs, come about when relevant “fundamental particles,” which have given “tendings, present in each microparticle,” act jointly with other micro-particles to produce the instantiation of a simple or non-structural property.43 They claim that they believe it to be an open question what types of emergent states exist. “For all anyone knows on present evidence,” they say, “some perfectly respectable biological and chemical features are ontologically emergent in this way.”44 But, they continue, “we do not think that there is any clear positive reason to suppose so.”45 The situation is otherwise when it comes to consciousness. “Things are different with respect to psychology,” they say.46 Unlike any existent biological, chemical (or, presumably, other physical) states that O’Connor and Jacobs are aware of, mental states clearly do, they believe, fit their criterion for emergence: they are instantiations of simple, or nonstructural properties. The thesis that they defend in the paper is that individuals are emergent entities in virtue of being the bearers of such states. O’Connor and Jacobs call their view “substance emergentism.”47 It may seem outlandish to level a charge of atomism against self-consciously Aristotelian proponents of agent causation who are defending a thesis called substance emergentism. If the assessment is correct, however, it speaks to the hold

Agents, Powers and Events 89 of the reductive, atomist aspect of Humeanism on the contemporary philosophical imagination. Manifestly, O’Connor and Jacobs are not reductionists with respect to individuals. In what sense, then, does their position betray an underlying atomism? The following: except in the case of individuals, they can think of no reason to take wholes to be full-fledged objects, metaphysically. They exist only by courtesy. Why are other wholes not full-fledged objects? Because they are thought to be ontologically reducible to their parts. Why do they reduce to their parts? Because the properties that they bear are themselves wholes (i.e., “structural properties”) that are thought to reduce to their parts. In order to count as emergent, as non-reductively existing in its own right, a relational entity must be a “unity,” to use O’Connor and Jacobs’ term, in a way that regular, not-really-existent wholes are not. And what is it that unifies a whole in the necessary way? Being the bearer of a property that is itself simple. For an atomist, the paradigm of a genuine entity is the fundamental particle, a singular unit bearing a simple non-structural property. O’Connor and Jacobs’ emergent entities are macro-level analogues of such entities. But even if we didn’t know any of the details of O’Connor and Jacobs’ positive argument, we could ask: what would it take for a property or entity to count as emergent, in the context of a presumptively atomist ontology? Being a functionally integrated relational phenomenon wouldn’t do it, since the claim is that structural phenomena are, in the end, just parts. What would the question of emergence even be about, if wholes and pluralities are all just aggregates of parts, in the end? It looks as though it would have to be a question about whether or not atomistic units that are purely physical are the only kind of atomistic unit that exists. And the answer from O’Connor and Jacobs is no: in addition to fundamental physical particles, bearing non-structural physical properties, there are emergent-butontologically-simple48 individuals, bearing non-structural mental properties. The standard analytic atomist will want to add that the parts to which she contends all wholes reduce (and that O’Connor and Jacobs believe structural phenomena reduce) must be “duly arranged.” She will insist on it, in fact, since arrangement is necessary in order for a relational phenomenon to exist at all. And given that this is so, i.e., given that it is agreed that arrangement plays a metaphysically constitutive role, one might be tempted to say that the defender of structural properties and/or relational entities can’t be an atomist. After all, atomism is an ontology that bottoms out in parts. Structural properties and relational entities are not equal to their parts, not even – it would seem – for the type of atomist in question. Rather, they are equal to their parts duly arranged. But again, the permitted move is to say that arrangement, while constitutive, does not count ontologically. Set aside the question of whether or not the position is persuasive. The point that I want to underscore is that all that genuinely exists on this view are parts: single parts and pluralities of parts. To hold such a view just is to be an atomist. For if atomism is anything, it is the belief that what exists – at the level of ontological inventory – are simple, discrete, singular bits. From a holist perspective – or, at least, from my holist perspective – this approach conflates the question of emergence with the question of whether or not mental properties are of the same kind as physical properties. The holist will first

90

Ontology Revisited

want to determine whether or not a putative emergent whole is a genuinely new entity, with properties not had by its parts considered singly or as a mere plurality. If it is, she will then go on to determine what kind of entity it is, bearing what kinds of properties. Is it animate or inanimate? Biological, embodiedpsychological or social? One might object that if persons and trees alike – let alone chairs or computers – are thought to be emergent, then there is no way to block reductive physicalism, since there would seem to be no novelty associated with the emergence of persons.49 But the demand that emergence itself be a novel phenomenon is misplaced. It is the power of consciousness that is novel, not the emergence of non-reducible wholes. If there is an argument to be had with a physicalist concerning whether or not embodied, conscious persons are the same thing as computers or chairs, simply because computers and chairs, like people, are emergent wholes with emergent properties, the argument should be settled via the distinction in kind between biological wholes and non-biological wholes, not via a tacit appeal to atomism. Similarly, in the tree case, there is a distinction in kind to be made between biological entities that are sentient and those that are not. Certainly, if one is sympathetic to O’Connor and Jacobs’ account, one would have no trouble making these distinctions, since they are adamant that there is qualitative divide between inanimate objects and conscious individuals. Like O’Connor and Jacobs, the holist will argue that agents and bodies are different kinds of entity, and she will maintain that it is embodied agents, not physiological processes, who bear the power of consciousness. But she will not take it to follow from this that only entities endowed with mental properties are emergent. This all matters because there are consequences of adopting the atomist ontology. I suspect that O’Connor and Jacobs’ model will allow one to say all that one needs to be able to say about individual agents, if one is a proponent of agent causation. But as we saw in Chapter 3 with John Stuart Mill, if one denies “chemical composition” in the case of persons, one will be unable to say anything about sociological phenomena. Let me be very clear. By “sociological,” I mean irreducibly relational. And by “irreducibly relational,” I mean to rule out anything that will be counted in an ontological inventory as “just parts,” be they duly arranged or not. Atomists have no place for genuine wholes in their ontology. At the level of social and political theory, this means that the atomist is not entitled to talk of the polis, as Aristotle does – or of societies, or families, or corporations, etc. They are entitled to talk only of pluralities of individuals, of “aggregates,” as Rousseau puts it. And speaking of Rousseau, even if we make an exception for O’Connor and Jacobs on the grounds that unlike some atomists they allow for emergence, on their account of emergence we would at very best be allowed only sociological phenomena such as a completely unified General Will. (I’m not suggesting that Aristotelian, Hegelian or Marxist conceptions of social wholes are necessarily superior to Rousseau’s in virtue of retaining a place for individuals [i.e., parts]. The point is only that if one were to endorse a non-Rousseauian conception, one would have to reject O’Connor and Jacobs’ account of emergence at the level of metaphysics.)

Agents, Powers and Events 91 My goal in this chapter was to trace the influence of Humeanism on the shape of the contemporary free will debate. The very formulation of the problem, I said, runs through Kant, who saw not just that a Cartesian-style materialism leaves no place for the mental, but also that a deterministic environment precludes autonomous acts. Libertarianism, however, turned out to have problems of its own – what we might call “attachment” or “traction” problems, already evident in Kant’s work in the difficulties associated with trying to integrate transcendental agency and empirical agency. The appeal to agent causation would seem to resolve those problems, and readily coheres with our pre-philosophical experience of ourselves as self-determining actors, but it has met with resistance. What I’ve tried to show is that it is at least in part a set of entrenched Humean commitments that has made agent causation so difficult to defend. At every turn, Humeanism undercuts an account of the world in which active entities, including but not limited to intentional beings, do things. If I have succeeded in making my case over the course of the last several chapters, the groundwork will now be laid for the thesis of Chapter 6, which is that anyone hoping to anchor a normative political theory in the concept of capabilities is going to need a metaphysics that allows for powers, persons and agent causation.

6

Metaphysics and the Capabilities Approach Martha Nussbaum, Political Liberalism and the Ideal of Metaphysical Neutrality

Martha Nussbaum and Amartya Sen are the authors of a general theoretical framework known as the capabilities approach. The framework has been put to use both inside and outside of academia, in areas ranging from development economics to human rights advocacy to feminist theory to ethics and liberal political theory. Sen’s version of the approach is primarily intended to introduce new categories of evaluation into debates over economics and social welfare policy. Nussbaum shares these concerns, especially as they relate to women, but has arguably gone even further than Sen in working out a capabilities-based political philosophy. Her initial formulations were self-consciously Aristotelian, and to a certain extent Marxist. In recent years, however, she has taken on the mantle of Rawlsian political liberalism, and so has undertaken to articulate a form of the view that will not commit its proponents to any comprehensive theory of the Good, as Rawlsians like to put it. Nussbaum’s work is more familiar to political theorists and philosophers than is Sen’s. Accordingly, I shall direct my attention to her version of the approach rather than to his. Of course, to the extent that there is agreement between the two, the analysis will pertain to Sen’s thinking as well as to Nussbaum’s. The question that I want to pose is this: is the capabilities approach metaphysically neutral? This is the question that I have posed to thinkers in previous chapters, but the question is a particularly pointed one in this instance because Nussbaum has combined an appeal to political liberalism (the “political” of which is short for Rawls’ “political not metaphysical”) with what certainly appears to be realist talk about dispositions. Given that political liberalism precludes dependence upon Aristotelian norms (or, for that matter, those of any other comprehensive theory of the Good) for the purposes of justification, one might want to say that it also prohibits reliance upon an implicit Aristotelian ontology. With respect to the former, i.e., to the issue of justification, Nussbaum maintains that the capabilities approach is or may reasonably be expected to become the content of a Rawlsian-style overlapping consensus. I shall address this contention only in passing. For the purposes of my own argument, nothing of consequence hangs on the capabilities approach being consistent, if it is, with what amounts to a meta-theoretical principle of normative universality. Proponents of the capabilities approach are free to stipulate that when they

Metaphysics and the Capabilities Approach 93 affirm metaphysical neutrality, what they mean by the term metaphysics is meta-ethics. But I am not concerned with meta-ethics. I am concerned with metaphysics. Specifically, I am interested in whether or not Nussbaum’s version of the capabilities approach is properly available to Humeans. I shall argue that it is not.

The Capabilities Approach as a Political Philosophy Nussbaum has recently published a concise treatise entitled Creating Capabilities: The Human Development Approach.1 Given that her position has evolved over the years, I will take this statement to be the source of record for her present thinking. According to Nussbaum, the capabilities approach may be summed up as the idea that “the key question to ask, when comparing societies and assessing them for their basic decency or justice is ‘What is each person able to do and to be?’ ”2 Her version of the approach centers on ten central capabilities, as she calls them, which she argues provide the content of any compelling answer. These are as follows: 1 Life. Being able to live to the end of a human life of normal length; not dying prematurely, or before one’s life is so reduced as to be not worth living. 2 Bodily health. Being able to have good health, including reproductive health; to be adequately nourished; to have adequate shelter. 3 Bodily integrity. Being able to move freely from place to place; to be secure against violent assault, including sexual assault and domestic violence; having opportunities for sexual satisfaction and for choice in matters of reproduction. 4 Senses, imagination, and thought. Being able to use the senses, to imagine, think, and reason – and to do these things in a “truly human” way, a way informed and cultivated by an adequate education, including, but by no means limited to, literacy and basic mathematical and scientific training. Being able to use imagination and thought in connection with experiencing and producing works and events of one’s own choice, religious, literary, musical, and so forth. Being able to use one’s mind in ways protected by guarantees of freedom of expression which respect to both political and artistic speech, and freedom of religious exercise. Being able to have pleasurable experiences and to avoid nonbeneficial pain. 5 Emotions. Being able to have attachments to things and people outside ourselves; to love those who love and care for us, to grieve at their absence; in general, to love, to grieve, to experience longing, gratitude, and justified anger. Not having one’s emotional development blighted by fear and anxiety. (Supporting this capability means supporting forms of human association that can be shown to be crucial in their development.)

94

Ontology Revisited 6 Practical reason. Being able to form a conception of the good and to engage in critical reflection about the planning of one’s life. (This entails protection for the liberty of conscience and religious observance.) 7 Affiliation. (A) Being able to live with and toward others, to recognize and show concern for other human beings, to engage in various forms of social interaction; to be able to imagine the situation of another. (Protecting this capability means protecting institutions that constitute and nourish such forms of affiliation, and also protecting the freedom of assembly and political speech.) (B) Having the social bases of selfrespect and nonhumiliation; being able to be treated as a dignified being whose worth is equal to that of others. This entails provisions of nondiscrimination on the basis of race, sex, sexual orientation, ethnicity, caste, religion, national origin. 8 Other species. Being able to live with concern for and in relation to animals, plants, and the world of nature. 9 Play. Being able to laugh, to play, to enjoy recreational activities. 10 Control over one’s environment. (A) Political. Being able to participate effectively in political choices that govern one’s life; having the right of political participation, protections of free speech and association. (B) Material. Being able to hold property (both land and moveable goods), and having property rights on an equal basis with others; having the right to seek employment on an equal basis with others; having the freedom from unwarranted search and seizure. In work, being able to work as a human being, exercising practical reason and entering into meaningful relationships of mutual recognition with other workers.3

Having enumerated the ten central capacities, Nussbaum goes on to introduce three further concepts: combined capabilities, internal capabilities and core capabilities. The notion of a combined capability is designed to capture the fact that capabilities are “not just abilities residing inside a person.”4 They are the substantive “freedoms or opportunities” that arise from the “combination of personal abilities and the political, social, and economic environment.”5 Internal capabilities, meanwhile, are characteristics of individual persons, but like combined capabilities they “are not innate equipment.”6 Rather, they are “trained or developed traits and abilities”; in that sense, they have to be acquired.7 Finally, Nussbaum uses the term basic capabilities to refer to those abilities that are, in fact, had as a matter of innate equipment. Human dignity requires that our capabilities be nurtured, says Nussbaum – and, as per combined capabilities, permitted to be expressed, once acquired. Fully actualized capabilities are called functionings. It is capabilities rather than functionings, however, that are the relevant norms. Governments are obliged to promote capabilities, in Nussbaum’s view, but the choice of whether or not to express them, and in what way, rests with individuals. The account involves the idea of freedom in at least two ways. First, capabilities can be seen to give content to the notion of autonomy, beyond that of a Kantian notion of free moral agents

Metaphysics and the Capabilities Approach 95 being bearers of pure practical reason; second, Nussbaum insists that people ought to be free (and able) to decide for themselves whether or not they wish to be (substantively) free. Nussbaum connects the second point to the principle of political liberalism. Political liberalism allows us to agree upon the list of ten central capabilities, without agreeing upon what a life in which they would figure might look like, or upon the norms that would fix such specifics.

The Ontology of Capabilities The question for the purposes of my own argument is whether or not the capabilities approach carries with it any ontological presuppositions. Manifestly, of course, it commits one to a belief in the existence of whatever it is that one takes capabilities to be, and to the bearers thereof. There is also the issue of the nature and source of the dignity that entitles us to realize our capacities, and whether or not the freedom that is identified with full human functioning is consistent either with determinism or with libertarian free will construed in event-causal terms. Nussbaum herself certainly seems to assume that capabilities are real powers, borne by real, self-determining causal agents. But the issue is not whether Nussbaum herself implicitly endorses a dispositional realist model of agent causation. The issue is whether or not a Nussbaum-style capabilities theorist must endorse such a view. Perhaps it will be possible to be a Humean and still endorse a recognizable version of Nussbaum’s position. Powers Let’s begin with powers. Must a proponent of the capabilities approach believe in irreducibly dispositional properties? One might be tempted to say that the answer is yes, but that it doesn’t have anything to do with the capabilities approach: proponents of the capabilities approach must be realists about powers because everyone must – or, at least, anyone who wants to talk about agency must.8 I’m sympathetic to the idea that Humeanism fails generally on this point, but the capabilities theorist who rejects powers will also, I believe, run into difficulties that arise in relation to her adherence to the capabilities approach specifically. As I’ve just said, the meta-theoretical issue is what a capabilities theorist must believe, not what Martha Nussbaum does believe. Still, Nussbaum’s capabilities approach just is the position under consideration. It’s worth noting, therefore, that Nussbaum distinguishes her notion of a capability from Sen’s. For Sen, a capability is a set of circumstances. For Nussbaum, by contrast, even a combined capability has an explicitly dispositional component. In the combined case, the capability is indeed an “opportunity,” à la Sen, but it is an opportunity borne of external environment and of personal ability, as she puts it. The other classes of capability, meanwhile, core and internal, are defined in even more exclusively dispositional terms. They are conceived as being a matter of raw ability and developed or acquired ability, respectively. The question, then, with respect to powers, is what happens if we replace “Smith’s ability to do x” with “the fact that state of

96

Ontology Revisited

affairs y regularly follows state of affairs x”? Do we lose anything that matters for the capabilities approach if abilities are swapped out for states of affairs? It seems to me that we do. First of all, as we saw with Marcuse and with Mill too, if there are no powers, then strictly speaking there can be no realist talk of arresting, or impeding, or preempting, in relation to energetic, dynamic or dispositional psychological phenomena. This is significant. As I have argued throughout, the Humean can’t have it both ways. If one denies the existence of powers, one is not in a position to criticize the restraint or suppression thereof. One might object that I have conflated vocabulary and metaphysics. Philosophically, one might say, talk of thwarted or stunted capabilities is only a bit of metaphor. Nothing hangs on its loss. But metaphorical language deployed for expository rather than expressive purposes must, in order to be intelligible, be intended to capture something about the object of analysis. In this case, what Nussbaum says about the object of analysis is this: [s]ocial, political, familial, and economic conditions may prevent people from choosing to function in accordance with a developed internal capability: this sort of thwarting is comparable to imprisonment. Bad conditions can, however, cut deeper, stunting the development of internal capabilities or warping their development.9 Nussbaum, like Mill, emphasizes unrealized potential rather than repressed desires, but, as with Marcuse, the contention is that something active has been held back. This is the philosophical content of the claim, not its rhetorical dress. And it is a contention for which there is no adequate Humean translation, since the idea of holding back a genuinely (or irreducibly) dynamic force cannot be expressed in a conceptual vocabulary in which dispositional terms are disallowed as a matter of ontology. At the normative level, meanwhile, if the human capabilities that stand to be thwarted are not real powers, it is less apparent that we should care about the thwarting. Nussbaum stipulates that the mere fact that human beings have capabilities does not in itself imply that the capabilities ought to be cultivated – and I shall address that point further in a moment – but whatever else it trades upon, the “ought” gains traction from the presumption that it is real efficacy that is at stake, real potency. As with Mill and Marcuse both, Nussbaum would have us defend and foster the potentiality to which she directs our attention. We are expected to be moved by the threat of its truncation, by the foreclosing of “nascent” abilities, as she puts it, citing Adam Smith.10 For Mill, we should care because the exercise of our powers in pursuit of the higher pleasures just is that which is in our permanent interest as progressive beings, to use his phrase. Moreover, in addition to being good for its own sake, full functioning will make us happy. For Marcuse, we should care because the free expression of our powers in accordance with a nonrepressive reality principle will make us happy and free. For Nussbaum, we should care because the central capabilities are linked to human dignity. In all three cases, it is not at all clear that we would care if capabilities were not understood to be real powers, powers that can be blocked, stunted, curtailed or compromised. At the

Metaphysics and the Capabilities Approach 97 very least, if we still cared, we would be caring about something else; we would not be caring about what capabilities theorists would have us care about. Moreover, it’s hard to attach a sequence of states of affairs to an agent in a manner that is comparable to the sense in which a power that is had by an agent attaches to her. This is to echo the complaints that agent causal theorists voice in relation to nondeterministic event-causation. If I know how to read, I have a cognitive ability that is anchored in me as a subject. Reading isn’t something that happens to me; it is something that I myself do. I am the reader. This fact authorizes a claim about my identity, too: I am a reader. The Humean can say only that the world is such that there is a constant conjunction between my looking at (or in the case of Braille, feeling) a text – where looking and feeling are themselves understood non-dispositionally, as sequences of states of affairs – and my thinking certain words. I don’t mean to argue the case here that a Humean construal of reading is false. I mean only to draw attention to the degree to which the agent necessarily drops out of the picture, on an account of reading (or of any other activity) that is consistent with there being no powers. And the point that I want to make is that such account, one in which the agent is not the locus, cuts against the grain of the capabilities approach in particular. For it is not powers as such that the capabilities theorist would have us care about; it is the powers or capabilities of the persons who bear them. Once again, even if one were to find oneself moved by constant conjunctions, the object of one’s moral concern would not be that presumed by the capabilities approach, which is not constant conjunctions but persons, persons qua bearers of capabilities. Finally, it seems to me that, having rejected the existence of powers, the Humean will have trouble with the distinction between capabilities and functionings. A functioning is an actualized ability, a successfully executed bit of doing. As defined by Nussbaum, a capability is either (a) an existing but unexercised ability to do something or (b) a potential (and therefore unexercised) ability to do something. The dispositional realist will think that the idea of unrealized potentiality is part and parcel of the concept of a power, since a power need not be expressed.11 The Humean has no comparably direct access to the idea of potentiality. Instead, she can only stipulate that there are circumstances in which x is followed by y, and observe that such circumstances do not presently obtain. Some will think that with the help of David Lewis she can go on to offer a contemporary Humean account of the word “would,” such that she will be able to equate a capability with a sequence (technically, an instantiation of a regularity) that would occur, given suitable circumstances.12 And if she does, then perhaps she will be able to preserve the distinction after all. Capabilities, she will say, are regularities that aren’t instantiated, whereas functionings are regularities that are instantiated. The gain is short-lived, however, as we are now back to being asked to care about ordered pairs of states of affairs. To make matters worse, now the conjoined states in question (i.e., those that underwrite the “would” that distinguishes a capability from a functioning) – now they are features not of the actual world but only of what Lewis calls possible worlds. I am not going to weigh in on the metaphysics of whether or not a regularity that isn’t instantiated in the actual world is what a potentiality is. What I do mean to suggest, however, is that how things are with people in worlds other than this one are not the

98

Ontology Revisited

sort of thing in relation to which human dignity plausibly stands or falls. At least, they are not what plays that role for a capabilities theorist. This said, Nussbaum’s appeal to dignity warrants further discussion. Her claim is that a life characterized by human dignity necessarily includes the exercise of the central capabilities, and that human beings are intrinsically entitled to have the opportunity to live out such lives. But why? What’s the basis for the dignity? This needs to be clarified before we can say more about whether or not the account can be meaningfully translated into a Humean register. Contemporary dignity talk is most often Kantian. On the Kantian model, human beings compel respect insofar as we are transcendental egos, noumenal selves that just are the fact of pure practical reason. In fact, one might argue that for Kant it is reason that commands respect, concrete persons only derivatively. Nussbaum rejects this model. For one thing, she explicitly cites the Stoics, rather than Kant, as the primary influence for her concept of human dignity.13 Moreover, entities lacking human intellectual powers, she says, contra Kant, also have dignity – the dignity that attaches to their kind specifically. She is explicit about this: “Unlike expanded Kantian approaches, which see duties to treat animals well as derivative from duties to support our own human animality, the Capabilities approach regards each type of animal as having a dignity all it’s own.”14 Thus, “[t]he main conclusion of my approach is that animals are entitled to a threshold level of opportunity for a life characteristic of their kind.”15 What is it, then, about being born a human that, on the capabilities approach, confers specifically human dignity upon a person, such that he or she is entitled to actualize his or her central capabilities? I think that Nussbaum’s response to this question is not actually what she believes it to be. She introduces the concept of dignity because she thinks that not all capabilities need be objects of concern. The mere fact that we have a capacity to do or to be x, she says, does not imply that such a capacity ought to be nurtured. “The Capabilities Approach,” she writes, is not a theory of what human nature is, and it does not read norms off from innate human nature. . . . An account of human nature tells us what resources and possibilities we have and what our difficulties may be. It does not tell us what to value.16 What it does tell us what to value is “the notion of human dignity and of a life worthy of it.”17 The notion is an “intuitive” one, she says, the content of which “is by no means utterly clear”: its meaning emerges only when it is considered in relation to other aspects of the approach.18 And what we discover about it, in the course of Nussbaum’s “holistic and nonfoundational type of justification,”19 is that it is “closely related to the idea of active striving. It is thus a close relative of the notion of basic capability, something inherent in the person that exerts a claim that it should be developed.”20 The main difference, it turns out, between basic capabilities and dignity, is that the former may be differentially distributed amongst and across populations, whereas the latter is “equal in all who are agents in the first place.”21

Metaphysics and the Capabilities Approach 99 I believe that the account here needs clarifying. On a sharpened-up reconstruction of Nussbaum’s position, kind-specific dignity can be seen to be accorded to creatures on the basis of their distinctive capabilities. When it comes to humans, it looks as though the key capability to which a universal dignity might be pegged is that of practical reason, the ability to choose a course of action – an ability that, as it happens, Nussbaum singles out (along with affiliation) as architectonic.22 Alternately, a trans-species dignity may be thought to be grounded in a more general capacity for self-direction that humans share with other animals. In either case, it’s not that dignity is itself a power, an active striving; it’s rather that the active, intentional strivings that constitute us as agents entitle us to dignity. Why? I think the answer has to be that agency, the capacity for conscious self-determination, is inherently worthy of respect. And this idea – evocative of Kant, even if the autonomy in question is, in Kantian terms, empirical rather than transcendental – this idea is arguably the crux of the view. It is the central capabilities themselves, and in particular the “master” capacity of intentionality, that underwrites the dignity in virtue of which said capabilities warrant being cultivated. The Self I have already addressed the question of whether or not capabilities may be plausibly construed as regularities, rather than powers, by a capabilities theorist. The next issue to consider is that of the self. Is the capabilities theorist free to be an anti-realist about the agents who bear the dignity-conferring strivings that are characteristic of human beings? To be sure, realism about powers does not imply realism about selves. In principle, it would seem – even if we don’t often see it in practice – one could believe in real dispositional properties, but still think that so-called selves reduce to regular sequences à la Hume. It is also possible that there are good reasons for anyone to believe in the reality of selves, reasons unrelated to the capabilities approach. At issue, then, is whether or not the capabilities approach in and of itself commits its proponents to a belief in real selves, in addition to real powers. There are several reasons to think that the answer is yes. First, as we’ve seen, on the capabilities approach, capabilities are characteristics that are borne by given kinds of entity. In the human case, Nussbaum explicitly notes that it is persons (rather than groups) who bear them “first and foremost.”23 While it is only the core capabilities that are conceived as “innate equipment,” as she puts it, it is clear that all of the capabilities are thought of as being attributes of persons: indeed, of agents. It is agents who have the potential to live out lives of a normal length; agents who experience longing, gratitude and justified anger; agents who form conceptions of the good and who affiliate with others and who exercise control over their environment, when they can. The Humean is certainly free to think that the purported selves who can do such things are really only clusters of constant conjunctions. But if one adopts a Humean metaphysics, one will have to adjust the picture accordingly. At a minimum, on a Humean model there will be no more “of.” Instead, there will simply be constantly conjoined states of affairs, some of which occur only in possible worlds. As a matter of ontology, one is no

100

Ontology Revisited

longer in a position to say of a person that she has the ability to use her imagination, for example. Instead, one would have to say that the world is ordered such that it just does include certain patterns of mental states (or, for the reductive physicalist, physiological states), and that these patterns may be added to the pattern of regularities that just “are” the person in question. As it happens, I don’t find the Humean story to be persuasive. But I don’t need to defend my reservations for the purposes of the present argument. My claim is only that once one has transposed “capabilities borne by individuals” into “ordered pairs of actual or possible states of affairs,” one has lost contact with one of the governing intuitions of the capabilities approach, which is that persons (real persons, not the fact that certain regularities are or could be instantiated) are the bearers of certain normatively salient attributes. Again: it’s not that a Humean can’t prefer one state of affairs to another. It’s that the Humean can’t ascribe powers to substances, let alone to substances who are selves. Second, Nussbaum claims that “[t]he [capabilities] approach espouses a principle of each person as an end.”24 I think it will be hard for the Humean capabilities theorist to make sense of such an idea. For Kant, whom such ends-talk echoes, individuals are ends in virtue of their transcendental identity as moral actors. For Nussbaum, it looks as though individuals are Kantian-style ends in virtue of whatever it is that analogously grounds human dignity on her view. I argued above that according to the logic of the capabilities approach, dignity is best understood as being conferred by the capabilities themselves, in particular by the over-arching capability for selfdetermining intentional action. On the Humean model, by contrast, it will have to be a set of constant conjunctions – rather than Kant’s transcendental ego or Nussbaum’s real, embodied agent – that is the unqualified end in question. One question for the Humean will be why those regularities that constitute “selves” should be regarded as special. If selves just are constantly conjoined states, then it can’t be that some constantly conjoined states are special because they are selves. Indeed, apart from the issue of their status as ends, it is not clear what the basis will be for individuating regularity-selves in the first place. (Though if this is a problem it will be a problem for Humeanism as such, not for Humeanism specifically as it relates to the capabilities approach.) At the deepest level, I think, the difficulty for a capabilities theorist who wants to be an anti-realist about selves will be the same one that we saw in relation to the capabilities theorist who rejects real powers: regularities aren’t the kind of thing that people care about. The Humean is perfectly entitled to say that, in her view, constantly conjoined states are what so-called selves really are – just as she was free to say that constantly conjoined states are what so-called powers really are. But if it was not clear that constant conjunctions are morally compelling enough to stand in for real powers, it is even less clear that they are sufficiently compelling to play the role of selves that are ends. Finally, it’s not just that on the capabilities approach human capabilities are powers of human beings, and that this aspect of the theory – it’s attributive structure, I’ll call it – will be lost in the translation into a Humean register. It is also that, as with Mill and even Hume, the selves that are implicit in the capabilities approach have a lot of work to do – more, I think, than can be done by patterns of

Metaphysics and the Capabilities Approach 101 constant conjunctions. This point is similar to the earlier one, but it isn’t quite the same. My claim before was that the attributive structure of the capabilities approach necessarily disappears, in a Humean rendition of the view, such that the Humean version would be a different kind of theory. Here the claim is that a self that amounts to regular sequences isn’t robust enough to function in the ways that the selves that figure in the capabilities approach are defined as being able to function, assuming conditions of justice. The selves that are endowed with the central capabilities can do things such as imagine the situations of others, participate effectively in political activity, feel concern for plants, make art, form lasting bonds, pursue goals, share humor. Above all, the selves that are bearers of dignity in accordance with the capabilities approach, be they human or of another species, are agents. They don’t simply engage in kind-specific activities; they are the authors of such activities. They are not just objects, but subjects. Constantly conjoined states of affairs cannot do or be any of these things. Free Will and Agency The fact that agency is at the heart of the capabilities approach raises the issue of free will. Is the capabilities approach neutral with respect to the questions discussed in Chapter 5? I think that the answer has to be no. Let’s begin with compatibilism. If one is a compatibilist, one will think that so long as we are not coerced, we are metaphysically free. But, the compatibilist will say, even when we are not coerced, what we want to do is itself deterministically caused, just like everything else. Admittedly, Mill (as we have seen) tried to hold both (a) that we ourselves can be sui generis causes of our character and (b) that determinism is true. But his is not the standard view, and the problems with it, as flagged in Chapter 3, are not hard to spot. The typical compatibilist is likely to say that freedom and determinism are consistent because to be free is to be able to do what one wishes to do, regardless of how or why one has come to have the desires and objectives that one has. It may have followed deterministically from antecedent conditions that I should want to write this book, but in writing it I am nevertheless a free agent, exercising free will. Can such a view be combined with the capabilities approach? I don’t think so. To see why, consider Nussbaum’s logic in pegging justice to capabilities rather than to functionings. Some thinkers, she says, hold that the right thing for government to do is to make people lead healthy lives, do worthwhile activities, exercise religion, and so on. We deny this: we say that capabilities, not functionings, are the appropriate political goals, because room is thereby left for the exercise of human freedom. There is a huge moral difference between a policy that prohibits health and one that promotes health capabilities – the latter, not the former, honors the person’s lifestyle choices.25 The compatibilist (who thinks that we are metaphysically free insofar as we get to do what we want) will certainly agree with the idea that people ought to be able to

102

Ontology Revisited

live out their lives as they see fit. But the point of the passage is not quite that. What Nussbaum is at pains to establish is that, in her view, we must be able to choose for ourselves whether or not to function well. We may not be forced to be free, as Rousseau would have it – not even when freedom is defined as fully exercising one’s capabilities as a human being. Is the compatibilist in a position to insist upon this degree of choice? It seems to me that she is not. Consider a woman whose desire structure is causally determined by patriarchal norms such that she does not wish to exercise any of her capabilities. The compatibilist will have to say, of such a person, that she nonetheless enjoys the freedom of choice upon which Nussbaum insists. To be sure, from the perspective of the compatibilist herself, her view is consistent with the capabilities approach. But things look different from the other direction. For the capabilities theorist, the person who cannot even choose whether or not she desires to exercise her capabilities (let alone whether or not to actually do so) – such a person is in a situation that the theory is expressly designed to call out as wrong. Such a person is most pointedly not free. Moreover, for the compatibilist, it’s not just that some especially oppressive set of circumstances might shut down a person’s ability to choose what to want; rather, it’s always that way. No one is ever thought to have control over their foundational or originating desire structure; they are free only to do or to not do what they are deterministically caused to want to do. The capabilities approach has built into it the presumption that such a view of freedom is false: the person whose very desires are imposed upon them externally will not show up as free simply because the coercion is cast as metaphysical rather than political-economic. What about agent causation? If the capabilities approach does not sit well with determinism, is it at least equally consistent with different versions of libertarianism? Here again, I don’t think so. As discussed in Chapter 5, in the contemporary analytic context agent causation is often introduced as a solution to the intelligibility problem, viz., that acts of libertarian free will, insofar as they escape the hold of deterministic causal laws, must therefore be uncaused. No, says the agent causal theorist: they are caused by agents. As we saw, the intelligibility problem can be broadened into what might be called a “traction” problem: namely, that it seems as though event-causal libertarian accounts leave us with free actions that are either uncaused or caused indeterministically, but in either case are not caused by the free agents in question. And in neither case – whether actions are thought to be a-causal or to be indeterministically caused by antecedent conditions and probabilistic laws – in neither case does it look as though they will connect up with supposedly free agents in such a way that agents will be their authors.26 If this is correct, though, then it might seem as though all libertarians should be proponents of agent causation – and thus that there is nothing about the capabilities approach in particular that would commit its implicitly libertarian endorsers to agent causation.27 I do think that event-causation based libertarianism fails generally, but it seems to me that the traction problem has differential significance vis-à-vis the capabilities approach. While the lack of traction associated with event-causation libertarianism might persuade a Humean or a Kantian to opt

Metaphysics and the Capabilities Approach 103 for agent causation (though this would arguably issue in other difficulties for each), the event-causation model is at least consistent with an underlying Humean or Kantian-style Humean metaphysical framework. This is not so in the analogous case of the capabilities approach. Here’s why. A Humean who was not a compatibilist – an uncharacteristic Humean – would be most likely to defend an indeterministic event-causation version of libertarianism. And characteristic or no, the fit is in fact seamless. Indeed, if Hume hadn’t re-defined deterministic necessitation to mean “customary and expected,” we would say that indeterministic event-causation just is the ordering principle of Hume’s own ontology. The Kantian, meanwhile, is likely to defend a version of libertarianism that involves a-causal spontaneity. And here too, regardless of whether or not the position fails due to the traction problem, the notion of a-causal spontaneity is not only consistent with the underlying metaphysics in question; it is, as I argued in Chapter 4, a product of it. The situation is otherwise for the proponent of the capabilities approach. If contingently conjoined events are at the heart of Hume’s normative theory, and the autonomy of the noumenal self at the heart of Kant’s, the “active striving” of real, concrete selves is at the heart of the capabilities approach. I’ve already offered reasons for concluding that the abilities in question must be powers, and not regular sequences; the selves genuine entities, and not regular sequences. At a minimum, such an ontology fits as seamlessly with an agent-causation libertarianism as Hume’s and Kant’s do with indeterministic and a-causal libertarianism. But the point is not just that there is a natural affinity between the capabilities approach and agent causation, no matter how striking an affinity it may be. The point is that insofar as the capabilities approach does indeed hinge on the active striving of real, concrete selves, it is in fact at odds with accounts of free will in which events are causes and agents are not. Again, it may well be that what I’m calling event-causation libertarianism fails generally because of the traction problem. But it’s a position that is not available to the capabilities theorist in the first place. The capabilities approach commits one to agent causation all on its own.

Political Liberalism and the Capabilities Approach I have tried to show that the capabilities approach is not in fact metaphysically neutral. As a matter of coherence if not strict logical entailment, it requires a belief in real powers; in real selves; in libertarian free will and in agent causation. And because it does, it is not consistent with a Humean metaphysics or an eventcausation model of agency. The Kantian, who at least has recourse to spontaneous albeit noumenal selves, will get further than the Humean empiricist in an effort to embrace the approach, but in the end the over-all position will not hang together if the underlying ontology cannot sustain it. The question that I want to consider now is what to make of such a fact. Does what I’ve said have any bearing on Nussbaum’s meta-theoretical appeal to political liberalism? And more generally, what if anything does it imply about efforts to synthesize radically different lines of thought from the history of social and political philosophy?

104

Ontology Revisited

As I suggested at the outset, I don’t think that the argument that I’ve made precludes the capabilities theorist from appealing to political liberalism as a normative principle. Such an appeal does two things. Meta-theoretically, it allows proponents of the approach to connect their concern for the central capabilities to alternate comprehensive accounts of the Good, as they see fit. One capabilities theorist may see human capabilities as gifts from God, another may not. One may believe that the organizing principle of a well-lived life is compliance with duty; for another it may be responsiveness to the requirements of beauty, say, or alternately of victory. Be these differences of orientation as they may, it should be possible, says the political liberal, for everyone to agree to the defining terms of the capabilities approach. Within the frame of the theory itself, meanwhile, the appeal to political liberalism underwrites the prioritization of capabilities over functionings, and keeps it an open question, not to be decided by the theory itself, which capabilities will matter most, and what their various expressions will look like. The ontological commitments that I have identified do not decide any of these issues. Realism about powers doesn’t commit one to theism rather than to atheism, for example. Nor does it follow from the rejection of compatibilism that one must emphasize one capability over another, or one version of functioning over another. It would be a mistake, therefore, to think that the capabilities approach must be available to Humeans on pain of violating the principle of political liberalism. Instead, it would seem that the correct application of political liberalism to matters ontological would be to say that while the capabilities approach presumes realism about powers and selves, and agent causation rather than event-causal libertarianism or compatibilism, its proponents are free to defend such positions on grounds of their own choosing. For instance, one might say, the capabilities approach is properly neutral with respect to debates between Ellis, Mumford and Bird on the nature of dispositions; or between Lowe and O’Connor on selves and agency. Admittedly, it’s not as though there’s no connection at all between different comprehensive views of the Good and ontology. If one believes in the existence of deities, for example, one must endorse a metaphysics that includes such entities in its inventory. In a similar vein, if one values loyalty to family, at the expense of the lives of individual family members, then one will thereby be committed to the existence of the relational whole that is the object of such fidelity. But this isn’t the level at which the capabilities approach potentially sits uneasily with the principle of political liberalism. If there turns out to be a tension between the capabilities approach and political liberalism, it will not be because the position is unavailable to Humeans since they deny the existence of powers, and the selves pproach doesn’t square with determinism, or with that bear them. Or because the approac event causation. Rather, it will be because the capabilities approach hinges on the belief that the central capabilities are universal, and, even more to the point, that all persons are entitled to function as they see fit. Political liberalism is only applicable when everyone agrees to the normative precept in question, when the disagreement is merely over why it’s true. If there is a worry to be had concerning

Metaphysics and the Capabilities Approach 105 political liberalism and the capabilities approach, it’s that in this case the requisite normative consensus may not exist. I want to be as clear as possible about what I do and do not mean to have suggested in the foregoing analysis. I do hope to have shown that Nussbaum’s capabilities approach cannot be sat a-top a Humean metaphysics any more than can Mill’s call for the cultivation of distinctively human capacities in the energetic character type. A political theory centered on the good of fostering the powers of individuals to affect the conditions of their own agency is simply at odds with a metaphysics that disallows powers, selves and agent causation. But I do not mean to claim that all conceptual syntheses, or even all efforts at meta-theoretical pluralism, are bound to be incoherent. On the contrary, it is an open question which ideas may be successfully lifted out of context and paired with others. Nussbaum herself does not attach herself to Hume. Rather, she says: “[a]mong Euro-American antecedents, the most important sources for my version of the Capabilities Approach are works from ancient Greece and Rome, although Smith, Kant, Mill and Marx have also greatly influenced my formulations” – adding that “John Rawls’s work has been of the utmost importance, particularly in convincing me that the view ought to be expressed as a type of political liberalism.”28 Is this a viable amalgamation? The answer is yes and no, I think. Yes, if and insofar as the focus is on the substantive elements of the view. That is, it seems perfectly reasonable to claim that, drawing from these thinkers in various ways, there is an account to be given of various powers that all human beings have; the dignity and right to self-determination that such powers ground; the wherewithal for full functioning to which all are entitled; and the autonomy constituted therein. No, if and insofar as the view includes a further belief to the effect that such an account is ontologically free-floating. I hope at this point to have cast serious doubt on the myth of metaphysical neutrality, if not to have debunked it altogether. Social and political philosophy is always already metaphysically committed. I hope, too, to have brought out some of the difficulties that arise when there is an effort to theorize agency within the Humean framework. It’s not that there is nothing at all for the Humean to say. It’s that what the Humean has to say is so wildly implausible: there is just no good reason to think that agency, properly understood, is analogous to the apparent motion of figures in an animation flip-book. But to say something other than that, one must call upon a non-Humean ontology. In Chapter 1, I set out a minimal definition of such a position, called dispositional realism. By way of conclusion, I now want to raise the question of where, in the history of Western philosophy, one might do well to turn for insight (or not so well, as the case may be), if one were persuaded of the need for a powers-based metaphysics – and, further, just what such insight may or may not be expected to yield.

7

Powers, Ontology and the Appeal to Spinoza

I hope to have made good on the thesis that social and political theories have metaphysical commitments built into them, as do accounts of the nature of agency itself. These assumptions shape and constrain the express content of a theory, making it more or less natural, or possible, for its proponents to articulate or defend given concepts or lines of argument. Hume’s rejection of irreducibly dispositional powers is an especially dramatic instance of such a configuring of options, as it disallows recognizable talk of doing altogether – be this the doing of individuals, collectivities or social structures. Thus we saw that Mill lacks the necessary ontology to sustain the ideal of the energetic character type that he otherwise so ably champions; that, if implicitly, members of the Frankfurt School must rely upon non-Humean resources in order to carry out their respective appropriations of Kant; and that Nussbaum’s capabilities approach loses its meaning, and arguably its hold, if the capabilities in question are taken to be mere sequences of states of affairs, as they must be for the Humean. Even Hume himself, I observed in Chapter 2, cannot make do without reference to the active powers of the sentiments. Similarly, moving from the history of political thought to contemporary analytic philosophy, we found that the very notion of agent causation is foreclosed by the Humean framework of passivist event causation. We saw too that the rejection of powers removes a potential pillar of support for holism, although holding a powers-based metaphysics does not conversely entail that one will either ascribe powers to irreducibly relational wholes or believe the ascription of unique powers to a purported entity to be an indicator of its irreducibility. In Chapter 1, I contrasted Humeanism with dispositional realism, which I described there as loosely neo-Aristotelian. And in subsequent chapters I have pointed to Aristotle as a thinker whose metaphysics is more promising than Hume’s (or than Kant or Mill’s revised versions thereof) for the purposes of theorizing “doing,” as I put it above. To be fair, there are dispositional realists who would look to Locke for their powers before looking to Aristotle, and still others who view contemporary philosophy as unrelated to the history of philosophy. This is not the place to weigh in on family differences amongst analytic metaphysicians regarding the genealogy of the concept of a power. I do, however, want to take a moment to address the recent interest in powers (and in ontology more generally) that has emerged in the area of Left continental political theory. The

Powers, Ontology and the Appeal to Spinoza 107 appeal from this quarter is not to Aristotle or to Locke, but to Spinoza. Such an appeal is unlikely enough, yet borne of assumptions that are understandable enough, that I believe it warrants comment in the present context. The excitement about Spinoza stems from core works of Deleuze, Gauttari, Hardt and Negri, but for the sake of efficiency I shall treat as emblematic of this material a recently published book by Jane Bennett, which draws upon and crystallizes the continental literature. The book, called Vibrant Matter, has garnered considerable attention from an audience of political theorists newly concerned with fundamental metaphysical issues.

Vital Materialism Vibrant Matter is a pitch for what Jane Bennett calls vital materialism. Vital materialism is the view that the material world has “thing-power,” as Bennett calls it, such that all entities, and not just those that are human or animate, must be counted as “actants,” a term she takes from Bruno Latour. In naming the position as she does, Bennett means to invoke the tradition of vitalism proper. If Cartesian mechanists thought the universe to be like a clock, vitalists conceived of it in terms best captured by Dylan Thomas’ “The Force that Through the Green Fuse Drives the Flower.” But though she means to invoke vitalism, she doesn’t mean to endorse it. Henri Bergson’s notion of an élan vital and Hans Driesch’s notion of entelechy both fall short, she thinks, because they rest, paradoxically enough, on a mechanistic understanding of materiality. Minus the infusion of an external “not-quite-life-force,” she contends, matter is conceived by Bergson and Driesch alike as being inert. Bennett’s own view, by contrast, is that matter is itself already vital. It is a “vibratory effluescence that persists before and after any arrangement in space: the peculiar ’motility’ of an intensity,” she says, citing Deleuze and Guattari;1 “[i]n this strange, vital materialism, there is no point of pure stillness, no individual atom that is not itself aquiver with virtual force.”2 At the same time, despite flirtations with the reader to the contrary, she herself stops short of fullblown vitalism. For her, the vital force in question amounts to motion, in the end, not to a property of alive-ness. That matter “itself” is alive she equates with the view that material entities have souls, a thesis that she rejects. That all entities are actants is thus an identity claim. The idea is not, or not only, that there are entities, and that they are active; rather, it is that matter just is, in its essence, activity construed as movement, as velocity. Ordinary objects appear to be substances, but the appearance is misleading. In reality they are “mobile, internally heterogeneous materials whose rate of speed and pace of change are slow compared to the duration and velocity of the human bodies participating in and perceiving them.”3 Just what renders such “materials” genuinely material is left unexplained, and Bennett concedes that “[i]t is hard . . . to keep one’s mind wrapped around a materiality that is not reducible to extension in space, difficult to dwell with the notion of an incorporeality or a differential of intensities.”4 She references Spinoza in relation to this point, but it is important to note that Spinoza himself, contra Bennett, does indeed reduce materiality to extension in space,

108

Ontology Revisited

even if extension, in turn, is merely a conceptualized attribute of the one substance that is God. (For those who prefer to say that Spinoza’s one substance is nature rather than God, the contrast with Bennett’s proposed conception of matter as incorporeal will be even sharper.) While every entity is an actant, some actants look to be solid. Other actants are manifestly porous: composite entities that are obviously comprised of other actants. Again in keeping with Deleuze and Guattari terminologically, Bennett calls the latter type of actant an assemblage. In an assemblage, actants of different types (human actants and non-living physical actants are the ones that most interest her) come together to form “living, throbbing confederations,”5 which are the bearers of “congregational”6 rather than “atomistic”7 agency. Assemblages are “ad hoc groupings,”8 “swarms of vitalities”9 that may or may not come into being in the first place, and, once formed, may or may not dissolve back into their constituent parts. A major fluctuation and blackout involving the North American power grid in the August of 2003 is presented as a prime example of such an actant and the emergent powers thereof. Bennett writes: [The electrical power grid] is a material cluster of charged parts that have indeed affiliated, remaining in sufficient proximity and coordination to produce distinctive effects. The elements of the assemblage work together, although their coordination does not rise to the level of an organism. Rather, its jelling endures alongside energies and factions that fly out from it and disturb it from within. And, most important for my purposes, the elements of this assemblage, while they include humans and their (social, legal, linguistic) constructions, also include some very active and powerful nonhumans: electrons, trees, wind, fire, electromagnetic fields.10 For Bennett, the important idea with respect to the phenomenon of the assemblage is that agency is not unique to human beings. Non-human actants are agents, and so too are the assemblages that humans and non-human actants jointly constitute. Bennett describes such an approach to agency as “distributive,”11 in that agency is thought to be spread across an assemblage, rather than being a property only of human beings. She also makes use of the distinction, quoted above, between what she calls “atomistic” and “congregational” conceptions of agency. An atomistic agent is one who is conceived as being a single actant; a congregational agent is understood to be a grouping of actants, one that brings about emergent effects in virtue of the association of its members. Bennett draws on the distinction for the purpose of rejecting a human subject-centered orientation. Ultimately, however, she maintains that even individual persons are congregational agents. A distributive approach to agency brings with it the political problem of specifying which kinds of agents should have a say — which kinds of agents should have access to power. Bennett responds to this problem in two ways. First, she pursues the idea that the “appropriate unit of analysis for democratic theory” is not individual humans or groups of humans, but human–nonhuman assemblages.12

Powers, Ontology and the Appeal to Spinoza 109 Here she draws on Dewey and Ranciere, linking the ideas of both to the Spinozan metaphysics that she takes to underwrite the notion of an assemblage in the first place. Dewey, she writes, “imagines a public as a set of bodies affected by a common problem generated by a pulsing swarm of activities.”13 If we allow that the relevant bodies may be actants of any kind, as she would have us do, then “[w]e . . . then get this (Spinozist) version of Dewey’s theory of the public and of conjoint action: problems give rise to publics; publics are groups of bodies with the capacity to affect and be affected.”14 Ranciere, meanwhile, conceives of “the being of the demos: not as a formed thing or fixed entity, but as an unruly activity or indeterminate wave of activity,”15 and Bennett supports the view. Rightly understood, then, the body politic is an assemblage of heteronomous “materials” just as all actants are – in quotations to signal that materiality itself, for Bennett, reduces to motion. As noted, she understands such a conception to be Spinozan, writing “[t]his idea of a force that traverses bodies without itself being one resonates with Spinoza’s conatus and Deleuze’s notion of (the motility of) intensities.”16 Second, she nevertheless stipulates that there are relevant differences between kinds of actant. “Of course,” she says, to acknowledge non-human materialities as participants in a political ecology is not to claim that everything is always a participant, or that all participants are alike. Persons, worms, leaves, bacteria, metals, and hurricanes have different types and degrees of power, just as different persons have different types and degrees of power, different worms have different types and degrees of power, and so on, depending on the time, place, composition, and density of the formation.17 Vibrant Matter is not intended to be a work of metaphysics for its own sake. On the contrary, Bennett’s goal is to effect a shift in environmental theory and policy. The claim on behalf of vital materialism is that it leads us to conceive of our relationship to the natural world as one of subject-to-subject, rather than subject-toobject. Needless to say, there is a tradition of argument into which Bennett fits, according to which the objectification of nature is internally related to the objectification of humanity. But Bennett believes that there has not yet been sufficient recognition of the subjectivity of nature – better, of materiality itself – such that all objects, even inanimate ones, would be viewed as genuinely agential, even if not consciously purposive. If we were to see ourselves as co-actants in human– nonhuman assemblages, rather than as agents who act against a backdrop of mechanistically construed external conditions, she concludes, we would fare better in our dealings with our home planet.

Vitality, Metaphysics and Spinoza What are we to make of all of this? Can a Spinozan ontology move us beyond mechanism, onto the terrain of an open-ended, powers-based materialism? Especially, how does vital materialism compare to dispositional realism? Bennett’s

110

Ontology Revisited

goal is to dissolve the dichotomy between subject and object, agent and nonagent, by getting us to see all entities as actants, and ourselves always only as co-actants (even our physical bodies she sees as assemblages involving bacteria, food, etc.). Such an undertaking seems reasonable enough. As metaphysics, it amounts to an endorsement of substance causation couched in romantic terms; normatively, there is little question but that our imperious stance has brought us to the brink of global catastrophe. Surely an injunction to recognize our proper place alongside other elements of the natural world is warranted. And yet, it seems to me that those who would look to Spinoza have got the metaphysics wrong. To begin with, I would suggest that in Bennett’s case the present philosophical situation has been misdiagnosed. Specifically, as Bennett would have it a belief that humans are the sole agents is an integral component of the mechanistic mindset. From a mechanistic perspective, she claims, “[a]ctive action or agency belongs to humans alone.”18 She continues, citing Elizabeth Brumfield: “ ‘All agree that agency refers to the intentional choices made by men and women as they take action to realize their goals.” ’19 But – and this is crucial – “all” do not agree. In fact, as we saw in Chapter 5, agent causation is not at all an easy position to defend in the contemporary context. A statement of Bruno Latour’s that Bennett cites as radical, in that it is at odds with a human and subject centered conception of agency (“There are events. I never act; I am always slightly surprised by what I do. That which acts through me is also surprised by what I do . . .”20) is in fact entirely orthodox. Latour is simply voicing the dominant, event-causal view. The comment is notable only for Latour’s apparently being charmed with, rather than concerned by, what I called earlier the traction problem. Moreover, even if agents (or substances of any kind) were normally imagined to be causes, the physicalism that is part and parcel of what Bennett calls mechanism turns subjects into objects. Between event-causation and physicalism, human actants are erased by mechanism just as surely as are non-human actants. The call to de-humanize our thinking about agency therefore puts the emphasis on the wrong syl-lah-ble, as the expression goes. Humeanism in its standard form has already de-humanized our thinking about agency. (I will return momentarily to the question of how, or if, Bennett will be able to establish substances of different kinds, with different kinds of powers, on the basis of a materiality that bottoms out in motion.) This brings me to a second issue, which is the flirtation with vitalism. Bennett calls chapter 5 of her book “Neither Vitalism Nor Mechanism,” and she insists that the vitality that she means to assert as ontologically primary is not a life force – not “soul”-like, as she puts it. But she is coy about it all the same. I say this because nearly all of the rhetorical work of the book is carried out through the use of unreservedly vitalist language. The electricity of a power grid “sometimes . . . chooses its path on the spot.”21 “Thus spoke the grid. One might even say that it exhibited a communicative interest.”22 Non-organic actants have what Bennett, referencing Deleuze, calls a life. “[A] life is not only a negative recalcitrance but a positive, active virtuality: a quivering protoblob of creative élan.”23 “In this onto-tale, everything is, in a sense, alive.”24 Invoking Spinoza (or, believing

Powers, Ontology and the Appeal to Spinoza 111 herself to be invoking Spinoza): “an active becoming, a creative not-quite-human force capable of producing the new, buzzes within the history of the term nature. This vital materiality congeals into bodies, bodies that seek to persevere, or prolong their run.”25 There are two things to be said in response to Bennett’s equivocation. First: it’s understandable – predictable, almost. To return to the point that I flagged above, if one adopts a metaphysics in which all there is, ultimately, is activity construed as motion, one will not be able to get out of it the genuinely material entities, differentially endowed with qualitatively different kinds of powers, to which Bennett is committed. It is one thing to say that it is an essential feature of material objects that they are internally dynamic. It is quite another thing to claim (contra Spinoza, as it happens) that extension itself just is motion. At a minimum, such a view leaves behind nothing to be that which is in motion. If the sheer fact of motion is to be intelligible at all, therefore, it will have to be as an analytic abstraction only, and one with a very short half-life at that, since one could easily argue that movement cannot even be thought without reference to that which moves. It is very hard to see how such an abstraction could congeal into material bodies, let alone into ones with distinct and characteristic capabilities. Spinoza has a lesser version of this problem, insofar as what he calls bodies are thought to be constituted by the mere spatial arrangement of the attribute of God/Substance that he calls extension. Aristotle, by contrast, doesn’t – because Aristotelian substances are instantiations of form. In any case, it seems to me that Bennett is left with no choice but to tacitly embrace a richer metaphysics than the one that she explicitly affirms, if only to establish the range of qualitative properties that she needs at the level of actants. Second, and I think more important, the association with vitalism is simply unnecessary for the purposes of re-vitalizing causal bearers – to employ a term that I have used elsewhere.26 This is not to say that one couldn’t believe that inanimate entities are indeed infused with a life force, or even a “not-quite-human force capable of producing the new.” It’s just that the truth or falsity of vitalism – or of quasi-vitalism – is not decisive for the issue at hand, viz., the rejection of the dead world of mechanism. If one is independently convinced of the reality of a vitalist or quasi-vitalist life force, then one should certainly say so. But Bennett doesn’t propose vital materialism as a straight-forward ontological thesis. Rather, the suggestion is that it would behoove us to see the world in such terms. Once we are on the ground of rhetorical strategy, the costs of endorsing the existence of a life force become relevant to the discussion. Again, the costs are ones that must be borne, if one is committed to the truth of such a view. But vitalism is overkill for the purpose of rejecting mechanism. For one can be a dispositional realist without relying on the tradition or the tropes of vitalism at all. Instead of trying to explain dynamism by way of references to chi that then have to be retracted, entities can be seen to be active in virtue of possessing irreducibly dispositional properties. As argued in Chapter 5, such a view allows for all substances to be seen as causally potent, different only in the nature of their powers. As also noted, such an approach has already been extended into the philosophy of science (and social science), such that there is now a body of

112

Ontology Revisited

scientific essentialist (and critical realist) literature that has de-coupled mechanism as a philosophical thesis from the epistemic authority of scientific theory. If vitalism were thought to be true, of course, then the same would have to be done for it: it would have to be integrated into the philosophy of science and social science. But where the considerations are instrumental, there is no question but that dispositional realism does the job more neatly. It’s the attraction to Spinoza that most interests me though. Should Spinoza’s metaphysics be the ontology of choice for those in search of a dynamic, nondeterminist materialism, now that such talk is once again permissible? The answer simply has to be no. I’ll begin with what seems to me to be the most obvious point. Bennett, whom I am taking to have produced a crystallized expression of contemporary Left Spinozan political thought, is at pains to emphasize the theme of contingency, of indeterminacy. Even electrical current, as noted in the quotation above, can decide on the spot what it will do. Such voluntarism is just completely at odds with the ontology of Baruch Spinoza, who is unmistakably not only a determinist, but a necessitarian. This has nothing to do with Hegel’s alleged misreading of him. It is perfectly clear from the Ethics itself, the Ethics, Demonstrated in Geometrical Order. Although individual passages considered on their own can convey only a limited sense of an over-all theory, here is telling evidence of Spinoza’s complete lack of ambiguity on this point: Part 1, Proposition 29: “In nature there is nothing contingent, but all things are determined from the necessity of the divine nature to exist and act in a certain manner.”27 One who has undertaken to appropriate Spinoza for a non-deterministic materialism might respond that obviously they do not mean to embrace his determinism or his necessitarianism, only some other aspect of his metaphysics: for example, the vision that Bennett attributes to him of active, striving entities that form alliance-like aggregations in order to increase their powers. Such a move is trickier than it might seem, however. Especially, it may turn out that the part of a theory that one rejects is needed in order to sustain the part that one adopts. Here, for instance, there is good reason to think that Spinoza’s notion of conatus, of striving, is integrally related to his uncompromising determinism. But for the sake of argument, let’s imagine a reply to the effect that while a commitment to the contingency of extant assemblages runs contrary to a Spinozan metaphysics, the core notion of activity, of a vital rather than mechanistic materialism, does not. Vital materialists (and others) take from Spinoza only the idea that things “strive” to enhance their powers. Striving explains why bodies come together into assemblages in the first place, in addition to describing the internal dynamism of such entities. Indeed, Bennett might add, insofar as all actants consist essentially of motion, the view amounts to a perfectly recognizable variant of Spinoza’s own account of composite bodies. The question, then, is whether or not the Spinoza with whom Bennett, in this case, has presented us (though again, she is not alone) can be reconciled with the Spinoza who emerges from the text of the Ethics. I am not hopeful. The problem is this: Bennett, citing Deleuze and Guattari, has put forward an ontology that bottoms out in dynamism – a dynamism construed expressly as motion, but some-

Powers, Ontology and the Appeal to Spinoza 113 times implicitly as a quasi-spiritual, free-floating creativity or sui generis productivity. Spinoza, by contrast, proposes an ontology that bottoms out in a system of perfect, fixed, rational necessitation. His concept of striving – of a drive, inherent in things, to increase their power – is therefore not at all what it might appear to be. Let me try to explain why. Spinoza claimed that there is only one substance, God. God has an infinite number of “attributes,” he says, adopting and modifying the Cartesian terminology. Human beings, who are limited, can only experience God in two ways: either as extension or as thought. Bodies are taken to be “modes” of the former, modes of God when God is (or is cognized as) extension. For each body (each finite mode of extension), there is an exact corresponding idea in the parallel order of God-as-thought (or cognized as thought). The body and the idea are held to be identical, ultimately, insofar as they are modes of different attributes of the same substance, namely God. It is not obvious what it is that gives bodies their identities, on this picture; at a minimum, different bodies (a) are made up of differentially arranged parts, “join[ed] together to have certain effects,”28 and (b) follow (and do so necessarily) from different conditions (conditions that translate into different ideas, if and when God is experienced as the attribute of thought, rather than that of extension). Setting aside her reduction of extension itself to movement, Bennett’s notion of an actant is consistent with (a). But Spinoza’s own account of striving is connected also to (b), and to the rationalist framework with reference to which it must be understood. It’s important to be absolutely precise here. What things “strive” for, according to Spinoza, is to exist in a way that is determined as fully as possible by that in virtue of which their existence is conceptually necessary. Being entirely so determined Spinoza calls “acting.” Acting is mirrored, in turn, at the level of thought, by “adequate” ideas. The having of adequate ideas is pleasurable, Spinoza says. Also, things are only destroyed or harmed, he believes, when they are determined by something that is external to them, i.e., something that is not a part of their identity conditions. The notion of striving, in this philosophical context, has nothing to do with affirming dynamism over fixity. Rather, it is a way of expressing the idea that the order of rational determination that is derived from God in all directions at once, and that is God – it is a way of expressing the idea that that order, which is perfect, serves as the default for a kind of metaphysical inertia.29 Things “strive” to be what they already, necessarily and optimally are. Moreover, as Michael Della Rocca has brought out so nicely for contemporary readers, Spinoza takes causation itself to be that same relationship of conceptual necessitation which gives meaning to the notion of striving.30 Della Rocca observes that the striving in question is not psychological.31 But it’s not the exertion of a power, either. And that’s because causation, for Spinoza, is not a matter of activity, of doing. It is a matter of logic. When Hume declares that causes and effects are distinct, and that the latter are not contained within the former, it is Spinoza to whom the claim is directed. The problem for those who are taken with Spinoza, therefore, is not just that it may be impossible to jettison Spinoza’s determinism whilst retaining his concept of conatus. Even more fundamentally, it is

114

Ontology Revisited

just very hard to see where in Spinoza’s metaphysics one might find the real powers that would be needed to ground real striving. One might be tempted to think that God, at least, is properly powerful, on the Spinozan picture, and that it might be that bodies inherit dispositional properties from God.32 But it’s not a promising line of argument, I don’t think. At least, Spinoza insists that it is a mistake to conceive of God’s powers as being tantamount to those of an earthly king, only magnified.33 This is worth taking very seriously: not even God – who, as it happens, is co-terminus with every material entity – is an actant in the sense required by Bennett’s vital materialism. Not even God is a do-er. In sum, if one confronts the Ethics head on, the belief that Spinoza can help to deliver a realist, powers-based ontology begins to look a bit like wishful thinking. That Spinoza is not the historical figure for the job is ultimately a side point, however, relative to the argument that I’ve developed over the course of the last five chapters. I shall therefore leave its full elaboration for another day. Most relevant for my present purposes is that the turn to Spinoza, advisable or no, represents a welcome spark of interest amongst political theorists in ontology, and in powers specifically. In this vein, I want to conclude by offering some general thoughts about what it means, or ought to mean, for social and political theorists to revisit ontology, as I’ve put it. Let me begin with the point that I made above in relation to the concept of conatus, which is that one may not assume that an individual term can be lifted out of its original philosophical context. This is a more difficult claim to advance than it might seem. In some circles, it is likely to be received as a clichéd truism even whilst being ignored; in others it will be thought to express a conflation of logic with hermeneutics. Nonetheless I want to insist upon it. As we saw with Spinozan conatus, or striving, it may well be that the favored idea derives its content from other elements of the theory, and so does not mean what it might appear to mean at first glance. And even if we say that there are exceptions – that the idea of a genuine power, for example, is what it is regardless of whether one prefers Aristotle’s, Locke’s or even Reid’s over-all metaphysics – even so, there is reason to proceed with care. For one thing, while the level of abstraction sufficient to allow one to remain otherwise-uncommitted may sometimes provide one with a free-standing bit of thought, single bits of thought are not enough. In the case of the idea of a real power, for instance, we will immediately need to know what such powers are borne by, and why, and how we are to think about the expressions thereof. Moreover, even if we could get the isolated idea to do all of the important work on its own, the very fact of it being free-standing vis-à-vis other theoretical commitments suggests that it is not well-secured, either. Especially if one is analytically trained, one might counter that in fact it’s a virtue of a core concept that it be free-standing. If the idea of a real power is not entrenched in any given background theory, it will be equally available to the Lockean, the Reidian and the Aristotelian, say. But this, I think, lands us in a situation that is similar to the one we were in before. Perhaps some concepts may be analytically isolated in a manner that is not possible with respect to Spinoza’s notion of striving (vis-à-vis his necessitarianism). However, whatever virtue such

Powers, Ontology and the Appeal to Spinoza 115 a concept might have in terms of appealing to the widest possible audience is likely to be offset by how limited a role it will be consigned to play in an actual theory, give its necessary degree of abstraction. Once again we would need to ask follow-up questions, were we to make genuine philosophical headway. And the answers would immediately infuse the concept with further determinate content. The upshot of this in terms of appropriations of paradigmatic work in the history of philosophy is that such appropriations must be sensitive to the conceptual topography of the original theories. It is in the nature of the case, it would seem, that the constraints are different in this regard in philosophy than they are in the history of other forms of expression, e.g., visual art, music, literature – though perhaps one cannot so neatly excise single bits of thought or meaning from a larger framework in the context of those media either. At any rate, if one appeals to Spinoza for the idea of a power, one may do so only insofar as one is prepared to defend a metaphysics according to which what are called powers are actually relationships of conceptual necessitation. Conversely, if one wants powers that amount to activity, to doing, then one will need to draw from a thinker who understands powers in those terms. It is not enough to use the word; everything depends upon what can be meant by it, given the theory as a whole. A second general point concerns the question of what an attentiveness to metaphysics does and does not presume or entail. I began Chapter 1 by referring to the distinction that Aristotle makes between phronesis – wisdom concerning virtuous relations with others – and sophia, systematic knowledge of the essences of things. As Aristotle would have it, politics involves the former, philosophy the latter. In particular, ontology – literally: knowledge of being – does not translate into wise counsel, into knowledge of what course of action would be the expression of moral virtue in any given circumstance. Knowledge of the human form tells us only that engaging in virtuous relationships with others is one of the types of doing that constitutes flourishing for entities of our kind, and that a proper polis will have such excellence of activity as its purpose. Metaphysics cannot tell us what to do here, now, in this particular situation, faced with this particular set of options. The gap that Aristotle leaves between sophia and phronesis is appropriate, in my view. However, there is more to be learnt from Aristotle than that even he didn’t think it possible to get all the way from “is” to “ought.” While sophia may not yield phronesis, it is nevertheless true in Aristotle’s own case that his account of human beings, and of the nature of the polis, is broadly consistent with the ontology set out in the Physics and the Metaphysics. We find in the Metaphysics that entities are what they are in virtue of their form, and also that different kinds of entities have different kinds of characteristic powers, expressed in different kinds of characteristic activities. The underlying normative assumption is that, for any entity, it is good to exercise excellently those powers that are essential to its kind. Accordingly, the determinations to be made when it comes to human beings begin with the question “What is the nature of the human form?” And then, “What is the activity or activities that are therefore characteristic of our kind, at which it is (therefore) good to excel?” Followed by: “What are the structural political-economic conditions of possibility for

116

Ontology Revisited

flourishing so conceived?” Aristotle’s response is that it is the essence of our kind to have powers of rational consciousness, such that our behavior is that of a reflexive, contemplative substance, capable of abstraction, explanation and moral deliberation – rather than that of an purely appetitive or merely vegetative substance, or of a non-living substance. The characteristic activities of our kind are therefore precisely those the excellences of which are sophia and phronesis respectively. Though again, knowledge of the human form tells us only that, along with philosophy, the conducting of virtuous relationships with others is one of the kinds of doing that constitutes flourishing for human beings, and that a proper polis will have such excellence of activity as its purpose. What we learn by example from Aristotle, then – apart from the limitations of sophia – is two-fold. First, the ontology that is implicit in a social-political theory must be compatible with what its author claims to be true at the level of metaphysics. Aristotle himself, of course, was both a social-political theorist and a metaphysician, as were many other canonical philosophers. Most contemporary thinkers are not. But the point is not that one may not specialize. Rather, it is that the compatibility in question is not guaranteed. Social-political theories cannot be mixed and matched with any and all ontological commitments. It is a mistake, therefore, for social and political theorists to believe themselves to be disinterested parties vis-à-vis debates in metaphysics. Second, the metaphysics that a social-political theorist does endorse will have to be explanatorily adequate relative to her analysis at the level of the social and/or political. Here the important point is that a good social-political theory will have to rest upon an ontology that does not undercut its content. Humeanism, with its flip-book model of agency, is bound to fare poorly in this regard: at a minimum, it is likely to be a detriment to any account of social and political life that involves the exercise of real causal power(s), which I daresay all actually do, though I am happy to leave it at “most, including that of Hume himself.” Having reached such a conclusion, it seems to me that we are invited to articulate more systematically the metaphysical demands associated with the nature of social and political phenomena. To develop and defend such an account would require a separate book, but it is possible on the basis of the present discussion, I think, to offer a sketch of how things are likely to go. At the level of persons, an adequate metaphysics will have to be able to account for real individuals who are capable of real agency. Meeting this demand calls for an acceptance of real powers, borne by real substances who are able to bring about intended changes. It also looks as though substances who are so endowed will have to be understood to be both irreducibly conscious and metaphysically free, in the senses proposed in Chapter 5. At the level of sociology and political economy, meanwhile, an adequate metaphysics will have to allow for emergent wholes that do not reduce to their parts, and that are the bearers of emergent powers not held by their individual members considered singly. As suggested in Chapter 5, there is a further issue to do with whether or not emergent wholes even have parts, as we normally understand the term. If one is a proponent of Aristotelian, Hegelian or Marxist social-political theory, for example, one will have to adopt an ontology that

Powers, Ontology and the Appeal to Spinoza 117 permits individuals to retain their integrity as substances even whilst constituting the “matter” of an irreducibly relational entity such as the polis (as even Aristotle does) or of other structural phenomena. Also depending upon one’s social theory, one may have to allow for sociological wholes that are collective subjects, i.e., bearers of mental powers. Note that these two variables are independent: Rouseau’s Sovereign is the bearer of a General Will, but it arguably has no parts. Conversely, the Aristotelian polis clearly has parts but is arguably not a seat of intentionality. It may seem to some readers that I have got the order of explanation backwards, in suggesting that an adequate metaphysics will have to account for features of the world as they are experienced and/or theorized at the levels of psychology and sociology or political-economy. We – or at least some of us – are accustomed to thinking that it is up to metaphysicians to tell us what every-day phenomena really are, and whether or not they actually exist. But philosophy, no less than any other form of inquiry, must do epistemic justice to its object. A metaphysics of agency that cannot tell a compelling story about the doings of people, acting singly or collectively, or of the social relations in which we find ourselves, lays no claim to our intellectual allegiance. Luckily, after a very long run, one such metaphysics is no longer sacrosanct.

Notes

1 The Myth of Metaphysical Neutrality 1 See, especially, articles collected in Charles Taylor, Philosophical Papers, Vols. 1 and 2 (Cambridge: Cambridge University Press, 1985). 2 For example, in Brian Ellis, The Philosophy of Nature: A Guide to the New Essentialism (Montreal and Kingston: McGill-Queen’s University Press, 2002), and Scientific Essentialism (Cambridge: Cambridge University Press, 2001). 3 Ellis, op. cit., note 2, The Philosophy of Nature, pp. 59–60. 4 For discussions of Kant and dispositional realism, see Roy Bhaskar, A Realist Theory of Science (Atlantic Highlands, NJ: Humanities Press, 1978), and Ruth Groff, Critical Realism, Post-Positivism and the Possibility of Knowledge (New York: Routledge, 2004). 5 Some may want to say that an essence is something over and above this. I return to the issue below. 6 John Locke, An Essay Concerning Human Understanding (abridged and with notes by A. S. Pringle-Pattison) (Ware, Herts: Wordsworth Editions Limited, 1998). See in particular Book 111, chapters 3 and 6. 7 See Irving Copi, “Essence and Accident,” Journal of Philosophy, Volume 51, Number 23, 1954, pp. 706–19. 8 Thanks to Noah Efron, for first putting it to me in just this way. 9 Though Alexander Bird, for example, distinguishes powers from what he sometimes calls mere dispositions, and does not define either of them as I have, in terms of activity. See Alexander Bird, “Limitations of Power”, in Powers and Capacities in Philosophy: The New Aristotelianism (eds., John Greco and Ruth Groff) (New York: Routledge, forthcoming). Stephen Mumford has suggested that powers so construed are not properly powers (Stephen Mumford, “The Power of Power,” in Powers and Capacities in Philosophy (eds., John Greco and Ruth Groff). 10 Ellis, op. cit. note 2, The Philosophy of Nature, p. 60. 11 Thanks very much to John Greco, for a helpful conversation regarding these distinctions. 12 D. H. Mellor, 1974, The Philosophical Review 83: 157–81, cited in Ellis, The New Naturalism. 13 It is not clear that Alexander Bird, despite his commitment to what he calls “potencies,” believes in powers so understood. See Alexander Bird, Nature’s Metaphysics: Laws and Properties (Oxford: Oxford University Press, 2007). 14 I don’t mean by this that causes necessarily bring about that which they have the power to bring about, only that when they do, they do so in virtue of what they are. 15 One might think that Hume’s inductive circle holds for epistemic reasons alone: i.e., one might think that the commitment to kinds itself necessarily presupposes a thesis about regularity. It is not clear that this is so, though this is not an argument that I can develop here.

Notes 119 16 Sections of the following analysis appeared in Ruth Groff, “Introduction,” Journal for Critical Realism, Volume 8, Number 3, 2009 (special issue on causal powers), pp. 267–76. 17 In Scientific Essentialism (op. cit, note 2), Ellis held that powers are a kind of property, viz., dispositional (as opposed to categorical), and that they in turn constitute the essences of what he called process kinds. 18 Bhaskar, op. cit., note 4, A Realist Theory of Science, p. 171. 19 Ibid., p. 212. 20 Ibid., p. 174. 21 Anjan Chakravartty, personal correspondence, November 2008. See also, Anjan Chakravartty, “Inessential Aristotle: Powers Without Essences,” in Revitalizing Causality (ed., Ruth Groff)(London: Routledge, 2008) and A Metaphysics for Scientific Realism: Knowing the Unobservable (Cambridge: Cambridge University Press, 2007). 22 Stephen Mumford sees natural kinds as “object types, existing only in . . . objects, just as a property or attribute exists only in its instances. These kinds can be characterized by attributes but a further assumption that any of these characterizing attributes is essential is neither required nor independently motivated.” See Stephen Mumford, “Kinds, Essences, Powers,” Ratio, Volume 18, Number 4, 2005, p. 435. 23 Heil says there that there are such things as kinds and essences in the heuristic notion of a “Locke world,” but maintains (personal correspondence) that the kinds are nominal, and that he does not consider himself to be an essentialist. Cf. John Heil, “Kinds and Essences,” Ratio, Volume 18, Number 4, 2005, pp. 405–19. 24 See Ruth Groff, “Getting Past Hume in the Philosophy of Social Science,” in Causality in the Sciences (eds., Phyllis McKay Illari, Federica Russo and Jon Williamson) (Oxford: Oxford University Press, 2011). 2 Hume 1 David Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals (reprinted from the 1777 edition with Introduction and Analytical Index by L. A. Selby-Bigge), 3rd edn., with text revised and notes by P. H. Nidditch (Oxford: Clarendon Press, 1975), pp. 29–30. 2 Ibid., p. 30. 3 David Hume, A Treatise of Human Nature (reprinted from the original edition in three volumes and edited with an Analytical index by L. A. Selby-Bigge) (Oxford: Clarendon Press, 1975), p. 161. 4 Hume, op. cit., note 1, Enquiry, p. 29. 5 Ibid., p. 76. 6 Ibid., p. 75. 7 Hume, op. cit., note 3, Treatise, pp. 165–6. 8 Hume, op. cit., note 1, Enquiry, p. 76. 9 See Georges Dicker, Hume’s Epistemology and Metaphysics: An Introduction (New York: Routledge, 1998) for a useful discussion. 10 Such a move is made possible by off-loading the modal force of the counterfactual onto logically possible worlds, leaving this world as Hume described it. 11 Thomas Reid, Essays on the Active Powers of Man, Essay I, chapter 4, in The Works of Thomas Reid, Vol. 2 (Chestnut Hill, MA: Adamant Media Corporation, 2005), p. 521. 12 Rupert Read and Kenneth A. Richman (eds.), The New Hume Debate, revised edition (New York: Routledge, 2008). 13 The footnote reads: “The word, Power, is here used in a loose and popular sense. The more accurate explication of it would give additional evidence to this argument. See Sect. 7.” Hume, op. cit., note 1, Enquiry, p. 33. The more accurate explication of Section 7, in turn, is therein summarized by Hume as follows: “We have sought in vain for an idea of power or necessary connexion in all the sources from which we could

120

14 15

16

17 18

19 20 21 22 23 24 25 26

27 28 29 30 31 32 33 34 35 36 37 38 39

Notes

suppose it to be derived. It appears that . . . we never can . . . discover any thing but one event following another; without being able to comprehend any force or power by which the cause operates, or any connexion between it and its supposed effect.” Hume, op. cit., note 1, Enquiry, pp. 73–4. Dicker, op. cit., note 9, see especially chapter 4. Given this, it is no surprise that New Humeans hold that Hume didn’t really mean what he said about meaning, either. See Kenneth Richman, “Debating the New Hume” and Kenneth Winkler, “The New Hume”, in The New Hume Debate, revised edition (eds., Rupert Read and Kenneth A. Richman) (New York: Routledge, 2008). The correct response to the Humean at this point is to ask if her powers can do what non-reducible powers can do. To return to my outlandish example in the footnote above, if unicorns exist but reduce to grapes, it is hard to see how they could be such that they can fly and have a sharp horn, good for doing the sorts of things that sharp horns can do. If powers are really something passive, it’s hard to see how they can be active. Walter Ott, Causation and Laws of Nature in Early Modern Philosophy (Oxford: Oxford University Press, 2009). There were differences among proponents as to whether or not causal interactions involve any additional causal input from God. On the end of the spectrum closest to Aristotle, the answer was no; substances are, as Ott terms it, causally autonomous. This is the conservationist view. The concurrentist view, by contrast, was that causation requires both the (God-given) powers of substances and the on-going intervention of God. Common to the range of bottom-up positions, however, was the idea that the material world is imbued with vis, with potency. Though Ott seems to believe that a more robust version of this is Locke’s move. In this sense, contemporary analytic Humeans who want their powers while denying them are perhaps best seen as thinned-out Lockeans. Hume, op. cit., note 3, Treatise, p. 409. Hume, op cit., note 1, Enquiry, p. 82. Hume, op. cit., note 3, Treatise, p. 255. Ibid., p. 252. Ibid., p. 259. Ibid. It is not obvious what the criterion would be for an impression being “of” something. At best, it would seem to depend upon how unified such impressions can be made to be, by the imagination. An upshot of this is that it is not clear what the difference is, ontologically, for Hume, between objects that exist, such as billiard balls, and objects that do not exist, such as selves. Hume, op. cit., note 3, Treatise, p. 261. Ibid. Ibid., pp. 408–9. Hume, op. cit., note 1, Enquiry, p. 96. To prevent any Humean finessing of this point, I will add the qualifier “irreducible”; by a dispositional property, I mean one that is irreducibly dispositional. Alasdair MacIntyre, Hume on “Is” and “Ought,” in MacIntyre, Against the Self-Images of the Age: Essays on Ideology and Philosophy (Notre Dame, Indiana: University of Notre Dame Press, 1978). Hume, op. cit., note 3, Treatise, p. 490. Ibid., pp. 501–2. Ibid., p. 502. Ibid., p. 527. Ibid. Ibid. As an aside, Hume observes that the circularity of his thinking (defining justice in terms of property, then property in terms of justice) lends credence to the idea that justice is

Notes 121

40 41 42 43 44 45 46 47 48 49 50 51

52 53 54 55 56 57 58 59 60 61 62 63 64 65

66 67 68 69 70

an artificial virtue. “This deceitful method of reasoning,” he says, “is plain proof that there are contain’d in the subject some obscurities and difficulties, which we are not able to surmount, and which we desire to evade by this artifice.” Hume, op. cit., note 3, Treatise, p. 528. Ibid., p. 506. Ibid., p. 503. Ibid., pp. 503–4. Ibid., p. 504, fn. Ibid., p. 509 and fn. 1. John Locke, The Second Treatise of Government (ed. Peter Laslett) (Cambridge: Cambridge University Press, 1988), chapter 5, paragraph 28. Hume, op. cit, note 3, Treatise, p. 510 and fn. 2, pp. 509–10. Ibid., p. 510. Ibid., p. 511. Ibid., p. 526. As John Hawthorne points out, there are uses of “do” and “does” that do not pose a significant problem for the Humean; the problem is most pointed in the case of non-metaphorical uses of causal verbs. Personal conversation, September 2010. Brian Ellis talks about the distinction between categorical and dispositional properties in just these terms. See Brian Ellis, “The Categorical Dimensions of the Causal Powers,” in Properties, Powers and Structures: Issues in the Metaphysics of Realism (eds., Alexander Bird, Brian Ellis and Howard Sankey) (Abingdon: Routledge, 2012). Hume, op. cit., note 3, Treatise, p. 455. Ibid., p. 415. Thanks very much to John Greco for conversation about this point, especially for his effort to rationally reconstruct Hume’s position. Hume, op. cit., note 3, Treatise, p. 457. Ibid. Ibid., p. 458. Ibid. Ibid., p. 462. Thanks to John Greco for conversation about this point. Hume, op. cit., note 3, Treatise, p. 252. Ibid., p. 411. Ibid., p. 575. Ibid., p. 286. In fact, flirting with circularity, he actually goes further, claiming that “nature has given to the organs of the human mind, a certain disposition fitted to produce a peculiar impression or emotion, which we call pride: To this emotion she has assign’d a certain idea, viz. that of self, which it never fails to produce.” Hume, op. cit., note 3, Treatise, p. 287. But he can only be teasing, of course, since when he uses the verb “to produce,” it can mean only “be followed by.” Thus pride is both preceded by the idea of the self and followed by it. Hume, op. cit., note 3, Treatise, p. 288. Ibid., p. 320. Ibid., p. 317. Ibid., p. 318. Stephen Mumford, Laws in Nature (Abingdon and New York: Routledge, 2004), p. 21.

3 J. S. Mill 1 John Stuart Mill, A System of Logic, Ratiocinative and Inductive: Being a Connected View of the Principles of Evidence and the Methods of Scientific Investigation, Book 3,

122

2 3 4 5

6 7 8 9 10 11

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

29 30 31 32 33 34

Notes

chapter v, section 1; in Collected Works of John Stuart Mill, Vol. 7 (ed. J. M. Robson) (Indianapolis, IN: Liberty Fund, 2006), p. 324. Mill, op. cit. note 1, Logic, Book 3, chapter v, section 2, p. 327. Ibid., Book 3, chapter v, section 6, pp. 338 and 339. Mill notes in a footnote, here, that by the present constitution of things he means “the ultimate laws of nature.” Ibid., Book 3, chapter v, section 1, p. 308. Ibid., Book 3, chapter v, section 3, p. 328. And again, “[t]he cause, then philosophically speaking, is the sum total of the conditions, positive and negative taken together; the whole of the contingencies of every description, which being realized, the consequent invariably follows.” Ibid., p. 332. Ibid., Book 3, chapter v, section 2, p. 326 and footnote g. Ibid., Book 1, chapter iii, section 9, p. 66. Ibid., Book 3, chapter v, section 4, p. 335. Ibid., Book 3, chapter v, section 4, p. 337. Ibid. John Stuart Mill, A System of Logic, Ratiocinative and Inductive: Being a Connected View of the Principles of Evidence and the Methods of Scientific Investigation, Book 4, chapter vii, section 1; in Collected Works of John Stuart Mill, Vol. 8 (ed. J. M. Robson) (Indianapolis, IN: Liberty Fund, 2006), p. 712. Ibid. Mill, op. cit., note 11, Logic, Book 4, chapter vii, section 1, p. 713. Ibid., p. 714. Ibid. Ibid., pp. 718–19. John Stuart Mill, An Examination of Sir William Hamilton’s Philosophy (London: Longmans, Green, Reader, and Dyer, 1872), chapter XI, p. 233. If one has any lingering thought that this can’t be so, that it must be that Mill believes that there is a material source of sensations, ibid., “An Appendix to the Two Preceding Chapters.” Mill, op. cit., note 17, Examination, “Appendix to the Two Preceding Chapters,” p. 256. Mill, op. cit., note 17, Examination, chapter XII, p. 242. Ibid., p. 243. Alan Ryan, John Stuart Mill (New York: Pantheon Books, 1970), p. 100. Mill, op. cit., note 17, Examination, “Appendix to the Two Preceding Chapters,” p. 262. Mill, op. cit., note 11, Logic, Book 6, chapter ii, section 2, pp. 836–7. Ibid., Book 6, chapter vii, section 1, p. 879. John Stuart Mill, Utilitarianism, in Utilitarianism, Liberty and Representative Government (New York: E. P. Dutton, 1951), p. 43. Aristotle, Politics (ed. and revised, R. F. Stalley; trans., Ernest Barker) (Oxford: Oxford University Press, 1995), Book 3, chapter 9, 1280b6, p. 105. In Book 6, chapter v of the Logic, and passim, Mill classifies ethology, the science of character, as a derivative, mediating science of the individual; in Book 6, chapter ix, however, he makes reference to “political ethology,” which he counts as an element of social science. See, for example, Mill, op. cit., note 11, Logic, Book 6, chapter ix. Ibid., chapter iv. G. W. Smith, “The Logic of J. S. Mill on Freedom,” Political Studies, Volume 28, Number 2, 1980, pp. 238–52. Mill, op. cit., note 26, Utilitarianism, p. 11. Ibid., p. 13. John Stuart Mill, On Liberty, in Utilitarianism, Liberty, and Representative Government (New York: E. P. Dutton, 1951), pp. 155–6.

Notes 123 35 36 37 38 39 40 41 42 43 44 45 46 47 48

49 50 51 52

Ibid., p. 159. Ibid., p. 160. Ibid., p. 97. John Stuart Mill, Representative Government, in Utilitarianism, Liberty and Representative Government (New York: E. P. Dutton, 1951), p. 272. Mill, op. cit., note 34, On Liberty, p. 156. Ibid., p. 158. Ibid. Ibid., p. 156. Ibid., p. 160. Mill, op. cit., note 38, Representative Government, p. 287. Ibid., p. 288. Ibid., p. 330. Mill, op. cit., note 11, Logic, Book 6, chapter ii, p. 841. “It is not, therefore, so much the understanding that constitutes the specific difference between the man and the brute, as the human quality of free agency. Nature lays her command on every animal, and the brute obeys her voice. Man receives the same impulsion, but at the same time knows himself at liberty to acquiesce or resist: and it is particularly in his consciousness of this liberty that the spirituality of his soul is displayed. For physics may explain, in some measure, the mechanism of the senses and the formation of ideas; but in the power of willing or rather of choosing, and in the feeling of this power, nothing is to be found but acts which are purely spiritual and wholly inexplicable by the laws of mechanism.” Jean-Jacques Rousseau, “A Discourse on a Subject Proposed by the Academy of Dijon: What is the Origin of Inequality Among Men, and is it Authorized by Natural Law,” in The Social Contract and Discourses (trans., G. D. H. Cole)(London: Dent – Everyman’s Library, 1973), p. 52. Richard Sorabji, Necessity, Cause and Blame: Perspectives on Aristotle’s Theory (Chicago, IL: The University of Chicago Press, 1980), especially chapters 2 and 14. Mill, op. cit., note 26, Utilitarianism, p. 41. Mill, op. cit., note 38, Representative Government, p. 291. One may or may not also believe that the collective good must be good for individuals considered singly; I think that Mill does not have a rule for this, other than that he seems to believe that individuals who are motivated only by their “private partialities” are morally impoverished. (One is tempted to come right out and use Rousseau’s language of the “particular will.”)

4 Kant and the Frankfurt School 1 In saying this, I am not talking about what Kantian phenomenal objects are like “in and of themselves,” as the phrase goes. To do so would be to ask what objects of experience would be like if we were to consider them as something other than potential objects of experience. By contrast, I’m drawing attention to what the objects that we do experience are like, on Kant’s account. And what they are like is this: they are not bearers of real powers, powers that could underwrite causation; they are not causally connected to each other in any naturally necessary way; they are not as or what they are in virtue of having real essences. 2 Immanuel Kant, Groundwork of the Metaphysics of Morals (trans., H. J. Paton) (New York: Harper & Row, 1964), p. 88. 3 My thinking on this point has benefitted from Angelica Nuzzo, Kant and the Unity of Reason (West Lafayette, IN: Purdue University Press, 2005), chapter 3. 4 The norms related to freedom that Kant invokes in relation to the state, and more generally in relation to action that is deemed political, don’t seem to be the same as those associated with the practical activities of individuals qua moral agents – though

124

5

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

28 29 30 31 32 33

34 35

Notes

not all commentators would agree with this assessment. At a minimum, some maintain that so-called “political freedom” is no different in Kant’s view from moral freedom (both being governed by the law of freedom rather than the law of causality); others hold that political freedom may be heteronomous, but that this does not undermine the coherence of Kant’s position. There are also those who would emphasize Kant’s hope that a realm of ends will evolve over time, out of heteronomous conditions. Katrin Flikschuh, Kant and Modern Political Theory (Cambridge: Cambridge University Press, 2000). See her Freedom: Contemporary Liberal Perspectives (Malden, MA: Polity Press, 2007) for a useful study of the connections between contemporary liberal theory and broad positions within the free will debate in philosophy to be discussed in Chapter 5. Herbert Marcuse, Eros and Civilization: A Philosophical Inquiry Into Freud (Boston, MA: Beacon Press, 1966), p. 228. Ibid., p. 35. Ibid. Ibid., see chapter 7. Ibid., p. 142. Ibid. Ibid., p. 143. Ibid., p. 205. Ibid. Ibid., p. 201. Ibid., p. 202. Ibid. Ibid., p. 211. Ibid., p. 218. Ibid., p. 224. Ibid., p. 177. Ibid. Ibid., p. 178. Ibid. and fn. Ibid., p. 178. Herbert Marcuse, “The Concept of Essence,” in Negations: Essays in Critical Theory (with translations by Jeremy J. Shapiro) (Boston, MA: Beacon Press, 1968). Jean-Jacques Rousseau, “A Discourse on a Subject Proposed by the Academy of Dijon: What is the Origin of Inequality Among Men, and is it Authorized by Natural Law,” in The Social Contract and Discourses (trans., G. D. H. Cole) (London: Dent – Everyman’s Library, 1973), p. 52. Max Horkheimer, “Materialism and Morality,” in Between Philosophy and Social Science: Selected Early Writings (trans., G. Frederick Hunter, Matthew S. Kramer and John Torpey) (Cambridge, MA: MIT Press, 1993), pp. 19–20. Ibid., p. 25. “Reason cannot become transparent to itself as long as men act as members of an organism which lacks reason.” Max Horkheimer, “Traditional and Critical Theory,” in Critical Theory: Selected Essays (New York: Continuum, 1989), p. 208. Horkheimer, op. cit., note 28, “Materialism and Morality,” p. 34. Ibid., p. 35. “Materialism,” he writes, “has in common with positivism that it acknowledges as real only what is given in sense experience.” Max Horkheimer, “Materialism and Metaphysics,” Critical Theory: Selected Essays, New York: Continuum, 1989, p. 42. Ibid., p. 24. Ibid., p. 42.

Notes 125 36 Theodor W. Adorno, Negative Dialectics (trans., E. B. Ashton) (New York: Continuum, 1992), p. 5. 37 Perhaps more generally, Adorno says, “The question of freedom does not call for a simple Yes or No, but a theory that rises above society as well as above the individuals existing in it.” Theodor W. Adorno, History and Freedom (ed., Rolf Tiedemann; trans., Rodney Livingstone) (Cambridge: Polity Press, 2006), Lecture 28, p. 264. 38 See Adorno’s sustained, very clear discussion of this point in the later lectures in Theodor Adorno, History and Freedom: Lectures 1964–65 (ed., Rolf Tiedemann; trans., Rodney Livingstone) (Cambridge: Polity Press, 2006). These particular ideas are also emphasized in Part Three, section I of Negative Dialectics, though they figure throughout Adorno’s work. 39 Adorno, op. cit., note 36, Negative Dialectics, p. 249. 40 Adorno, op. cit., note 37, History and Freedom, Lecture 21, p. 197. 41 Ibid., p. 198. See also, Theodor W. Adorno, Problems of Moral Philosophy (ed. Thomas Schroder; trans., Rodney Livingstone) (Stanford, CA: Stanford University Press, 2000), Lecture 12. 42 See especially the later lectures of Adorno, op. cit., note 37, History and Freedom. 43 Theodor W. Adorno, Kant’s Critique of Pure Reason (ed. Rolf Tiedemann; trans., Rodney Livingstone) (Stanford, CA: Stanford University Press, 2001), Lecture 13, p. 140. Kant’s rejection of a more realist, object-centered account, Adorno explicitly likens to Hume’s: “[t]he criticism he makes here is the same criticism as Hume makes of the traditional concept of causality.” Ibid., p. 141. 44 Adorno, op. cit., note 37, History and Freedom, Lecture 27, p. 253. 45 Ibid. 46 Adorno, op. cit., note 36, Negative Dialectics, p. 251. 47 Marcuse, op. cit., note 6, Eros and Civilization, p. 178. 48 Adorno, op. cit. note 43, Kant’s Critique of Pure Reason, p. 98. 49 Adorno, op. cit., note 41, Problems of Moral Philosophy, p. 50. 50 Adorno, op. cit., note 36, Negative Dialectics, p. 248. 5 Agents, Powers and Events 1 Thanks to John Greco for help with this general formulation of the free will problematic. See also Roderick Chisholm, “Human Freedom and the Self,” reprinted in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). 2 I don’t mean to suggest that the view is that the laws themselves hold of necessity, only that given what they are, there is a relationship of necessity between causes and their effects. Thanks to Jonathan Lowe for pointing out this potential ambiguity. 3 Chisholm, op. cit., note 1, “Human Freedom and the Self,” p. 26. Chisholm frames the debate in this way despite the fact that he himself undertakes to defend agent causation; that this would be so is the very point that I want to make. 4 David Wiggins, “Towards a Reasonable Libertarianism”, reprinted in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). 5 Ibid., fn. 16. 6 Kai Nielson, “The Compatibility of Freedom and Determinism,” reprinted in Free Will (ed., Robert Kane) (Malden, MA: Blackwell Publishing, 2002), p. 43. 7 Peter Van Inwagen, An Essay on Free Will (Oxford: Oxford University Press, 1983), p. 65. 8 Ibid., p. 16. 9 David Lewis, “Are We Free to Break the Laws?”, reprinted in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). 10 Of course, I do not mean to suggest by this that every individual philosopher involved in the free will debate is a Humean.

126

Notes

11 Thanks to John Greco for raising the question of triviality. 12 Thanks to Stephen Mumford for reminding me that one could attempt to defend the principle on the basis of the constant conjunction of A & B, B & C – though, as Mumford notes, Hume himself might have trouble with the “contiguity” requirement, i.e., that causes be contiguous with their effects. 13 Thomas Reid, Essays on the Active Powers of Man, Essay IV, chapter 2, in The Works of Thomas Reid, Vol. 2 (Chestnut Hill, MA: Adamant Media Corporation, 2005), p. 603. 14 Reid contrasts active powers with what he terms speculative powers, by which he means powers of speculation, powers of speculating: “The powers of seeing, hearing, remembering, distinguishing, judging, reasoning, are speculative powers; the power of executing any work of art or labor is active power.” Ibid., p. 515. William Hamilton maintains, in an editorial footnote to the work, that Reid should have contrasted active powers with passive powers. Reid, however, explicitly criticizes Locke for using the term “passive power,” charging that the concept of a “powerless power” is “a contradiction in terms.” Ibid., p. 519. Here, in another footnote, Hamilton defends Locke but grants that “Reid understands by power merely Active Power, Efficacy, Force, Vis; and in this exclusive sense Passive Power is certainly a ‘contradiction in terms.’ ” Ibid. 15 Reid, op. cit., note 13, The Works of Thomas Reid, Introduction, chapter 1, p. 515. 16 Ibid., Essay IV, chapter 2, p. 603. 17 Ibid., Introduction, chapter 5, p. 523. 18 Cf. Brian Ellis’ depiction of dispositional properties as those that affect what their bearers can do, in contrast to categorical properties, which affect only what their bearers are. Brian Ellis, “The Categorical Dimensions of the Causal Powers,” in Properties, Powers and Structures: Issues in the Metaphysics of Realism (eds., Alexander Bird, Brian Ellis and Howard Sankey) (Abingdon: Routledge, 2012). 19 Roderick Chisholm, “Agents, Causes and Events: The Problem of Free Will,” in Agents, Causes and Events (ed., Timothy O’Connor) (New York: Oxford University Press, 1995), p. 95. 20 Robert Kane, “Some Neglected Pathways in the Free Will Labyrinth,” in The Oxford Handbook of Free Will (ed., Robert Kane) (Oxford; Oxford University Press, 2002), p. 416. 21 Ibid., p. 415. 22 Ibid., p. 426. 23 Ibid. 24 Randolph Clarke, “Toward a Credible Agent-Causal Account of Free Will,” in Agents, Causes and Events (ed. Timothy O’Connor)(New York: Oxford University Press, 1995). 25 Ibid., p. 208. 26 As John Greco has observed, the point can be made in an epistemic register, too, since empiricists hold to an epistemology that mirrors the ontology of event causation, i.e., one in which to explain is to show how a sequence of events comes under a law – not how an agent and an event come under a law. Agent causation is thus literally inexplicable for the Humean, on epistemic grounds alone. John Greco, personal conversation, May 2011. 27 Jaegwon Kim, Philosophy of Mind (Boulder, CO: HarperCollins, 1996). 28 Thanks to John Heil for conversation on this point. 29 Donald Davidson, “Mental Events (1970),” in Essays on Actions and Events (New York: Oxford University Press, 1980), p. 208. 30 In more technical language: there is token identity between mental and physical events – which, for Davidson, allows for causal interaction – but there is no type identity, which allows for the view to be characterized as non-reductive. 31 Kim, op. cit., note 27, Philosophy of Mind, pp. 227–8.

Notes 127 32 Kim, op. cit., note 27, Philosophy of Mind, pp. 232–3. 33 For a meticulous discussion of this point as it pertains to the philosophy of mind, see Charles H. Kahn, “Aristotle vs. Descartes on the Concept of the Mental,” in Metaphysics, Soul, and Ethics in Ancient Thought: Themes from the Work of Richard Sorabji (ed., Ricardo Salles) (Oxford: Clarendon Press, 2005). 34 My thinking on this point is informed by Stephen Mumford and Rani Anjum, Getting Causes from Powers (Oxford: Oxford University Press, 2011), chapter 7. I do not mean to suggest that they would agree with me, though they might. 35 Thanks to Scott Berman for helping me to clarify my thinking on this point. 36 I don’t mean to say that Humeanism logically entails the claim that only basic particles exist. Humeanism is associated with atomism, in my view, because the fundamental units of the so-called Humean mosaic, at whatever level of complexity they are fixed, will be discrete, atomistic bits, bits that do not combine to form relational wholes. 37 Analytic readers might find it useful to think about the significance of arrangement in terms of the concept of “special composition”; the idea is that certain arrangements of parts give rise to new, non-reducible, relational entities, i.e., wholes. Thanks to Alexander Bird for this suggestion. 38 For a useful critique of an eliminativist version of this type of position, and its reliance on the “duly arranged” stipulation, see Jonathan Lowe, “In Defense of Medium-Sized Specimens of Dry Goods,” Philosophy and Phenomenological Research, Volume 68, 2003, 704–10. 39 Timothy O’Connor and Jonathan D. Jacobs, “Emergent Individuals,” The Philosophical Quarterly, Volume 53, Number 213, 2003, p. 547. 40 E. Jonathan Lowe, personal correspondence, June 2011. See also, E. Jonathan Lowe, A Survey of Metaphysics (Oxford: Oxford University Press, 2002). 41 O’Connor and Jacobs, op. cit., note 39. 42 Ibid., p. 541. 43 Ibid., pp. 541–2. 44 Ibid., p. 542. 45 Ibid. 46 Ibid. 47 Ibid., p. 543. 48 I don’t mean the term simple in a technical sense; I mean it to include O’Connor and Jacobs’ idea of individuals as exhibiting “objective, substantial unity.” O’Connor and Jacobs, op. cit., note 39, p. 547. 49 This is a worry expressed by John Greco. Personal conversation, May 2011. 6 Metaphysics and the Capabilities Approach 1 Martha Nussbaum, Creating Capabilities: The Human Development Approach (Cambridge, MA: The Belknap Press of Harvard University Press, 2011). 2 Ibid., p. 18. 3 Ibid., pp. 33–4. 4 Ibid., p. 20. 5 Ibid. 6 Ibid., p.23. 7 Ibid., p. 21. 8 John Greco has raised this possibility in conversation. 9 Nussbaum, op. cit., note 1, Creating Capabilities, p. 31. 10 Ibid., p. 137. 11 For an early, excellent treatment of this issue, see Roy Bhaskar, A Realist Theory of Science (Atlantic Heights, NJ: Humanities Press, 1978). 12 Thanks to Alexander Bird for conversation on this point.

128

Notes

13 Nussbaum, op. cit., note 1, Creating Capabilities, p. 130. For a more detailed discussion, see Martha Nussbaum “Kant and Stoic Cosmopolitanism,” Journal of Political Philosophy, Volume 5, Number 1, 1997. 14 Ibid., p. 161. 15 Ibid., p. 162. 16 Ibid., p. 28. 17 Ibid., p. 29. 18 Ibid. 19 Ibid., p. 30. 20 Ibid., p. 31. 21 Ibid. 22 Ibid., p. 39. 23 Ibid., p.35. 24 Ibid. 25 Ibid., pp. 25–6. 26 Thanks to John Greco for this way of expressing the point. 27 Thanks again to conversation with John Greco. 28 Nussbaum, op. cit., note 1, Creating Capabilities, p. 124. 7 Powers, Ontology and the Appeal to Spinoza 1 Jane Bennett, Vibrant Matter: A Political Ecology of Things (Durham, NC: Duke University Press, 2010), p. 57. 2 Ibid. 3 Ibid., p. 58. 4 Ibid. 5 Ibid., p. 23. 6 Ibid., p. 34. 7 Ibid., p. 20. 8 Ibid., p. 23. 9 Ibid., p. 32. 10 Ibid., p. 24. 11 Ibid., p. 31. 12 Ibid., p. 108. 13 Ibid., p. 101. 14 Ibid. 15 Ibid., p. 106. 16 Ibid. 17 Ibid., pp. 108–9. 18 Ibid., p. 29. 19 Ibid. 20 Ibid., p. 103. 21 Ibid., p. 28. 22 Ibid., p. 36. 23 Ibid., p. 61. 24 Ibid., p. 117. 25 Ibid., p. 118. 26 Ruth Groff (ed.), Revitalizing Causality: Realism About Causality in Philosophy and Social Science (New York: Routledge, 2008). 27 Benedict Spinoza, Ethics (trans., W. H. White and A. H. Stirling) (Ware, Herts: Wordsworth Editions Limited, 2001), p. 28. 28 Michael Della Rocca, Spinoza (New York: Routledge, 2008), p. 107. 29 Thanks very much to Scott Raglan for a number of extremely helpful conversations about Spinoza, including reflection on this point.

Notes 129 30 Della Rocca, op. cit., note 28, Spinoza, chapter 2, especially pp. 43–4. Thanks to Scott Ragland and with Eric Schliesser for conversation on this point. 31 Ibid., 147. 32 Thanks again to Scott Ragland. 33 Spinoza, op. cit., note 27, Ethics. See especially Part I, Appendix.

Bibliography

Adorno, Theodor W., History and Freedom (ed., Rolf Tiedemann; trans., Rodney Livingstone) (Cambridge: Polity Press, 2006). Adorno, Theodor W., Kant’s Critique of Pure Reason (ed., Rolf Tiedemann; trans., Rodney Livingstone) (Stanford, CA: Stanford University Press, 2001). Adorno, Theodor W., Problems of Moral Philosophy (ed., Thomas Schroder; trans., Rodney Livingstone) (Stanford, CA: Stanford University Press, 2000). Adorno, Theodor W., “Subject and Object,” in The Essential Frankfurt School Reader (eds., Andrew Arato and Eike Gebhardt) (New York: Continuum, 1993). Adorno, Theodor W., Negative Dialectics (trans., E. B. Ashton) (New York: Continuum, 1992). Aristotle, Politics (ed. and revised, R. F. Stalley; trans., Ernest Barker) (Oxford: Oxford University Press, 1995). Aristotle, Metaphysics, in The Basic Works of Arisotle (ed., Richard McKeon) (New York: Random House, Inc./Modern Library, 2001). Aristotle, Nicomachean Ethics (trans., Martin Ostwald) (Indianapolis, IN: Bobbs-Merrill, 1962). Ayers, Michael, Locke: Epistemology & Ontology (New York: Routledge, 1993). Bennett, Jane, Vibrant Matter: A Political Ecology of Things (Durham, NC: Duke University Press, 2010). Bhaskar, Roy, The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences, 3rd edn. (London: Routledge, 1998). Bhaskar, Roy, A Realist Theory of Science (Atlantic Highlands, NJ: Humanities Press, 1978). Bird, Alexander, “Limitations of Power”, in Powers and Capacities in Philosophy: The New Aristotelianism (eds., John Greco and Ruth Groff) (New York: Routledge, forthcoming). Bird, Alexander, Nature’s Metaphysics: Laws and Properties (Oxford: Oxford University Press, 2007). Buroker, Jill Vance, Kant’s Critique of Pure Reason: An Introduction (Cambridge: Cambridge University Press, 2006). Chakravartty, Anjan, A Metaphysics for Scientific Realism: Knowing the Unobservable (Cambridge: Cambridge University Press, 2007). Chakravartty, Anjan, “Inessential Aristotle: Powers Without Essences,” in Revitalizing Causality (ed., Ruth Groff) (New York: Routledge, 2008). Chisholm, Roderick, “Agents, Causes and Events: The Problem of Free Will,” in Agents, Causes and Events (ed., Timothy O’Connor) (New York: Oxford University Press, 1995).

Bibliography 131 Chisholm, Roderick, “Human Freedom and the Self,” reprinted in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). Clarke, Randolph, “Toward a Credible Agent-Causal Account of Free Will,” in Agents, Causes and Events (ed., Timothy O’Connor) (New York: Oxford University Press, 1995). Collins, Arthur, Possible Experience: Understanding Kant’s Critique of Pure Reason (Berkeley, CA: University of California Press, 1999). Copi, Irving, “Essence and Accident,” Journal of Philosophy, Volume 51, Number 23, 1954, pp. 706–19. Davidson, Donald, “Mental Events (1970),” in Essays on Actions and Events (New York: Oxford University Press, 1980). Della Rocca, Michael, Spinoza (New York: Routledge, 2008). Dicker, Georges, Hume’s Epistemology and Metaphysics: An Introduction (New York: Routledge, 1998). Ellis, Brian, “The Categorical Dimensions of the Causal Powers,” in Properties, Powers and Structures: Issues in the Metaphysics of Realism (eds., Alexander Bird, Brian Ellis and Howard Sankey) (Abingdon: Routledge, 2012). Ellis, Brian, The Philosophy of Nature: A Guide to the New Essentialism (Montreal and Kingston: McGill-Queen’s University Press, 2002). Ellis, Brian, Scientific Essentialism (Cambridge: Cambridge University Press, 2001). Flikschuh, Katrin, Freedom: Contemporary Liberal Perspectives (Malden, MA: Polity Press, 2007). Flikschuh, Katrin, Kant and Modern Political Theory (Cambridge: Cambridge University Press, 2000). Frede, Michael, A Free Will: Origins of the Notion in Ancient Thought (ed., A. A. Long) (Berkeley, CA: University of California Press, 2011). Groff, Ruth, “Getting Past Hume in the Philosophy of Social Science,” in Causality in the Sciences (eds., Phyllis McKay Illari, Federica Russo and Jon Williamson) (Oxford: Oxford University Press, 2011). Groff, Ruth, “Introduction,” Journal for Critical Realism, Volume 8, Number 3, 2009 (special issue on causal powers), pp. 267–76. Groff, Ruth (ed.), Revitalizing Causality: Realism about Causality in Philosophy and Social Science (New York: Routledge, 2008). Groff, Ruth, Critical Realism, Post-Positivism and the Possibility of Knowledge (New York: Routledge, 2004). Harre, Rom and Madden, E. H., Causal Powers: A Theory of Natural Necessity (Oxford: Blackwell, 1975). Heil, John, “Kinds and Essences,” Ratio, Volume 18, Number 4, 2005, pp. 405–19. Horkheimer, Max, Between Philosophy and Social Science: Selected Early Writings (trans., G. Frederick Hunter, Matthew S. Kramer and John Torpey) (Cambridge, MA: MIT Press, 1993). Horkheimer, Max, Critical Theory: Selected Essays (New York: Continuum, 1989). Hume, David, Enquiries Concerning Human Understanding and Concerning the Principles of Morals (reprinted from the 1777 edition with Introduction and Analytical Index by L. A. Selby-Bigge), 3rd edn., with text revised and notes by P. H. Nidditch (Oxford: Clarendon Press, 1975). Hume, David, A Treatise of Human Nature (reprinted from the original edition in three volumes and edited with an Analytical Index by L. A. Selby-Bigge) (Oxford: Clarendon Press, 1975).

132

Bibliography

Kahn, Charles H., “Aristotle vs. Descartes on the Concept of the Mental,” in Metaphysics, Soul, and Ethics in Ancient Thought: Themes from the Work of Richard Sorabji (ed., Ricardo Salles) (Oxford: Clarendon Press, 2005). Kane, Robert, “Some Neglected Pathways in the Free Will Labyrinth,” in The Oxford Handbook of Free Will (ed., Robert Kane) (Oxford: Oxford University Press, 2002). Kant, Immanuel, Critique of Pure Reason (trans. and eds., Paul Guyer and Allen W. Wood) (Cambridge: Cambridge University Press, 1997). Kant, Immanuel, Critique of Practical Reason (trans. and ed., Mary Gregor) (Cambridge: Cambridge University Press, 1997). Kant, Immanuel, Prolegomena to Any Future Metaphysics that can Qualify as a Science (trans., Paul Carus) (La Salle, IL: Open Court Classics, 1988). Kant, Immanuel, Groundwork of the Metaphysics of Morals (trans., H. J. Paton) (New York: Harper & Row, 1964). Kim, Jaegwon, Essays in the Metaphysics of Mind (Oxford: Oxford University Press, 2010). Kim, Jaegwon, Philosophy of Mind (Boulder, CO: HarperCollins, 1996). Kim, Jaegwon, Supervenience and Mind: Selected Philosophical Essays (Cambridge: Cambridge University Press, 1993). Lewis, David, “Are We Free to Break the Laws?”, reprinted in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). Locke, John, An Essay Concerning Human Understanding (abridged and with notes by A. S. Pringle-Pattison) (Ware, Herts: Wordsworth Editions Limited, 1998). Locke, John, The Second Treatise of Government (ed., Peter Laslett) (Cambridge: Cambridge University Press, 1988). Lowe, E. Jonathan, Personal Agency: The Metaphysics of Mind and Action (Oxford: Oxford University Press, 2008). Lowe, E. Jonathan, Locke (New York: Routledge, 2005). Lowe, E. Jonathan, “In Defense of Medium-Sized Specimens of Dry Goods,” Philosophy and Phenomenological Research, Volume 68, 2003, pp. 704–10. Lowe, E. Jonathan, A Survey of Metaphysics (Oxford: Oxford University Press, 2002). MacIntyre, Alasdair, Hume on “Is” and “Ought,” in Against the Self-Images of the Age: Essays on Ideology and Philosophy (Notre Dame, IN: University of Notre Dame Press, 1978). Marcuse, Herbert, Negations: Essays in Critical Theory (with translations by Jeremy J. Shapiro) (Boston, MA: Beacon Press, 1968). Marcuse, Herbert, Eros and Civilization: A Philosophical Inquiry Into Freud (Boston, MA: Beacon Press, 1966). D. H. Mellor, “In Defence of Dispositions,” The Philosophical Review, Volume 83, 1974, pp. 157–181. Mill, John Stuart, An Examination of Sir William Hamilton’s Philosophy (London: Longmans, Green, Reader, and Dyer, 1872). Mill, John Stuart, Utilitarianism, Liberty and Representative Government (New York: E. P. Dutton, 1951). Mill, John Stuart, A System of Logic, Ratiocinative and Inductive: Being a Connected View of the Principles of Evidence and the Methods of Scientific Investigation, in Collected Works of John Stuart Mill (ed., J. M. Robson) (Indianapolis, IN: Liberty Fund, 2006). Mumford, Stephen, “The Power of Power,” in Powers and Capacities in Philosophy: The New Aristotelianism (eds., John Greco and Ruth Groff) (New York: Routledge, forthcoming).

Bibliography 133 Mumford, Stephen, “Kinds, Essences, Powers,” Ratio, Volume 18, Number 4, 2005, pp. 420–36. Mumford, Stephen, Laws in Nature (Abingdon and New York: Routledge, 2004). Mumford, Stephen and Anjum, Rani Lill, Getting Causes from Powers (Oxford: Oxford University Press, 2011). Nielson, Kai, “The Compatibility of Freedom and Determinism,” reprinted in Free Will (ed., Robert Kane) (Malden, MA: Blackwell Publishing, 2002). Nussbaum, Martha, Creating Capabilities: The Human Development Approach (Cambridge, MA: The Belknap Press of Harvard University Press, 2011). Nussbaum, Martha, “Kant and Stoic Cosmopolitanism,” Journal of Political Philosophy, Volume 5, Number 1, 1997, pp. 1–25. Nuzzo, Angelica, Kant and the Unity of Reason (West Lafayette, IN: Purdue University Press, 2005). O’Connor, Brian, Adorno’s Negative Dialectic: Philosophy and the Possibility of Critical Rationality (Cambridge, MA: MIT Press, 2005). O’Connor, Timothy, Persons and Causes: The Metaphysics of Free Will (Oxford: Oxford University Press, 2000). O’Connor, Timothy and Jacobs, Jonathan D., “Emergent Individuals,” The Philosophical Quarterly, Volume 53, Number 213, 2003, pp. 540–55. Ott, Walter, Causation and Laws of Nature in Early Modern Philosophy (Oxford: Oxford University Press, 2009). Read, Rupert and Richman, Kenneth A. (eds.), The New Hume Debate, revised edition (New York: Routledge, 2008). Reid, Thomas, Essays on the Active Powers of Man, in The Works of Thomas Reid, Vol. 2 (Chestnut Hill, MA: Adamant Media Corporation, 2005). Rousseau, Jean-Jacques, “A Discourse on a Subject Proposed by the Academy of Dijon: What is the Origin of Inequality Among Men, and is it Authorized by Natural Law,” in The Social Contract and Discourses (trans., G. D. H. Cole) (London: Dent – Everyman’s Library, 1973). Ryan, Alan, John Stuart Mill (New York: Pantheon Books, 1970). Smith, G. W., “The Logic of J. S. Mill on Freedom,” Political Studies, Volume 28, Number 2, 1980, pp. 238–52. Sorabji, Richard, Necessity, Cause and Blame: Perspectives on Aristotle’s Theory (Chicago, IL: The University of Chicago Press, 1980). Spinoza, Benedict, Ethics (trans., W. H. White and A. H. Stirling) (Ware, Herts: Wordsworth Editions Limited, 2001). Taylor, Charles, Philosophical Papers, 1: Human Agency and Language (Cambridge: Cambridge University Press, 1985). Taylor, Charles, Philosophical Papers, 2: Philosophy and the Human Sciences (Cambridge: Cambridge University Press, 1985). Van Inwagen, Peter, An Essay on Free Will (Oxford: Oxford University Press, 1983). Wiggins, David, “Towards a Reasonable Libertarianism”, in Free Will, 2nd edn. (ed., Gary Watson) (Oxford: Oxford University Press, 2003). Yaffe, Gideon, Manifest Activity: Thomas Reid’s Theory of Action (Oxford: Oxford University Press, 2004).

Index

A Realist Theory of Science 7 actants 107–8 action 29, 53; autonomous 53; heteronomous 53 active power 77 Adorno, T. 54, 63–71 aesthetic judgment 55, 56 agency 101–3 agent causation 95, 102; agents 81–91; causality 76–81; free will problematic 73–6; powers and events 73–91 alienation 67 Anomalism of the Mental 83 anti-essentialism 3–4, 19–20 Aristotle 40, 44, 45‒6, 48, 81, 87, 111, 115‒17; Aristotelian 5, 51, 81, 85‒6 assemblage 108‒12 atomism 41, 49 basic capabilities 94, 98 Bennett, J. 107, 110 Bhaskar, R. 7 capabilities approach 2; free will and agency 101–3; metaphysics 92–105; ontology 95–103; political liberalism 103–5; political philosophy 93–5; power 95–9; self 99–101; ten central capabilities 93–4 capitalism 61–2 Cartesian dualism 85, 86 Causal Powers 7 causality 7; agent causation 76–81; Hume on 11–15; Kant on 50–2; Mill on 33–5 causation 13; productive 12, 13 cause 14 character 29–30, 39–40, 45–8; energetic 45‒7 Chisholm, R. 75, 79

classification 35 combined capabilities 94, 95 compatibilism 80–1, 104 conatus: see Spinoza 112‒14 Considerations on Representative Government 45 core capability 94, 99 Creating Capabilities: The Human Development Approach 93 custom 14 Davidson, D. 83–4 Della Rocca, M. 113 determinism 39, 74, 101 Discourse on the Origin of Inequality 47 dispositional realism 7, 105 Doctrine of Liberty 21 Doctrine of Necessity 21, 38–9 domination 66 double relation 30 downward causation 83 drive 58; see also Marcuse 54‒9 dynamism 112–13 Ego 38 Ellis, B. 2, 8 emergence 84, 86–91 empirical realism 52 empiricism 3 eros 56 essence 3‒5, 8‒9, 58 essentialism 8–9 ethology 45 eudaimonia 44 event causation: 74, 78, 102–3; agent causation and agents 81–91; agent causation and causality 76–81; agents and powers 73–91; free will problematic 73–6

Index 135 flourishing 44 form 44‒5, 57‒8, 65 Frankfurt School 2, 54‒72 free will 73–91, 101–3; problematic 73–6 freedom: Kant 52–4 functioning 97 General Will 48 habit 14 happiness 43–4 Harre, R. 7 hexis 45 higher pleasure 43–4 holism 20, 106 holists 40–1 Horkheimer, M. 54, 59–63 human actants 108 human dignity 94, 96, 98 human knowledge 115–16 Hume, D. 11–32, 74, 77–8, 82, 100, 103, 106, 113; causality, powers and necessity 11–19; metaphysics and the social 26–32; persons, collectivities and norms 19–26 Humeanism 1, 3, 11, 49, 73–91, 106, 110, 116 Humeans 5–6 idealism 62 imagination 24, 55–6 impulse 58 incompatibilism 74 intelligibility problem 78 internal capability 94, 96 invariant order 6 justice 23, 31 Kant, I. 50–72, 74; Kantian metaphysics and the Frankfurt School 54–72; on causation and freedom 50–4 Kantian approach 6, 74 Kantian materialism 70 Kantian metaphysics 54–72 Kantian universalism 68 Kim, J. 82–6 kinds: natural 8‒9; social 9 Latour, B. 110 Law of Causation 34, 47, 50 Law of Value 60 Laws 8, 41, 73‒4, 76–7

libertarianism 91, 102–3, 104 libidinal morality 55 Locke, J. 3‒4 Lowe, E. J. 78, 88 MacIntyre, A. 2, 18 Madden, E. H. 7 Marcuse, H. 54–9, 96 Marx, K. 58‒9, 62 materialism 59, 62‒3, 82–3; vital 107‒9 matter 18, 35‒8, 51 mechanism see Humeanism mental causation 79, 82 metaphysical neutrality 92–105; myth of 1–10 metaphysics: capabilities approach 92–105; capabilities approach and political philosophy 93–5; ontology of capabilities 95–103; political liberalism and capabilities approach 103–5; vitality 109–17 Mill, J. S. 6, 33–49, 74, 96, 100, 101; clash of categories in his thought 42–9; on causation, kinds and material entities 33–8; social entities 38–42 modernity 4 morality 62 Mumford, S. 8, 32, 104 Natural Classification 35–6 necessitarianism 114 necessitation 39 necessity: Hume on 18–19; Kant on 18, 50‒4; Mill on 37 nominal essences 3–4 non-deterministic materialism 112 non-human actants 108 non-reductive physicalism 82–3, 84, 85, 86 Nussbaum, M. 2, 92–105, 106 ontology: capabilities and 95–103; powers and appeal to Spinoza 106–17; vital materialism and 107–9; vitality and metaphysics 109–17 original natural agents 34 performance principle 55 permanent causes 34 phantasy 55 phenomenal objects 50 phenomenalism 15, 24, 36, 37 Philosophical Necessity 39 physicalism 81–6

136

Index

play impulse 56 plurality 87 political liberalism 92–105; capabilities approach 103–5 political philosophy: capabilities approach 93–5 positive liberty 66 powers 95–9; agent causation and agents 81–91; agent causation and casuality 76–81; agents and events 73–91; and capabilities approach 95‒9; free will problematic 73–6; Hume on 15–18; Mill on 37‒8, 46, 47; Spinoza 106–17; vital materialism 107–9; vitality and metaphysics 109–17 pride 30 Principle of Causal Interaction 83 Principle of the Nomological Character of Causality 83 Principle of Utility 42 properties: dispositional 4‒5; Hume on 21‒2 property 23; by accession 24–5 public sentiment 48–9 purposes 40 qualities: primary 5; secondary 5 quasi-vitalism 111 real essences 3–4 realism: critical 7; dispositional 7–8 reality principle 55 reason 27–8 regularity 6, 14‒15, 29 Reid, T. 15, 77–8 reification 62 Relations of Ideas 20 repression 58; basic 55; surplus 55

Rousseau, J. J. 48, 59 Ryan, A. 38, 42, 46 Scientific Essentialism 8 self 19, 20–1, 28–31, 38–40 Sen, A. 2, 92, 95 social solidarity 48–9 society 20–1, 22, 31–2, 40–1 Sorabji, R. 48 Spinoza, B. 106–17 spontaneity 53‒4, 68 sublimation 58 substance 13, 51, 111, 113, 116; substances 15, 36, 51, 78, 86, 110‒11, 116‒17; substance causation 12 Taylor, C. 2 “The Concept of Essence” 58 the new Hume 15 The Possibility of Naturalism 7 The Problem of Free Will 79 traction problem 91, 102‒3 transcendental ego 52, 53 transcendental idealism 52, 70 transcendental object 69 transcendental subject 52, 53 utilitarianism 42, 49 utility 44–5 Vibrant Matter 107, 109 vital materialism 107–9 vitality: metaphysics 109–17 Watkin, E. 50 Wiggins, D. 75 Winkler, K. 16

Taylor & Francis

eBooks FOR

liBRARIES

Over 23,000 eBook titles in the Humanities, Social Sciences, STM and Law from some of the world's leading imprints.

Choose from a range of subject packages or create your own! ~ Free MARC records

~ COUNTER-compliant usage statistics

~ Flexible purchase and pricing options ~ Off-site, anytime access via Athens or referring URL

~ Print or copy pages or chapters

~ Full content search ~ Bookmark, highlight and annotate text ~ Access to thousands of pages of quality research

at the click of a button For mo re information, pricing enquiries or to order a free trial, contact your local online sales team . UK and Rest of World: [email protected] US, Canada and Latin America: e-reference®taylorandfrancis.com

www.ebooksubscriptions.com