The Theory of the Individual in Economics
The concept of the individual is central to the understanding of behavior in...
78 downloads
1126 Views
910KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
The Theory of the Individual in Economics
The concept of the individual is central to the understanding of behavior in economics. Different approaches in economics implicitly rely on different theories of the individual. Yet economics is guilty of using this very important concept without questioning how it is theorized. This superb book remedies this oversight. The new approach put forward by John B. Davis employs identity analysis to understand theories of the individual in economics. It combines philosophy and economics to determine when theories of the individual are successful. With both heterodox and orthodox economics receiving a thorough analysis, this book is at once inclusive and systematic. Davis has produced a startlingly original book that should become essential reading for all those interested in the study of economic philosophy and methodology, but will also be of great interest to philosophers and social scientists outside economics. John B. Davis is Professor of History and Philosophy of Economics at the University of Amsterdam, the Netherlands, and Professor of Economics at Marquette University, Milwaukee, USA.
Advances in social economics Edited by John B. Davis
Professor of History and Philosophy of Economic, University of Amsterdam, the Netherlands, and Professor of Economics, Marquette University, Milwaukee, USA This series presents new advances and developments in social economics thinking on a variety of subjects that concern the link between social values and economics. Need, justice and equity, gender, cooperation, work, poverty, the environment, class, institutions, public policy and methodology are some of the most important themes. Among the orientations of the authors are social economist, institutionalist, humanist, solidarist, cooperatist, radical and Marxist, feminist, post-Keynesian, behavioralist, and environmentalist. The series offers new contributions from today’s foremost thinkers on the social character of the economy. Published in conjunction with the Association of Social Economics. Books published in the series include Social Economics Premises, findings and policies Edited by Edward J. O’Boyle The Environmental Consequences of Growth Steady-state economics as an alternative to ecological decline Douglas Booth The Human Firm A socio-economic analysis of its behaviour and potential in a new economic age John Tomer Economics for the Common Good Two centuries of economic thought in the humanist tradition Mark A. Lutz Working Time International trends, theory and policy perspectives Edited by Lonnie Golden and Deborah M. Figart The Social Economics of Health Care John B. Davis Reclaiming Evolution A Marxist institutionalist dialogue on social change William M. Dugger and Howard J. Sherman The Theory of the Individual in Economics Identity and value John B. Davis
The Theory of the Individual in Economics Identity and value John B. Davis
For Bryan Davis
First published 2003 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2003.
© 2003 John B. Davis 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. 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 Davis, John, 1947 June 13– The theory of the individual in economics : indentity and value / John Davis. p. cm. — (Advances in social economics) Includes bibliographical references and index. ISBN 0–415–20219–1 (hbk) — ISBN 0–415–20220–5 (pbk.) 1. Economics—Philosophy. 2. Individuality. 3. Identity (Psychology)—Economic aspects. 4. Value. I. Title. II. Series. HB72 .D247 2003 330'.01'9—dc21 ISBN 0-203-45768-4 Master e-book ISBN
ISBN 0-203-76592-3 (Adobe eReader Format)
ISBN 0–415–20219–1 (hbk) ISBN 0–415–20220–5 (pbk)
2002023992
Contents
Preface and acknowledgments 1
Framing the issues
1.1 1.2 1.3 1.4 1.5
viii 1
Modernism as dualism: origins of the modernist concept of the individual 2 Contemporary critiques of the modernist conception of the individual 6 Individuals in economics 11 Orthodox economics and heterodox economics 16 Plan of this work 18
PART I
Orthodox economics
21
2
23
The atomistic individual
2.1 2.2 2.3 2.4 2.5 2.6 3
Locke’s legacy 24 The evolution of the atomistic individual conception 27 The requirements of atomism 32 Methodological individualism and reductionism: a last-ditch defense? 35 Game theory to the rescue? 38 What kind of individual? 42
Reidentification: preferences and human capital
3.1 3.2 3.3 3.4 3.5 3.6 3.7
The standard pure preferences view of the individual 47 Reidentification à la Locke and in the pure preferences account 49 Locke’s critics 50 The Butler–Hume critique applied to the neoclassical view 53 The time allocation conception of the individual 55 New problems 58 Individuation reconsidered 61
45
vi
Contents
4
Individuation: multiple selves
4.1 4.2 4.3 4.4 5
63
Single versus multiple utility functions 65 Internal preference structures 69 Endogenously changing preferences 77 A unity of selves or a plurality of selves? 79
After the fall: the machinery of choice
81
5.1 Mind as a computer 84 5.2 Economics as cognitive science 88 5.3 After the fall 100 PART II
Heterodox economics
105
6
107
The embedded individual
6.1 6.2 6.3 6.4 6.5 7
Individuation: collective intentionality
7.1 7.2 7.3 7.4 7.5 8
The legacy of Marx and Durkheim 109 Social theory’s structure–agency models of the individual and society 111 Social psychology’s self-referent behavior and individual self-concept 114 The socially embedded individual conception in economics 117 The socially embedded individual conception 127 130
Collective intentionality analysis 132 Collective intentionality and the structure–agent framework 136 A revised view of individual economic behavior 139 The individuation of embedded individuals 143 Individuation in the atomistic and socially embedded individual conceptions 147
Reidentification: capabilities
8.1 Sen’s capability framework 152 8.2 The capability framework applied to the reidentification of embedded individuals 155
8.3 The embedded individual conception as an ideal conception 160 8.4 Sen’s thinking about the individual versus the embedded individual conception 163
8.5 Concluding remarks on the atomistic and embedded individual conceptions 164
150
Contents 9 Before the fall: value in economics
vii 167
9.1 Facts and values 168 9.2 Welfarism, utilitarianism, and the atomistic individual conception 171 9.3 Normative reasoning and the socially embedded individual conception 174
9.4 Individual identity as a normative concern 180 9.5 A classical world 182 10 Revisiting the issues
10.1 10.2 10.3 10.4
184
Ontology in economics 185 Individual identity in economics and personal identity 187 Two historical traditions 188 Significance 190
Notes References Index
195 199 211
Preface and acknowledgments
This book is a product of my longstanding interest in the nature of the individual in economics. Its subject matter lies at the intersection of philosophy, economics, and the history of economics, and perhaps for that reason the book will be thought to involve a highly specialized sort of discussion. However, throughout my work on this subject I have been motivated by the belief that not only the topic of the individual in economics does have a wider importance but also that this topic has been relatively neglected. Thus, I hope that readers will approach what follows in a spirit like my own, and that this will help generate interest in the book’s issues, arguments, and conclusions. I am indebted to many people for comment upon my work in its various stages over a number of years, but I particularly appreciate the assistance of Wilfred Dolfsma, Zohreh Emami, Nancy Folbre, Philippe Fontaine, Edward Fullbrook, Wade Hands, Geoff Harcourt, Shaun Hargreaves Heap, John Henry, Geoff Hodgson, Tony Lawson, Paul Lewis, Mark Lutz, Uskali Mäki, Robert McMaster, Lance Minkler, Phil Mirowski, Philippe Mongin, Leon Montes, Phil O’Hara, Steve Pressman, Geert Reuten, Jochen Runde, Malcolm Rutherford, Warren Samuels, Esther-Mirjam Sent, Alex Viskovatoff, and Jack Vromen. I have also benefited from individuals I cannot name, including anonymous referees of my work, who in some instances have been very helpful, and participants at various seminars and conferences at which I have made presentations. I also thank my students from Spring 2001 at the Erasmus Institute for Philosophy and Economics, where parts of this book were presented. I am very grateful to Marquette University for research support and to members of the Marquette Department of Economics for collegial support. I also thank Robert Langham and Alan Jarvis at Routledge and Alison Kirk formerly of Routledge for supporting this book and for support generally for the Review of Social Economy and social economics. Thanks also to Leila Davis for editorial assistance. In parts of this book I have adapted material previously published in Cambridge Journal of Economics, Journal of Economic Methodology, Review of Political Economy, The Economic World View edited by Uskali Mäki, and Intersubjectivity in Economics edited by Edward Fullbrook. Finally, most thanks of all to Zohreh.
1 Framing the issues
[T]he real self is “extensionless”; it is nowhere but in this power to fix things as objects. (Taylor 1989: 172) No entity without identity.
(Quine 1969: 23)
This is a book about our understanding of the individual in economics. The concept of the individual is one of the most fundamental in contemporary society. It may even be the most fundamental of all our concepts. We cannot understand the historical evolution of political systems in terms of democracy, freedom, and human rights, the development of knowledge and science, and the quality and meaning of life without recognizing the centrality of the individual to our thinking. Human society could conceivably have developed differently in this regard. However, one thing we can know with certainty at this point in history is that individuality is a fundamental preoccupation of contemporary human society. Yet in economics, with its tremendous influence on society, very little attention is given to the theory of the individual. The theory of the individual concerns our most basic assumptions regarding what explains individuality. Economics, in fact, takes the individual as given, and operates on the implicit assumption that one particular conception of the individual, indeed one of well-established lineage—the subjectivist view—successfully explains individuality. In this respect it is remarkably alone since, almost everywhere else in science, the humanities, the arts, and law, most believe this conception of the individual in economics to be naive. At the same time, it is difficult for those outside economics to say how individuals should be approached in economics, since the theory of individual choice—in which the understanding of the individual within economics is elaborated, and which economists generally regard as the centerpiece of scientific achievement in the field—places a forbidding array of technical issues in the way of any inquiry into the nature of individuals in economic life. Indeed, the theory of choice
2
Framing the issues
blocks economists’ own investigation of the nature of individuality in economics as well, and most economists accordingly take the view of the individual in economics as a given. This book, then, attempts to expand the space in which individuals can be discussed in economics. It does so by examining the requirements for a theory of the individual in economics, and by investigating the two main conceptions of the individual in economics—the familiar orthodox conception associated with neoclassical and mainstream economics and a less sharply articulated conception associated with dissident traditions in heterodox economics. Though this is an investigation carried out within the philosophy and history of economics, it should be seen as having considerable practical significance. Thinking about the individual in economics is not just important for the future development of that field. How economics understands individuals has extremely important social consequences. In particular, if the way most economists understand the individual actually contributes to a decline and weakening of contemporary society’s commitment to the integrity of the individual, then a closer look at conceptions of the individual in economics is surely in order. In this chapter, I seek to frame the issues this book investigates. I begin in Section 1.1 with a review of the seventeenth century philosophical origins of the modernist conception of the individual that ultimately came to underlie neoclassical economics at the end of the nineteenth century. Section 1.2 outlines the main contemporary critiques of this conception in order to explain what is problematic about this understanding of the individual. Here, I distinguish between the social science critique and the postmodernist critique. Section 1.3 introduces the systematic framework which I use in the book to examine the theory of the individual in economics, namely, an individual identity analysis that has parallels to personal identity analysis in philosophy. In Section 1.4, I briefly differentiate the two main conceptions of the individual in economics, that associated with orthodox economics and that associated with heterodox economics. Finally, Section 1.5 closes with a summary of the plan and organization of the book.
1.1 Modernism as dualism: origins of the modernist concept of the individual The modern concept of the individual has its origins in the Cartesian–Newtonian dualism of human subjectivity and objective nature associated with Enlightenment conceptions of science and society.1 Contemporary history of science emphasizes the Newtonian vision and the nature side of this dualism, but the inseparable accompaniment of this vision was Descartes’ view of the individual as a disengaged, subjective inwardness. In my view, like most dualisms, this Cartesian–Newtonian one was fundamentally problematic and therefore ultimately unsustainable in each of its two aspects. Just as we cannot understand nature purely as a mechanism, so neither can we understand the human individual purely as a disengaged subjectivity. Thus, looking ahead,
Framing the issues
3
that neoclassicism took up this particular concept of the individual planted the seeds of its ultimate dissolution as a theory of the individual in economics. In his Sources of the Self: The Making of the Modern Identity, Taylor traces the development of the modernist concept of the individual from Descartes to Locke. Descartes’ image of the self as a disengaged subject identifies the self with the power of reason by virtue of the self not being “in” the material world. Locke carries this image further in ascribing a power to the self to objectify the world. As not being in space— as an extensionless point—the “punctual” self, as Taylor puts it, has the power to set aside the influences that opinion, custom, and desire can have upon us, so as continually to remake itself in a manner that magnifies its own happiness. For Locke: To take this stance is to identify oneself with the power to objectify and remake, and by this act to distance oneself from all the particular features which are object of potential change. What we are essentially is none of the latter, but what finds itself capable of fixing them and working on them. This is what the image of the point is meant to convey, drawing on the geometrical term: the real self is “extensionless”; it is nowhere but in this power to fix things as objects. (Taylor 1989: 171–2)
What is it that makes this particular image of the individual a characteristically modern one? The answer lies in the unique response it permits to the rise of mechanical science in the seventeenth century. For Descartes, seeing the individual as disengaged from the world provided both an independent basis for scientific reason and also a means of reconciling science and theology. In the prescientific medieval world, before nature came to be understood as a mechanism, all was thought to be governed by eternal Ideas imperfectly embodied in the physical world. Plato originally developed this conception, but his basic understanding was later systematically integrated into Christian theology, with the eternal Ideas reinterpreted as the Thoughts of God. According to this view, the things of earthly life are imperfect exemplifications of a cosmic order organized as a constellation of Ideas, and human understanding, embodied in the fallen human form, could never be more than a striving to imperfectly grasp these Ideas through their manifestation in the phenomena of the world. In contrast, then, to the later, modernist conception of the human mind as a “mirror of nature” (Rorty, 1979), before the Enlightenment “nature was the image of mind,” that is, the mind of God.2 Moreover, in contrast to our later understanding of the world as ruled by cause and effect, in both Plato’s philosophy and Christian theology the world was organized teleologically, exhibiting a Reason and Goodness for Plato and then the Wisdom of God in Christian theology. Human understanding and behavior had their own particular role in all this. As Taylor puts it, “As humans we are to conform to our Idea, and this in turn must play its part in the whole, which among other things involves our being ‘rational’, i.e., capable of seeing the self-manifesting order” (Taylor 1989: 161).
4
Framing the issues
Adopting a view of the world as mechanism dethroned this medieval picture of the world. Descartes still preserved a place for God outside nature by supposing the world operated according to the axioms of mathematics and analytic geometry, themselves determined by divine fiat. More importantly, he supposed that our capacity to understand the world as mechanism depended upon our being able to form clear and distinct ideas which only God could guarantee. The connection to the new concept of the individual was direct and immediate. Individual existence, rationality, and God were all tied together by Descartes in his famous cogito ergo sum argument, in which he withdraws into himself in doubting his ordinary beliefs, finds certainty in his clear and distinct ideas made possible only by God’s goodness, and simultaneously proves his own existence as consciousness, that is, as pure subjective inwardness. From this interior vantage point, Descartes the scientist might proceed to explain Newton’s world through exercise of a reason that was certain and powerful by virtue of its joint authorization by God and removal from the earthly world. However, God only guaranteed us a capacity for clear and distinct ideas, and it was still possible that we might have obscure and indistinct ones. How, then, were scientists to proceed in producing scientific truths about the world? Descartes’ solution was an extension of his basic solution in that it involved distinguishing between the primary and secondary qualities of things. The sources of our obscure and indistinct ideas, he reasoned, are the distortions of sensation and perception, which concern secondary qualities such as color, taste, smell, sound, and temperature, as compared with primary qualities such as figure, number, position, and size which are amenable to mathematical analysis. Descartes’ view was that only primary properties are “in” nature, and that secondary qualities are “in” our senses. Attaining scientific truth thus requires that we “disengage” ourselves from our senses and rely upon our Godgiven ability to form clear and distinct ideas of the way the world is in itself, or in terms of its primary qualities. Reason, reposited in the subjectively inward individual, possesses the capacity to see the inner nature of the world. Descartes’ famous cogito argument consequently gave the Enlightenment an epistemology for science, a new ontology of nature without God’s participation, and a new view of the individual—one that withdrew from the world to understand and control the world’s mechanical laws. This fundamental division of the world into an inner subjective domain and an outer objective domain was Descartes’ great contribution to modern thinking, and it remains the foundation for contemporary thinking about the individual as a disengaged subjectivity and the modern view of nature as a “spiritless” domain. However, Descartes’ commitment to the idea of disengagement and recourse to an inner reason contained tremendous ambiguities. In a seventeenth century England that combined the Puritan regicide, hatred of Catholicism, and the science of Newton and Boyle, Locke sought to resolve Descartes’ ambiguities by rejecting Descartes’ doctrine of innate ideas. The clear and distinct ideas that Descartes thought our inner reason made possible were innate ideas, because God had created our capacity
Framing the issues
5
to grasp them. In the climate of his time, however, Locke feared that what might be represented as ideas God guaranteed us might just be the opinions of men. He consequently proposed a thorough clearing away of “the rubbish that lies in the way to knowledge” (Locke 1975 [1694]: Epistle, 14), and called for a rebuilding of knowledge out of the simple ideas that came to us through sense experience—ideas he regarded as inalterable atoms of understanding. This rebuilding itself took the form of an assembly and reassembly of simple ideas in ways that would create complex ideas, so that a quasi-mechanical, associational organization of the mind would parallel the mechanical organization of nature. Of course, Locke’s view of knowledge as being built up through association of simple ideas hardly amounts to an apparatus likely to support the new ambitions of science in the seventeenth and eighteenth centuries. However, Locke was less interested in science than in the relationship between the individual and the state, and thus it seems that the important thing about his view of our assembly and reassembly of simple ideas was that it presupposed an empowered disengagement from the world, a “double movement of suspension and examination [in which] we wrest the control of our thinking and outlook away from passion or custom or authority and assume responsibility for it ourselves” (Taylor 1989: 167). This ability of individuals to make their mental activities their own, free of the influence, even the despotism, of others, was what seems to have motivated Locke most strongly. Indeed, he characterized this capacity to suspend judgment and dispassionately examine the credentials of our ideas as a question of freedom versus slavery. This in turn tied in with his theory of legitimate government that presupposed a prior state of nature in which all persons had incontrovertible title to their own persons and labor. For Locke, that is, a capacity for subjective disengagement and title to oneself were part and parcel of a single political theory of the individual. Thus, whereas Descartes laid individualist foundations for science and knowledge, Locke extended these foundations to the subject of individuals and their relation to society. Arguably, a consequence of this extension was that Locke was more radical than Descartes in the degree to which he understood subjectivity as constituting individuals as independent, autonomous beings. Since Descartes relied on God to guarantee clear and distinct ideas, the existence of individuals in effect depended on God. However, in abandoning innate ideas and reason, and in supposing that individuals had a natural ability to recognize the simple and inalterable ideas of sense experience, Locke effectively made individuals responsible for their own existence. Thus, he advanced what is generally regarded as the first philosophical theory of personal identity—a theory that for him involved defining the self reflexively and self-identically: “For it is by the consciousness it has of its present thoughts and actions, that it is a self to itself now, and so will be the same self, as far as the same consciousness can extend to actions past and to come” (Locke 1975 [1694]: 2.27.10). The self reflexively constitutes itself, because as nothing but consciousness it must always be a “self to itself” at any point in time. The self is also the same through time, or self-identical,
6
Framing the issues
because it can never be anything but consciousness through all “actions past and to come.” I return to this conception of personal identity in Chapter 3. For now, let us simply note another important feature of Locke’s view: the individual understood as consciousness is not only an autonomous being but also an inescapably private being. A being understood simply as consciousness must be a private being, both because the self as pure consciousness can only be conceived in firstperson terms, and because consciousness, by virtue of its intentional character, must always be separate from that which consciousness is of. For Locke, individuals are confined within a first-person world, with the world of real things only available to them as intentional objects. That is, his conception of the individual is solipsistic and idealist. Locke thus gave the strongest possible interpretation to the Cartesian–Newtonian dualism of subject and object worlds. However, there exists a long history of critique of Locke’s conception of the individual, the outline of which can briefly be introduced here. Basically, the idea of subjectivity understood as a complete disengagement from the world borders on being incoherent or even self-contradictory. Wittgenstein (1958 [1953]) made essentially this critique in his arguments against the possibility of there being a private language. The Lockean individual presumably employs a private language generated from names privately given to sense experience events. But how would such an individual know that a name given to one experience applies also to a like experience? Something like this would be required if the assembly and reassembly of simple ideas depended upon relations of similarity and likeness. Wittgenstein’s view was that, unless we want to assume some innate faculty for detecting similarity and likeness (which Locke of course ruled out), we must depend on how we learn to use names from others in a shared, public language. Language for Wittgenstein is social. This, however, is incompatible with Locke’s view of the individual as a subjectively disengaged being. But if the individual cannot realistically disengage, then the self is either undefined or defined through identification with others (termed “social identification” in contemporary social psychology). This Scylla and Charybdis was inherited in the neoclassical appropriation of Locke’s view and goes a considerable distance towards helping explain neoclassicism’s history of continued efforts to escape from Locke’s contradictions. Final escape, I will argue, was finally achieved, but at the cost of bringing to an end neoclassicism having any theory of the individual for economics.
1.2 Contemporary critiques of the modernist conception of the individual Locke’s view relies, in Taylor’s words, on the idea that we are able to “wrest the control of our thinking and outlook away from passion or custom or authority and assume responsibility for it ourselves” (Taylor 1989: 167). Contemporary thinking about the breadth and depth of political, economic, technological, and cultural
Framing the issues
7
change in modern society has created a wide-ranging skepticism regarding Locke’s assumption. However, if Locke was mistaken in his estimation of our ability to suspend belief, then our thinking loses the security of inwardness and loses its presumed autonomy, so that our ideas and also ourselves conceivably become merely products of social forces. Where does this leave the individual? Two versions of this critique are now widely held, one more moderate and one more extreme. First, many social scientists and philosophers argue that the individual remains an important fixture of contemporary society, but say that the individual’s nature and identity are constituted for it in various social processes, such as consumerism, the mass media, and identity politics. The self exists, but its identity is determined for it from without by broad social forces. I call this the social science critique. Second, postmodernists, deconstructionists, critical theorists, and poststructuralists situate the self within discourse, and argue that the shifting and fragmentary nature of discourse “decenters” the self, undermining its status as a self-subsistent real entity, thus dissolving it as a fixture of contemporary life. According to this view, the self is a modernist illusion. I call this the postmodernist critique. 1.2.1
The social science critique
The more moderate social science critique of the Lockean individual derives from the simple idea that, if the individual is not self-directed, then the individual can only be “other-directed” (e.g., Riesman et al. 1950). If our experience is not essentially private and inward, as Locke believed, then we and it must be understood in terms of our relation to society. However, if we define ourselves in the terms that society creates for us, then individuality rests in identifying with others rather than distinguishing ourselves from others. Philosophers have interpreted the concept of individual identity in the latter sense as personal identity. But the concept of “identity” itself is ambiguous, and also includes the idea of our identifying with others. Individual identity, then, might be better understood in terms of the ways that individuals seek “social identification.” Indeed, from this perspective, we can explore countless possible means and vehicles by which individuals have historically sought to bind themselves to others in groups and to groups, communities, and movements, often in order to replace the burden of maintaining personal identities for the security of group identities. Thus, a wide range of different literatures in sociology, psychology, communications, cultural studies, and other fields have investigated the various roles that consumerism, religion, ethnicity, mass media, mass culture, mass movements, etc., have played in constituting individual identity in terms of social identities. In most such accounts, however, the view is not that the individual ceases to exist, but rather that individual identity is continually created and re-created in terms of external group associations. Thus, this critique is still a moderate one in allowing individual identity to remain a pre-eminent concern, though not in its traditional Enlightenment guise as the personal identity of an independent, self-determining being.
8
Framing the issues
For example, one broad approach within this overall social identity literature starts off with the Lockean view of the individual as subjective inwardness to argue that social identification maintains individual identity but only in a distorted way. If the hallmarks of the traditional conception are that the individual is self-directed and independent, then individuals who acquire identities by identifying with others must be internally conflicted, and prey to external influences. The origins of this approach go back to earlier nineteenth century literature on alienation, for example, in Marx’s early writings, and have received continued expression in later existentialist writings. One more recent, influential example from social science is Lasch’s theory of the “narcissistic personality” which sees individuals as pathologically individualistic in consumer cultures structured around commodity consumption (Lasch 1979). A widely accepted orientation in psychology uses the idea of an inner conflict brought on by a drive to identify with others to advance the notion of individual identity crises. Therapy strategies in psychology and psychiatry thus often aim at restoring a sense of self, finding the “authentic” self, and eliminating a false sense of self on the assumption that the patient’s problem is a disorder in personal identity brought on by fragmentation caused by social identification processes. Another focus of the social science critique of the Lockean individual is in work that explains individual identity in terms of identity politics. Ethnic, racial, gender, and other social group identities are said to be shaped by alien views of “otherness” projected upon their members by individuals from other groups intent upon social dominance. These imposed social identities, however, are contested by individuals in dominated groups through their joint elevation of “difference” and creation of preferred social identities for self-recognition and self-defense. One recent example of this from sociology is the work on Native American ethnicity which investigates how tribal identities are constructed through markets for “genuine” Indian art and through US federal government “official” measures of “Indianness” in terms of blood quantum rules (e.g., Nagel 2000). Another example is “standpoint epistemology” developed by feminist philosophers (e.g., Harding 1986), which examines how individuals are shaped by their “social locations.” In general, the strategy of identity politics is to problematize an imposed, “official” group identity by redefining “otherness” as difference. However, contra Locke, individual identity is still primarily social identity, and not an identity apart from others. That is, these strategies simply seek alternative social identifications. 1.2.2
The postmodernist critique
The more radical, postmodernist critique of the Lockean individual has its origins in the work of Nietzsche, who regarded the concept of the self as a fiction we impose upon ourselves through language to veil the world’s terrifying nature as ceaseless change and endless becoming. Thus, to expose the illusoriness of the self, Nietzsche abandoned traditional writing styles to produce collections of disjointed, aphoristic
Framing the issues
9
texts that fragmented his own identity as author, while simultaneously undermining the reader’s own sense of self ordinarily built up through a feeling of engagement with a single author. Both were “decentered” when one exposed, as he put it in “Twilight of the idols”: the basic presuppositions of the metaphysics of language, in plain talk, the presuppositions of reason. Everywhere it sees a doer and doing; it believes in the will as the cause; it believes in the ego, in the ego as being, in the ego as substance, and it projects this faith in the egosubstance upon all things—only thereby does it first create the concept of a “thing”. (Nietzsche 1968 [1889]: 483)
For Nietzsche, without the illusions of language, there simply are no doers, no egos, and no selves. Descartes and Locke supposed there must be, because a world without individuals—indeed a world without God—was too terrible for them to contemplate. However, for Nietzsche, language was a changeable play of masks behind which there was simply nothing at all. Nietzsche’s ideas reappear in Derrida, Foucault, Lyotard, Jameson, Rorty, Baudrillard, and others who follow him in emphasizing the self’s inescapable decenteredness. The issue for them is also not how individual identity might be “other-directed” and transformed by social processes, as in the social science critique, but rather how individual identity always dissolves away whenever we attempt to locate it. For example, in Jameson’s arguments about the decline of “depth models” of human subjectivity in late capitalism, the commodification of society is said to separate the construction of subjectivity from the social-economic process, leaving subjectivity to be constituted out of an unstructured play of whatever free images or “signifiers” a culture may happen to possess (Jameson 1984). However, a discourse of “signifiers” as a shifting, fluid affair offers no resources for settled ontologies, particularly for an ontology of the self. Consequently, the “individual” can at best be only a loosely associated, changing collection of fleeting images. That is, since imagery occupies the surface of things, a subject constituted out of images cannot have depth, or, in the case of Locke’s individual, an inwardness. Baudrillard takes this argument a step further and argues that even the distinction between surface and depth must be abandoned (Baudrillard 1983). Within the world of discourse, signifier and referent inevitably merge with one another, since the term “referent” is itself simply another piece of discourse. Thus, though we might speak of “ontological depth” and “subjective inwardness,” these terms do not lay claim to a world beyond discourse. The postmodernist critique of the individual also has its proponents in the margins of economics (cf. Cullenberg et al. 2001), particularly in post-Marxism, among feminists, in the French Intersubjectivist School, and among discourse theorists in the field of economic methodology. Amariglio of the Rethinking Marxism School follows Foucault in arguing that humanism’s decline and contemporary loss of faith in the modern episteme signals “the death of Man” (Amariglio 1988, 1990).
10
Framing the issues
Economic exploitation, political oppression, and the disempowerment of individuals and entire peoples are not historical anomalies, but rather reflect a postmodern condition that fragments and dislodges man from the center of the universe. Individuals are “overdetermined” by cross-cutting, conflicting social, economic and cultural forces that undermine the old essentialist view of the individual as a selfsubsistent unitary being. Among feminists, Nelson (1996) has created a gender/value compass of positive and negative characteristics arrayed across the dualism of masculine and feminine. The purpose is not to relocate or restructure male and female individuality, but rather to recognize the multiplicity of dimensions that govern our understanding of individuals. This in turn is part of a project of criticizing stereotypical roles (the “self-denying wife”) and perverse myths (“economic man”) that dominate economics. Dupuy of the French Intersubjectivist School sees the individual economic agent as “a being radically incomplete, in a state of lack” (Dupuy 1988: 85). Drawing on Keynes’ General Theory treatment of conventions and account of mimetic behavior in speculative markets (Keynes 1936: Ch. 12), Dupuy characterizes a “conventional order” as a system “that contains the principle of its own decomposition” (Dupuy 1994: 97). Individuals within such a system reciprocally imitate one another rather than act on economic “fundamentals,” reducing economic life to a continually shifting kaleidoscope of interdependent subjectivities (cf. Fullbrook 1997). Mehta’s discourse analysis addresses the dualism between academic and ersatz economics, and argues that the former, in the guise of modern knowledge, “constitutes a form of iconolatry which has subjugated, or rendered ersatz, the other knowledge of wo/man-in-the-street” (Mehta 1999: 45). Thus, Homo economicus as a type displaces ordinary individuals who lose their specificity in the requirement that they be seen only within a discourse’s system of classification. “It’s the death of the Subject—an embalmment” (Mehta 1999: 44), the elimination of the real individual for a fetish that fixes the eye. The postmodernist critique, like the social science critique, exposes a fundamental weakness in the Enlightenment conception of the individual as subjective inwardness, namely, that it lacks a specifically positive account of what subjective inwardness actually involves. Descartes’ emphasis on disengagement and Locke’s attention to individual suspension of belief make our understanding of the individual a matter of what the individual is not, rather than what it is. Indeed, the idea of “inwardness” itself is more the idea of a direction than a description of what subjectivity might amount to. This is not to say that the traditional view of the individual as somehow “out” of the world of nature does not still have serious adherents, for example, Nagel (1986), who argues in favor of subjectivity as the “view from nowhere,” or Swinburne (1987), who sees treating the individual as pure subjectivity as a correlate of the traditional Christian conception of individuals possessing an immaterial soul. In fact, it is probably fair to say that the idea that the individual is defined in terms of subjectivity, and that subjectivity is on an altogether different plane from principles that explain the natural world, continues to be held on some
Framing the issues
11
level by many. Yet, in light of the social science and postmodernist critiques, we need to consider the possibility not only that this view contains the seeds of its own destruction but also that this self-destruction constitutes the foundation for a darker view of the individual. The hothouse growth in recent years of cognitive science and expanding support for the view that the mind is only a type of computer suggest that instability in this conception of the individual as a subjectivist being might create the best possible conditions for a new view of the individual as a machine-like or “cyborg” being (Haraway 1991; Mirowski 2002). The logic for this belongs to Descartes and Locke. If the individual as subjectivity is a problematic idea, and subjectivity is simply “not” nature, then it is a short step to seeing the mind as merely another natural phenomenon. How, then, are we to begin thinking more systematically about the individual in economics?
1.3
Individuals in economics
The social science critique gives us reason to think that individuals are subsumed within groups. The postmodernist critique eliminates individuals altogether. If we are to think that individuals are nonetheless important in economics, we need to produce arguments that somehow demonstrate that they retain some form of integrity as individual agents in spite of those forces operating upon them that either collapse them into groups or remove them from the scene altogether. That is, we need arguments that might justify our thinking that individuals possess at least a relatively independent identity. In this section, then, I discuss what the idea of individual identity involves, explain how I will investigate individual identity in this book in terms of identity criteria, and finally situate this framework in the context of current thinking about economic methodology. Before doing this, however, let me briefly note a further rationale for investigating individual identity in economics in addition to its social significance. The lack of attention given to the theory of the individual in economics reflects economists’ widely held and generally unexamined belief that individuals are exogenous to the economic process. Most economists, that is, do not believe that individuals are changed in nature by the economic process, and much less believe that their status as individuals might itself be affected by that process, or that individuals and the nature of individuality might be endogenous to the economic process. They believe these things in spite of obvious historical evidence that the economic process is a record of the continual transformation and destruction of individuals. However, perhaps changes in the way economists are beginning to think about time, process, evolution, and cumulative causation herald a change in how in the future they will think about how individuals are affected by change in economic life. My argument is that beginning to think systematically about individuals as endogenous to the economic process depends upon first developing a framework for discussing individuality and the nature of individuals. By analogy to accounts of change in economic systems
12
Framing the issues
explained in terms of theorized structures within which that change can be said to occur, the theory of individual identity constitutes a theoretical structure of analysis within which an investigation of change in individuals can take place. This book accordingly introduces a structural framework of analysis for the theory of the individual in economics, employing the concept of identity criteria for that purpose. 1.3.1
Individual identity in economics
The topic of individual identity in economics can be introduced as an analog to the topic of personal identity in philosophy. When philosophers investigate personal identity, they ask what must we assume about a person in order to suppose that that person retains a single identity in spite of changes in any number of that person’s characteristics (cf. Noonan 1993). For example, one might argue that, though a person ages or changes in appearance over time, because the person still “has the same body,” or better perhaps, “has the same brain,” we are justified in saying that the person remains the same person. Or, were one to point out that bodies and brains also change over time at the cellular level, then we might say that the person “having psychological continuity” justifies our ascribing personal identity to that person. Thus, the basic idea that philosophers investigate in connection with the problem of personal identity is whether there can reasonably be said to be something unchanging about individuals in spite of their obviously undergoing change through time. In economics, as traditionally understood, it seems clear that something like personal identity is assumed about individual economic agents. Individual economic agents are clearly changed in some respects as a result of actions they undertake, as, for example, when individuals raise their “utility” levels when they consume commodities, or when individuals acquire new skills because they invest in human capital. However, to suggest that individuals who raise their utility or acquire new skills are somehow not the same individuals who consume commodities and invest in human capital borders on incoherence. Indeed, economists innocuously assume that, when an individual at time t1 engages in actions meant to produce certain consequences at time t2, the same individual, even if changed in important respects, is involved at both t1 and t2. Thus, implicitly economists follow philosophers in supposing that there is something unchanging about individual economic agents that would in principle explain their individual identity. This implies that it should be possible to say what explains, or at the very least say what is believed to explain, individual identity in economics. I characterize this concern with tracking the individual through change as a concern with reidentifying the individual. It can consequently be called the reidentification problem in individual identity analysis in economics. Traditionally, the reidentification problem has been the primary concern of philosophers investigating the problem of personal identity. Methodologically speaking, their procedure has been first to assume that they have a distinct individual of such-and-such characteristics, and then to investigate how one
Framing the issues
13
might reidentify that particular individual through a process of change that affects some but hopefully not all of that individual’s characteristics. However, the social science and postmodernist critiques indicate that addressing the reidentification problem alone is not enough. Since they conclude that the individual either disappears into the group or disappears altogether, for them there is simply no distinct or independent individual in the first place that we might then attempt to track and reidentify through change. It may seem odd to say that individuals might not be distinct from one another, since human beings can be individuated by virtue of their having different bodies. However, the issue here is not physical distinctiveness but rather whether individuals are distinct as economic agents. Economics, of course, often disregards individuals’ physical distinctiveness when it counts multiperson entities such as families and firms as individual economic agents.3 At the same time, economics (at least orthodox economics) also places considerable emphasis on single human individuals as economic agents. How, then, are we to know when individuals ought to be subsumed within groups of individuals and when they ought to be regarded as distinct agents? Since economics allows for both cases, it ought to have some implicit criterion for differentiating between them. Drawing this out would then be important, because that would tell us when and whether the social science critique can be resisted in economics. The issue, then, is the criterion by which we distinguish individuals as independent economic agents from one another. I call this the individuation problem in individual identity analysis in economics.4 How, then, can these two problems be systematically investigated? My strategy is to formulate identity criteria for both problems that particular conceptions of the individual in economics would need to jointly satisfy if the economic theories in which those conceptions operate are to be said to possess adequate conceptions of the individual. Since the goal of this investigation is to be able to say whether individuals exist according to a given conception of what individuals are, in the next subsection I justify these two criteria in ontological terms, and then comment on the general methodology involved. 1.3.2
Identity criteria
Talking about whether independent individuals exist as distinct beings through a process of change involves ontological analysis. Ontological analysis generally operates in terms of basic categories of existence. This implies that, if we are to investigate individuals in ontological terms, we need in the first place to be able to say in a most basic sort of way what individuals are by identifying them as being a certain kind of thing. That is, implicit in our concept of any object is that it is always some kind of object, in that “it is impossible to say anything about an object without allowing for the possibility of the question, ‘What kind of a thing is it?’” (Shwayder 1965: 41). Quine puts this in terms of the traditional ontological maxim: “no entity without identity” (Quine 1969: 23). This is to say that we decide what kinds of thing can be
14
Framing the issues
said to populate our conceptual universe according to whether they pass the test of being identifiable as a certain kind of thing. Or, “our conception of an object, regarded as a possible object of thought, is to be explained in connection with certain kinds of tests by application of which we are able to individuate particular objects and to establish the existence and identity of objects” (Shwayder 1965: 41–2). These tests might accordingly be thought of as existence tests for kinds of things. Here, I formulate existence tests specifically applicable to the concept of the individual in terms of identity criteria or identity conditions that need to be satisfied when we speak about individuals in general. The two tests or criteria are that it must be possible to show for a given conception of the individual: (a) how individuals are distinct from one another in terms of that conception’s key defining respect, and (b) how they may be tracked through change as unchanged in terms of that conception’s key defining respect. Whether different conceptions of the individual in economics satisfy these tests determines whether they involve satisfactory concepts of the individual, and consequently whether the kinds of individual they describe can be said to exist. The approach here, then, is normative in evaluating different possible accounts of individuals in economics according to a standard. The standard is an ontological one that assesses whether particular accounts of individuals fulfill the most basic requirements for talking about individuals. This strategy differs from a more familiar form of critique of concepts common in economics—one operating epistemologically in terms of what we can know—that juxtaposes a concept and information about the world. For example, it is often said that the subjectivist, preferences-based concept of the individual does not accord well with individuals’ observable reliance on conventions. This method of critique, however, is typically not persuasive to steadfast defenders of a concept, since they can always draw attention to different bodies of empirical evidence that better accord with their concept. The method here, in contrast, investigates the consistency of a concept according to whether it fulfills the basic presuppositions that apply to it as a particular kind of concept. This makes the critique in question one of a concept’s internal consistency, since the basic presuppositions of a concept are part of that concept. Of course, a critique in terms of the supposed basic presuppositions of a concept can always be challenged by questioning those basic presuppositions. Are they the right ones? Are they properly formulated? Did one carry out the ontological analysis that produced them correctly? These questions can be examined independently of their application to a subject matter being investigated, as they long have been by philosophers. However, in this book I concentrate on the task of application in the hope that it will illuminate the most basic requirements involved in talking about individuals in economics in an ontological way. To this end, I hope also to refocus attention on a key question too often neglected in economics: what is economics about?
Framing the issues 1.3.3
15
Theories, models, and reference
Let me place this strategy of ontological investigation in the context of current thinking about economic methodology. An important concern in current economic methodology is how theories, assumptions, and models in economics relate to the world— their realism or their realisticness (Mäki 1989). One view is that theories, assumptions, and models somehow “represent” economic relationships in the world. Yet how this “representation” occurs is controversial, since there are many different ways in which this contested notion can be understood. Consider models as theoretical vehicles in economics. One view of representation is van Fraassen’s (1980) that there is a strict isomorphism between models and the world they represent. Another view is that the relationship between a model and the world is one of similarity (Giere 1988). Departing from this emphasis, Cartwright regards models as idealizations or approximations in that they only describe reality in certain respects (Cartwright 1983). Cartwright’s view, however, suggests we consider what roles models play and more generally what they do. This opens up a range of views about models, such as that models are metaphors (McCloskey 1990), that the relation between the model and the world involves different kinds of analogy (Hesse 1966), that models caricature the world (Gibbard and Varian 1978), and that models may be understood as mediating instruments (Morrison and Morgan 1999). One moral of this history is that the relationship between theories, assumptions, and models and the world to which they relate is multifaceted and complex (Boumans 1999; Boumans and Morgan 2001). Is it still possible, then, to take a meaningful ontological perspective on the theory of the individual, including its main assumptions and the ways in which individuals are modeled? My approach to realism and realisticness is formulated in terms of referentiality. Rather than ask how entire theories and models connect to the world, I proceed somewhat more modestly by investigating the conditions under which their principal terms refer to the world. In the philosophy of language, the theory of reference concerns how linguistic terms and expressions designate or pick out things in the world. The dominant view of reference for much of the twentieth century was the Frege–Russell description theory which held that terms refer by expressing properties actually possessed by the things to which those terms refer. This view was challenged in the 1960s by the causalhistorical theory which explained reference in terms of historical causal chains that traced back the use of terms to an original act of referring. However, with a focus on identity conditions, I approach reference ontologically rather than linguistically by asking whether a theory’s terms can be said to satisfy appropriate identity conditions in regard to what they are meant to pick out. For example, the term “individual” used in a particular theory is used legitimately if it successfully refers to individuals in the world, the determinant of success being that the way the term is used in the theory is consistent with identity criteria for use of the term “individual.” Thus, I make successful reference in this sense a requirement of “good” theories. One motivation for doing so is my belief that theories and models are largely constructed around their objects, and that our theorizing about
16
Framing the issues
relationships between objects is secondary to and dependent upon our intuitions about the nature of objects themselves.5 I believe there is also good reason for thinking about reference in terms of identity conditions (cf. Davis 2001). Though the causal-historical theory of reference has important advantages over the description theory, its reliance on an original historical act of referring ultimately makes reference depend upon something that cannot be explained. An identity condition approach to reference, however, sets aside the origin question and any concerns about the intelligibility of a series of connections between our terms and the original world “out there” and simply asks whether terms in a theory or model justifiably refer to things in the world according to how those terms are conceptualized. Thus, we look at the meaning and content associated with a given term, say, how the individual is defined and explained, and then ask whether this account is consistent with our two identity conditions. This of course does not guarantee that terms that do fulfill identity conditions actually pick out things “out there.” This, however, is something we can never know. The best we can do, rather, is establish what our proper existence tests are, and then determine whether the things we believe exist measure up to those tests. This makes the ontological method pursued here neither the descriptivist sort (Strawson 1959) nor the revisionary sort (Parfit 1986). Rather, the method is Kantian in spirit by asking what must generally be assumed about different kinds of proposed existents, and by then inquiring critically whether those proposed existents satisfy our basic understanding of the kinds of things they are thought to be. As will become clear in the book, I do favor one of the two views of the individual discussed, but that theory is subjected to the same critical analysis as is the theory I do not favor. Let me, then, distinguish the two alternative conceptions of the individual in economics.
1.4
Orthodox economics and heterodox economics
There are of course many different ways to distinguish orthodox and heterodox economics. Nonetheless, an important distinction between orthodox economics and heterodox economics concerns their different conceptions of the individual and their different views regarding the importance of individuals in economics. Orthodox economics, which of course places great weight on individuals in economics, generally treats individuals as relatively autonomous or even atomistic beings. Within orthodox economics I distinguish primarily between neoclassical economics and contemporary mainstream economics, into which I argue the former has evolved, at least with respect to the theory of the individual (also cf. Davis 2002b). Austrian economics is orthodox in its view of the individual as an autonomous being, but it does not share neoclassical or mainstream views of what makes individuals autonomous. Heterodox economics, which places considerably less weight on individuals, generally regards individuals as being embedded in social and economic relationships. To sharpen this distinction, I differentiate between what I
Framing the issues
17
term internalist and externalist types of definition of the individual. Defining individuals in terms of their internal states of consciousness, such as subjective preferences (a form of subjective inwardness), involves what can be termed an internalist definition of the individual. Defining individuals as socially embedded explains them in terms of their external relations to one another—what can be termed an externalist definition of the individual. The differences between the two main approaches reflect their opposed attachments to two long-established, competing methodological traditions in economics, namely, methodological individualism and methodological holism. Methodological individualism and methodological holism can both be characterized epistemologically as strategies of explanation and also ontologically as views about the nature of agency in economic life. As strategies of explanation, an important goal is the unification of thinking under a central explanatory paradigm (Kitcher 1993). Both methodological individualism and methodological holism can be understood in this way as unifying explanatory paradigms. As to views about the nature of agency in economic life, an important goal is to provide causal accounts of the world that specify mechanisms that bring about events (Kincaid 1996). Both methodological individualism and methodological holism can also be understood as ways of accounting for cause-andeffect relationships. One important conclusion of this book is that neoclassical and mainstream economics, which make the individual central to their analysis, lack an adequate conception of the individual. If this conclusion is true, it would be significant for our understanding of orthodox economics as both an explanatory and cause-and-effect framework. However, it would also be significant for our understanding of individuals in contemporary society, since critiques of the modernist concept of the self represent challenges not just to a set of historical ideas adopted by orthodox economists but also to one of the principal defenses of our inherited normative vision of individuals as being centrally important. Indeed, because orthodox economics is widely thought to offer a particularly strong defense of the individual, wrapping an inadequate account of the individual in the banner of liberalism would doubly risk jeopardizing society’s defense of the individual. A second important conclusion of this work is that heterodox economics, which does not generally emphasize the individual, in fact offers elements of an adequate theory of the individual. If this conclusion is true, it would also be significant for our understanding of economics as both an explanatory and cause-and-effect framework. However, it arguably would be more significant for our wider view of individuals, because it would demonstrate that the individual can remain a primary normative focus in spite of being socially embedded. Indeed, successfully relocating our understanding of the individual into a conceptual space generally claimed by the critics of the individual would offer an especially promising strategy for sustaining the individual as an epistemological, ontological, and normative concern.
18
Framing the issues
Thus, more is at stake in this investigation than competition in the history of economics. The influence of economics today is the more important matter— indeed, its influence in a domain of central human concern. The goal of this book, then, is to bring the theory of the individual into the foreground in economics, and in the process bring some illumination to the subject.
1.5
Plan of this work
This book is divided into two halves. The first is devoted to the orthodox conception of the individual as a relatively autonomous atomistic being, and the second is devoted to the heterodox conception of the individual as a socially embedded being. The chapters in the two parts correspond to one another, though not quite in a parallel manner. Chapters 2 and 6, which open Parts I and II respectively, provide overviews of the theory of the atomistic individual and the theory of the embedded individual. Chapters 5 and 9 close Parts I and II and provide accounts of what I believe to be the destination towards which each of the theories is tending. In between these bookend chapters I apply the identity analysis described above. Chapters 3 and 8 investigate the problem of reidentifying individuals in orthodox and heterodox economics respectively. Chapters 4 and 7 investigate the problem of individuation in orthodox and heterodox economics respectively. Chapter 10 concludes the book with a summary and discussion of issues left unaddressed. The nonparallelism between the middle chapters of the two parts concerns the order with which the problems of individuation and reidentification are addressed in each. In Part I, devoted to orthodox economics, I address the problem of reidentification first (Chapter 3). Reidentification presupposes some conception of the individual, since one must first have a distinguishable individual if one is to reidentify that selfsame individual across change. Since orthodox economics is built around the assumption that the individual as relatively autonomous is a distinguishable being, I begin by granting the theory the benefit of the doubt, and initially determine whether the individual as conventionally understood is reidentifiable through change. Since problems arise in this reidentification analysis, I then go on to ask whether the atomistic individual can even be distinguished as a separate being (Chapter 4). In Part II, devoted to heterodox economics, I investigate the problem of individuation first (Chapter 7). Because the idea of the individual as a socially embedded being raises questions from the very outset regarding whether individuals can be distinguished from one another, the problem of individuation is primary. Indeed, many would no doubt be inclined to say that heterodox economics cannot even have a theory of the individual. However, I argue that there is a way of understanding embedded individuals as distinct beings. From this vantage point I go on in the next chapter to ask whether socially embedded individuals can also be reidentified across change (Chapter 8).
Framing the issues
19
Overall, the framework of analysis in this book involves examining how successfully two different conceptions of the individual stand up to two different identity criteria tests. Thus, the book makes four arguments, and finds that orthodox economics fails both to individuate and to reidentify individuals as it understands them, and that heterodox economics succeeds in individuating and reidentifying individuals as it understands them. These conclusions are certainly sensitive to how the two theories and views of the individual are represented in connection with each of the arguments. It is a fair response to say that there are many different currents within orthodox economics and also many different currents within heterodox economics, and that, under closer inspection of these other currents, the conclusions reached here may not bear up for them. However, it is my view that the different streams within orthodox and heterodox economics at bottom share the basic alternative views of the individual that I have set forth as broadly orthodox and heterodox. I thus leave it to others to demonstrate whether the conclusions reached here do or do not apply to other views of the theories.
Part I
Orthodox economics
2 The atomistic individual
The limit of objectivity with which I shall be most concerned is one that follows directly from the process of gradual detachment by which objectivity is achieved. An objective standpoint is created by leaving a more subjective, individual, or even just human perspective behind; but there are things about the world and life and ourselves that cannot be adequately understood from a maximally objective standpoint, however much it may extend our understanding beyond the point from which we started. A great deal is essentially connected to a particular point of view, or type of point of view, and the attempt to give a complete account of the world in objective terms detached from these perspectives inevitably leads to false reductions or to outright denial that certain patently real phenomena exist at all. (Nagel 1986: 7) Polonius [behind] What, ho! help, help, help! Hamlet [drawing] How now! a rat? Dead, for a ducat, dead! Polonius [behind]
O, I am slain!
[Makes a pass through the arras]
[Falls and dies]
(Shakespeare, Hamlet, Act III, scene iv)
Reason is wholly instrumental. It cannot tell us where to go; at best it can tell us how to get there. It is a gun for hire that can be employed in the service of any goals we have, good or bad. (Simon 1983: 7–8)
This chapter provides an overview of the atomistic individual conception, tracing its evolution from the neoclassical subjectivist conception of the individual to the current mainstream view of the individual as a bearer of strategies in contemporary game theory. The argument of the chapter is that neoclassical economics began with a Lockean understanding of the individual as a being defined in terms of a private psychology, but subsequently gave up this foundation in psychology for a purely formalist conception that I label the abstract individual conception of mainstream economics. Indeed, I differentiate mainstream economics from neoclassical
24
The atomistic individual
economics in terms of their two respective conceptions of the atomistic individual. I argue, moreover, that, because this evolution was driven by only a vague awareness of the problematic nature of Locke’s subjectivist view of the individual, the mainstream’s resulting abstract individual conception constituted a largely unintended outcome. Since Locke operated in a dualist framework, giving up the subjective side of that framework meant—in the absence of an alternative substantive way of reconceptualizing the individual—that mainstream economics ended up giving an explanation of the individual “from a maximally objective standpoint” (Nagel 1986: 7). This raises the question, is mainstream economics still about human individuals? I argue that it has no systematic way of saying that it is.
2.1
Locke’s legacy
One way to understand the autonomy of atomistic individuals—the way that Locke understood it—is to suppose that there must always be something about individuals that remains unchanged when they interact with others. In neoclassical economics, the private states of consciousness of individuals, to which they alone have access, remain unchanged—first individual tastes and desires in cardinal utility theory and then later preferences in ordinal utility theory. Individuals do not interact on this subjective level, but interact only with one another at a significant remove according to how their inner psychologies are manifested in their actions in markets. Thus, there always hangs a curtain, as it were, between the individual and the world of action and interaction with others. Locke, we saw, laid the groundwork for thinking about the individual in this way when he characterized consciousness as something disengaged from the world. However, Locke’s dualist separation of individuals from the world came with an inherent dilemma regarding how the inner subjective worlds of individuals linked up with the outer objective world. Which discourse should we use to explain the connection between inner and outer worlds, particularly when we want to talk about economic life and the social world? If the subjective side of things is emphasized, it might be difficult to resist the conclusion that a world made up of private, subjective individuals is a chaotic, Hobbsean sort of world. Alternatively, were Newton’s vision of the world given precedence, individuals might not need to play any part at all. This dilemma was addressed in The Wealth of Nations when Smith brought both worlds together in his use of the Scottish Enlightenment concept of unintended consequences and the idea that the market worked as if by an invisible hand (Smith 1976 [1776]). Individuals brought their private subjectivities to bear upon one another in competition and exchange, but only at a distance in the form of their actions in buying and selling. This made it possible to treat the market as a fundamentally natural, Newtonian-type process understood in terms of equilibria and causal connections, which nonetheless presupposed—behind the scenes, as it were—the private and subjective interests of individuals. The metaphor of the invisible hand ideally captured this
The atomistic individual
25
combination, because it allowed the market to be explained as an observably natural, mechanical process that depended upon subjective interest as the invisible force that drove it. However, in providing a metaphorical solution to the Enlightenment’s dualism problem, Smith did not resolve it. On Smith’s understanding, because the market is a systematic and unintended process fully within the realm of mechanical nature, it can and ought to be explained in the language of science in cause-and-effect terms. Yet Smith did not explain the precise mechanism by which human psychology produced its effect on market outcomes. Were one to focus exclusively on self-interest, one could of course explain competition. However, Smith also believed that individuals possessed a variety of different types of motivation, including sympathy for others. To give a fully natural account of the market, then, he would have had to explain the mechanics of human psychology. Thus, while Smith’s image of an invisible hand operating in markets is a commanding one, it does not really explain how private tastes and desires determine the operation of markets. Rather, what Smith really provided was an account of competitive behavior loosely linked to an underlying psychology. However, Newton could also ascribe “behaviors” to physical bodies in nature, such as the planets. And then, just as the planets lack inner lives but are fully explainable in terms of their “behaviors,” so, from the point of view of economics as a natural science, individuals could lack inner lives and yet be fully explained by their “behaviors.” Having gone this far, however, one might ask whether individuals need have any role in the explanation of economic systems. Individuals might seem to be the most important units of analysis, but without there being a specific role for their unique characterization as subjectively inward they might be no more important than other possible units of analysis. Much of classical economics as an aggregative form of explanation essentially drew this conclusion in making economic classes its preferred units of analysis. However, at the end of the nineteenth century a new strategy for linking Locke’s two worlds emerged. Arguably, the remarkable achievement of early neoclassical economics was to have found a way of reconciling individual subjectivity with objective nature via the theory of choice. From this perspective, Gossens, Jevons, Menger, Walras, and Marshall were specifically neoclassical in showing how Smith’s inadequate resolution of the problem of Enlightenment dualism could be better addressed by understanding individual behavior specifically as choice behavior. Their argument to this effect can be reconstructed as follows: 1 2
Individuals are unique among all the possible units of analysis in economic life because their behavior alone can be understood in terms of choice. Choice can only be explained by reference to individuals’ inner states, that is, their private tastes and desires, because this is the basis on which individuals discriminate among their options.
26
The atomistic individual
3
Individuals are able to discriminate among their options because they apply marginalist principles. Marginalist principles explain markets because they account for the determination of prices.
4
The key to this argument, of course, was in the last step, and in showing exactly how marginalist principles produced prices in a supply-and-demand framework. Smith had only provided a metaphorical treatment of how subjective interest was somehow at work in the market. However, the early neoclassicals explained how subjective interest became material, for example, in using the marginal utility concept to derive the individual demands for goods. They thereby made what the individual wants the explanation of what the individual does, and accounted for the objective world in terms of the subjective world. This linkage later found powerful expression in the notion that choice could be formalized as “rational” choice. The term “rational” can refer to both mental phenomena (in terms of logical soundness) and natural phenomena (in terms of natural laws). “Rational” choice, then, provides the solution towards which Smith had only gestured. Whereas Smith and the classics had left a curtain between subjectivity and the world, neoclassicism, as the Hamlet of economics, penetrated that curtain with the theory of choice. However, neoclassicism’s pass through the arras was not without consequences. To make choice truly rational so as to comport with a supply-and-demand market system explained in Newton’s language of equilibria and cause and effect, the hidden world of taste had to be exposed and tamed—indeed, ultimately regimented, as it was within an axiomatic structure. The diffident prince of Denmark thus never quite knew who was slain, but only that there was an enemy there to be dispatched, and how this could be done. Thus, I argue in the next section that mainstream economics emerged from neoclassical economics through a progressive realization that subjective inwardness was altogether incompatible with a scientific world view. However, with the elimination of subjectivity, the very basis on which individuals had been understood as independent beings was also eliminated. Smith’s problem was finally solved, then, not by making a better connection between the subjective and objective, but rather by doing away with the former, escaping Locke’s dualism, and emptying out all psychological content from the neoclassical conception of the individual. No longer required as an intermediary between taste and action, choice then became a general, all-purpose logic that could be applied to any sort of agent, single individual, multi-individual, human, or otherwise. In this chapter, I thus first look at the history of the atomistic individual from neoclassical to mainstream economics, and then consider questions that this account raises about the atomistic conception. Section 2.2 reviews the historical steps from cardinal utility theory to ordinal utility theory to the Chicago School, then from revealed preference theory to axiomatic preferences, and closes with brief attention to new mainstream views on ethics and altruism. Section 2.3 looks at what is involved
The atomistic individual
27
in understanding the individual as an atomistic being, and argues that the abstract individual is still the conception of an atomistic being. Thus, mainstream economics continues under the banner of individualism though without an account of what individuals are. Section 2.4 asks whether methodological individualism and its associated reductionist project might constitute a last-ditch defense for keeping a place for the human individual in mainstream economics. In Section 2.5, I look at the abstract individual conception in its principal current location in mainstream economics: game theory. Finally, Section 2.6 links the chapter’s conclusions to the thinking about individual identity introduced in the last chapter.
2.2
The evolution of the atomistic individual conception
The theory of choice is about being instrumentally rational. Instrumental rationality is defined as the choice of actions that best satisfy an individual’s ends or objectives however those ends or objectives may happen to be characterized. Instrumental rationality is a rationality of efficient means, and per se is completely agnostic regarding the nature of the ends those means serve. Utility analysis is accordingly only one form that rational choice theory can take when an individual’s aim is specified as achieving maximum utility. However, the theory of instrumentally rational choice does not require that utility or anything else related to an individual’s private psychology be the basis for choice. Historically, this flexibility in the theory facilitated a relatively seamless evolution from the neoclassical to the mainstream conception of the individual. In that history, the theory of choice was progressively emptied of any and all reference to individuals’ subjective ends, to the point that current mainstream choice theory refers to individuals’ ends entirely formally in terms of objective functions. Even the expression “objective function” signals final departure from Locke’s view of the individual as subjective and the adoption of a view of individuals more appropriate to Newton’s object world. 2.2.1
Cardinalism to ordinalism to the Chicago School
The nineteenth century neoclassical marginalists were quite explicit about applying a subjectivist understanding of the individual to the theory of choice. In the cardinal utility accounts of choice of Gossens, Jevons, Menger, Walras, and Marshall, an individual’s ends or objectives are associated with a psychology of wants and desires. There had been precursors. Hobbes, in terms of desire and aversion, and Bentham, in terms of a hedonistic, pleasure and pain calculus, had laid the original foundations for a utility understanding of choice, promoting the idea that human psychology thus understood represented scientific advance. The early marginalists understood utility as usefulness, satisfaction, or happiness, but still saw it as a psychological quantity that was measurable in principle just like an individual’s weight and height. Cardinal utility theory thus provided the foundations for neoclassicism’s early achievements in
28
The atomistic individual
connection with the explanation of demand and elaboration of the “old” welfare economics. However, there were also doubts about the scientific standing of the concept of utility at the end of the century. Jevons used a mechanical kind of reasoning to explain psychology, and there were a variety of psychophysiological accounts of mind and mental activity that were then being explored (Maas 2001). Pareto (1971 [1909]), however, took what is generally regarded as the first important step towards contemporary choice theory when he showed that ordinalist analysis made it possible to avoid making any essential reference to human psychology in explaining individual choice. Individuals were said to prefer one combination of goods to another, but the language of preferences contained very little in the way of psychology. In fact, as applied to combinations of goods, “prefer” only means that these combinations can be rank ordered, and rank ordering is a relation that has nothing in particular to do with human psychology. It can be applied to individuals, but it can also be applied to agents of any kind that can be claimed to discriminate options, whether these agents be individuals or groups. After Pareto, individual preferences were only nominally subjective. They could still be characterized as preferences of human individuals, but in comparison with the older cardinal utility framework there was not much that linked them to the psychology of human individuals. Subsequently, Hicks and Allen (1934) systematically developed Pareto’s insight in the indifference curve framework. Indifference curves originated with Edgeworth and were refined by Pareto, but Hicks and Allen showed they could be used to give a complete theory of consumer behavior on the basis of ordinal utility. They rediscovered Slutsky’s earlier contribution, and his symmetry analysis for maximizing utility subject to a budget constraint, coupled with the assertion that there was “[no] point of contact whatsoever between economics and psychology” (Slutsky 1953 [1915]: 43), gave additional impetus to the reformulation of utility theory as ordinalist (cf. Stigler 1950). A central part of this reformulation involved their explanation of the diminishing marginal rate of substitution. The earlier cardinal analysis had employed diminishing marginal utility to generate demand behavior, but Hicks and Allen demonstrated with convex indifference curve maps that the same results could be achieved without recourse to the cardinal utility concept. The concept of indifference as the underlying basis for understanding preference, moreover, had the virtue of describing the individual choice in terms of a position vis-à-vis possible consumption bundles. It thus shifted the focus from an internal to an external space in which individuals acted, and made observable relationships central to understanding economic behavior. In a different way, Robbins (1935 [1932], 1938) also reinforced Pareto with his well-known critique of interpersonal utility comparisons. Though the cardinal utility theorists had thought human psychology sufficiently intelligible to justify making such comparisons, and then using them to make normative recommendations, Robbins’ skeptical argument received an assist from changing views about the relationship between psychology and moral judgment in the first decades of the century.
The atomistic individual
29
Emotivist thinking in ethics (a companion movement to positivism in science and philosophy) treated moral judgments as expressions of attitude or emotion (Ayer 1946 [1936]; Stevenson 1937). Earlier, utilitarians had believed moral judgments were objective, because these judgments derived from an understanding of how human psychology operated. G. E. Moore (1903) broke this connection by charging that it committed the naturalistic fallacy by supposing that statements regarding what was good or bad followed from statements describing the world (in this case, statements regarding what produced satisfaction). However, Moore’s own Platonism and intuitionist basis for moral judgments was no more secure, and the emotivists concluded shortly thereafter that moral judgments, as expressions of attitude or emotion, were essentially irrational. In this way, they succeeded in discrediting both moral judgment and human psychology. Robbins’ arguments regarding the irrationality of interpersonal utility comparisons were made in this climate, and had the effect in economics of ruling out interpersonal utility comparisons in normative recommendation (Davis 1990). With interpersonal utility comparisons ruled out, the space for normative recommendation in economics was thus narrowed to the Pareto efficiency principle, and cardinal utility theory’s hedonistic psychology ceased to figure in the future development of the theory of choice. There were isolated rearguard efforts to sustain a psychological interpretation of utility (cf. Backhouse 2003; Giocola 2003). Cassell had given an account of demand without any reference to utility whatsoever, arguing that the latter added nothing to our knowledge of pricing (Cassell 1932 [1923]). In reply, Wicksell made a case for the psychological interpretation of utility by claiming that one could only understand the behavior of real-world economic agents when one thought in terms of the motivations behind their actions (Wicksell 1934–5 [1918]). Moreover, Pigou and the Cambridge School followed Marshall, who saw wants as the “motor-force” behind individual action (Marshall 1920 [1890]: 16). Pigou reasoned in terms of the “satisfactions and dissatisfactions” behind behavior, and continued to think of utility as measuring the intensity of desire (Pigou 1932 [1920]: 23). However, the dominant trend was away from psychology. In Chicago, Schultz organized a statistical laboratory to explain the demand not just of consumers but also of producers. Schultz was a student of Henry Moore, who had used regression analysis to explain demand, and had found the demand for pig iron to be upsloping (Moore 1914). Following Bridgeman, Schultz insisted that concepts were only meaningful when it was possible to identify operations on which they were based (Schultz 1928). Schultz later joined forces with Hotelling (cf. Hands and Mirowski 1997), and they built upon Slutsky, Hicks, and Allen. Schultz’s The Theory and Measurement of Demand (1938) attempted to provide a statistical explanation that justified the hypothesis of rational consumer behavior, but was unsuccessful. This created a dilemma regarding the future direction of utility theory which ultimately played out in such a way as to further reduce the role of psychology.
30
The atomistic individual
An important part of the story concerns the postwar Chicago School’s ambivalence towards utility theory, particularly in the form of Friedman’s “as if” instrumentalism. As Mirowski and Hands relate (Mirowski and Hands 1998), Knight had contested the Hicks–Slutsky classical mechanics-inspired approach to demand theory, insisting on a psychology-based account of demand rooted in human motives (Knight 1944). However, Friedman, in his famous article on the Marshallian demand curve (Friedman 1949), though he claimed Knight’s mantle, reasoned in an “as if” manner that it was unnecessary to think in terms of utility functions to be able to explain demand. The same “as if” thinking was then codified in Stigler’s The Theory of Price in which he further minimized the importance of utility functions (Stigler 1952), and was finally given definitive expression by Stigler and Becker (1977) in their paper “De gustibus non est disputandum” which made preferences essentially insignificant in the determination of price by assuming them neither to vary nor to differ importantly between people. Nonetheless, the Chicago School retained the appearance and sometimes the spirit of engaging in utility analysis, even as they undermined its role. This was not the case with Samuelson at M.I.T. or for those associated with the Cowles Commission. 2.2.2
Revealed preference and axiomatic preferences
The step most often associated with fully eliminating psychology from choice analysis was taken by Samuelson in his revealed preference approach (Wong 1978; Hands and Mirowski 1998). In his 1938 paper he proclaimed that his goal was “to develop the theory of the consumer behavior freed from any vestigial traces of the utility concept,” and thus he carefully avoided using the language of preferences or anything that referred to the subjectivity of individuals (Samuelson 1938). In his paper a decade later in which he introduced and defended the revealed preference concept as operationally observable, preferences do not dictate an individual’s choices, but rather an individual’s choices are said to be revealed preferred according to the observable choices the individual makes (Samuelson, 1948). Samuelson did shift his emphasis to argue that revealed preference theory was only observationally equivalent to ordinal utility theory rather than a substitute for it, but from the point of view of the profession the damage was done. One could make reference to human psychology in explaining economic behavior, but doing so was now seen as unnecessary. With his ambition of making economics “scientific,” Samuelson thus resolved the problem of Enlightenment dualism by dropping Locke’s subjective inwardness side and recasting the world in pure Newtonian terms (best modeled, in his view, on equilibrium thermodynamics). The influential implication of Samuelson’s framework was that it demonstrated that choice could be explained in purely formal terms. At the same time, formal representation of choice also had the advantage of being able to accommodate any and all possible kinds of economic agent that might be assigned objective functions. Moreover, since human individuals were one such
The atomistic individual
31
agent, they were not eliminated but merely reinterpreted. In particular, they were interpreted as one kind of abstract individual. Yet perhaps an even more important step taken in eliminating the human individual from economics came from those in the Walrasian general equilibrium theory tradition—Arrow, Koopmans, Marschak, Hurwicz, and others—many of whom passed through the Cowles Commission in Chicago in the immediate postwar period (Mirowski 2002). Here, the abstract mathematical reconceptualization of the neoclassical economic agent as an expected utility maximizer was developed in terms of the von Neumann– Morgenstern (1944) proof of a representation theorem for preferences in terms of the now familiar asymmetry, transitivity, continuity, and independence axioms (Guala 2000). With proof of the representation theorem, preferences finally lost their psychological characterization altogether, since their interpretation now depended on their formal specification rather than on their description as natural phenomena. Of course, preferences defined axiomatically can be associated with human preferences, but this can only be done on an ad hoc basis. Recognition of this has led most rational choice theorists to label the theory a normative rather than a descriptive theory. The issue had thus become not how best to represent human choice but whether human beings measured up to the standards of abstract rationality. Current mainstream economics, then, makes the discussion of individual psychology an incidental dimension of the explanation of choice. This has provoked considerable criticism in recent years from psychologists and a growing number of behavioral economists (Rabin 1998) who have argued that, to explain choice, one must begin with human psychology. However, in an important sense, these criticisms are irrelevant, because the theory is not about human psychology, at least in the view of most of its proponents. Rather, it is about rationality. This is reflected in the standard strategy of taking individual preferences as “given,” meaning that their character is irrelevant to the analysis of rational choice. Alternatively, preferences are said to be “exogenous,” meaning that they stand outside the patterns of determination that explain the economy. Either way, preferences as specifically human subjective phenomena now fall outside the domain of explanation. Were they influenced by economic events, or endogenous, we might be led to develop a better understanding of them and how human psychology explains choice. However, taken as given or exogenous, issues of this sort can never arise. Individuals thus located outside economics no longer played a role in economics. 2.2.3
Altruism and ethics
A final reflection of this development in choice theory concerns new thinking about altruism and ethics. An old criticism of neoclassical economics was that it portrayed the individual as selfish and therefore could not explain such things as altruism. If selfishness has roots in the Lockean idea of the individual as motivated by inner, private concerns, this criticism is basically correct. Since the 1970s (cf. Fontaine 2000),
32
The atomistic individual
however, mainstream economists have argued that there is no inherent reason to suppose that preferences in the ordinalist framework must be selfish (or selfregarding), or that they might not be altruistic (or other-regarding). Collard was instrumental in developing this view when he drew on Edgeworth to argue that individuals used “coefficients of effective sympathy” to measure the weight given to the “utility” of another in an individual’s “utility” function (Collard 1978). This made individual “utility” functions interdependent, and undermined the old view that Homo economicus was necessarily self-interested. Becker (1976a) used this idea of interdependent “utilities” in his famous rotten kid theorem which explained altruistic or cooperative behavior of individuals as adaptive (also cf. Hirschleifer 1976). In effect, there is nothing that commits rational choice analysis to any particular view of the nature or content of preferences, and the door is thus wide open to making any assumptions about their character that may suit the investigation at hand. It is true, of course, that there are problems with these new interpretations of altruism from both normative and evolutionary perspectives (cf. Fontaine 1997, 2000; Vromen 1995: 152ff.). However, the point here is that the later development of choice theory allowed for alternative explanations of such phenomena, whereas earlier formulations of the theory were too tightly tied to self-regarding if not “selfish” explanations of individual behavior to admit them. That earlier association was due to choice originally being explained in terms of Bentham’s hedonistic psychology. But the gradual distancing of choice theory from any essential connections to human psychology cut the theory of choice free of the limitations of utility analysis, and created significant new opportunities for mainstream economics that were not available earlier. What has been overlooked in this expansion of opportunities, however, was that no new view of the individual had been introduced to replace the abandoned Lockean view. Thus, there was no particular reason to think these alternative explanations of behavior ought actually to be ascribed to human individuals.
2.3
The requirements of atomism
In light of this history, recall Taylor’s characterization of Locke’s concept of the individual as a “punctual” self. For Locke, the self as subjective inwardness was an “extensionless” geometric point (Taylor 1989: 172), because a genuinely subjective nature cannot occupy space in a Newtonian world of objects and their relations. However, we have seen that there is another meaning to being “extensionless,” namely, that the subjective self is also inaccessible and without meaning in the language of science. Emphasizing this sense of the concept suggests that the inevitable destination of the idea of the individual as subjective inwardness is its own elimination from the world of nature. That is, the subjective must remain subjective. But if the abandonment of Locke’s subjectivist view undermined the conception of the individual itself, then is the abstract individual still even a conception of the individual? Or is the correct conclusion that the individual disappears altogether in
The atomistic individual
33
mainstream economics? In this section I argue that, while the abstract individual conception makes no reference to human individuals, it still functions satisfactorily as a purely formal account of atomistic individuals. This will ultimately be valuable for understanding the different kinds of referent the conception can be used to pick out in the world, and for understanding the basis on which this referring occurs. 2.3.1
Autonomy from other individuals
For Locke and neoclassical economics, individual autonomy is understood in terms of what is substantively unchanging about individuals. However, rather than concentrate on some unchanging characteristic of individuals, we might alternatively think of individual autonomy simply in terms of individuals’ degree of connectedness to one another. At one end of the spectrum this involves the idea that individuals are entirely explainable in terms of relationships to others, that is, that individuals are internally related to one another. Two individuals are internally related if their nature fully depends upon their relation to one another. For example, being a parent fully depends upon having a relationship to a child. At the opposite end of the spectrum, individuals may be externally related in that their nature is in some sense independent of their relation to others. For example, “adult” and “child” are externally related in that the nature of being an adult is independent of having a parental relationship to a child. This latter conception, then, represents a formal or nonsubstantive basis for treating individuals as autonomous. To use it, one simply provides a finite list of those respects in which the individual can be influenced through interaction with others, leaving the individual externally related in all other respects. Autonomy is thus defined residually rather than substantively. Consider the conception of the individual in general equilibrium models of market exchange. General equilibrium models are mathematical systems of interdependence exhibiting complex relationships between atomistic agents. In comparative statics exercises, only a finite list of things about individuals can change as a result of their interaction, such as the resources at their disposal and their post-trade demands and supplies. Having determined what goes on this finite list, however, there is no need to say what does not go on it, that is, in what specific respects individuals remain autonomous. Indeed, since general equilibrium models are almost always formulated mathematically, the autonomous individual is always represented in terms of some objective function whose arguments have no meaning other than to distinguish one individual from another. Of course, individual objective functions are often informally called utility functions, and their maximization is loosely associated with the idea that an individual’s utility is as high as possible. However, as mainstream economists would be the first to admit, this really has nothing to do with individual psychology. Utility functions are only a formal device used to differentiate economic agents. Thus, individuals can be treated as autonomous from one another without any substantive account of what makes them so.
34
The atomistic individual
2.3.2
Autonomy from supraindividual entities
Another dimension of atomism involves its contrast with holism, where holism is the idea that the parts of a whole exist only in relation to the whole, and are thus only properly conceptualized when understood in relation to the whole. For example, on a holist reading, individuals might be seen as citizens of a nation or as members of a class, and as having no identity apart from such membership. Thus, whereas organicism requires that individuals’ relationships to other individuals define who they are, holism requires that their relationship to supraindividual entities defines who they are. From this perspective, atomism in the traditional Lockean sense requires that there be something about individuals that is independent of such relationships and moreover is central to the understanding of the individual. However, this position can also be interpreted in an entirely formal manner so as to avoid any substantive reference to what specifically might provide individuals with a measure of independence from supraindividual entities. Arrow (1959) paved the way in this regard in connection with his holistic treatment of the concept of market equilibrium. Recognizing that individual action cannot itself account for how market equilibrium comes about, he treats the institutions of the market as a framework within which individual action occurs. Individuals who are (relatively) autonomous in this framework have preferences that can be explained in a purely formal way in terms of the idea of an ordered pair. Their autonomy is presupposed rather than explained, and the result is a “moderate” holism involving the institutions of the market (the system of private property, the apparatus of civil society, patterns of trust and moral behavior, etc.) that structures interaction between relatively autonomous individuals without our needing to say anything about them in particular. Nonetheless, it might be argued that a purely formal treatment of the individual eliminates another key element in the Lockean view, namely, that individuals are the only source of economic activity. Explaining individuals psychologically not only accounts for their autonomy but also attributes a source of motivation to them that can then explain action. Does Arrow’s formalist “moderate” holism create problems for thinking that individuals are the sole sources of economic activity? Those who allow supraindividual entities to play a role in individualist explanations in economics offer a straightforward answer to this question. Hayek may be the most clear in maintaining that supraindividual entities are merely concepts rather than real things (e.g., Hayek 1952). Only individuals really exist, and thus only they can be considered agents and sources of action in the economy. This “by default” answer clearly does not require any explanation of what individuals actually are—something about which Hayek and Arrow would disagree, while yet agreeing about the nature of the supraindividual. As Arrow succinctly puts it, “Society, after all, is just a convenient label for the totality of individuals” (Arrow 1984: 80; emphasis added). Later expressions of this view treat supraindividual concepts simply as a “convenience” in economics that economic theories may employ instrumentally in the form of “as if”
The atomistic individual
35
devices to produce good predictions (Friedman 1953). Lucas emphasizes the temporary nature of such “conveniences” when he claims that individual microfoundations will ultimately be developed for all nonindividual macroeconomic concepts, so that “the term ‘macroeconomic’ will simply disappear from use and the modifier ‘micro’ will be superfluous” (Lucas 1987: 108).
2.4 Methodological individualism and reductionism: a last-ditch defense? Thus, the abstract individual conception can explain individuals as atomistic beings on a purely formal basis. However, Hayek’s assertion that supraindividual entities are merely concepts suggests that mainstream economics might be indirectly about human individuals. Though as a purely formal conception the idea of the abstract individual itself makes no direct reference to human individuals, nonetheless it might be argued that this conception, as a matter of fact, can be applied only to human individuals. Suppose that we take firms, families, governments, and other multi-individual entities as paradigm examples of abstract individuals, since their “preferences” by nature cannot have psychological foundations. Then, since Hayek views supraindividual entities only as concepts, their choice behavior must involve some kind of shorthand for the ensemble of individual choices that go on within them. Thus, in the final analysis the mainstream conception of the abstract individual could still be fundamentally about human individuals, albeit only contingently so, because, as a matter of fact, only individuals are said to exist. This “back door” argument depends on our having some confidence that only individuals exist. Not everyone, of course, believes that only individuals exist. However, a prima facie case for thinking this might be true could be made by showing that the choice behavior of multi-individual entities can always be explained in terms of the behavior of individuals (but not the reverse). Then, individuals would at least be more fundamental than supraindividual entities, arguably implying that they alone are the world’s real existents. This, then, is the reductionist project of mainstream economics, the roots of which rest in methodological individualism, a stance taken originally in turn-of-the-century neoclassical economics and subsequently defended in the microfoundations reductionist project of mainstream economics. Thus, in this section I examine whether this last-ditch defense for mainstream economics still being about human individuals is successful. 2.4.1
Methodological individualism
The expression “methodological individualism,” it seems, was originally used by Schumpeter as early as 1908 (Blaug 1992 [1980]) to refer to what he later characterized as the type of investigation that began with “behavior of individuals without going into the factors that formed this behavior” (Schumpeter 1954: 889).
36
The atomistic individual
Schumpeter reserved the expression “sociological individualism” for what he regarded as the “untenable” view that “all social phenomena resolve themselves into decisions and actions of individuals that need not or cannot be further analyzed in terms of supraindividual factors” (Schumpeter 1954: 888). “Sociological individualism,” in fact, is what most people in economics and other social sciences today understand by the term “methodological individualism.” Whereas methodological individualism in Schumpeter’s original sense meant that explanations at the level of the individual were only especially informative in economics, methodological individualism in the stronger contemporary sense generally has a prescriptive meaning in recommending that economic explanation of supraindividual entities ultimately be reduced to or translated into accounts of the decisions and actions of individuals. For example, Elster defines methodological individualism as “the doctrine that all social phenomena (their structure and their change) are in principle explicable only in terms of individuals—their properties, goals, and beliefs” (Elster 1982: 453). Thus, contra Schumpeter, methodological individualism has come to be associated with reductionism, or the view that not only can one always in principle replace explanations of social entities by some individualist explanation, but one ought to do so whenever practically possible. 2.4.2
Reductionism
Reductionism according to philosophers of science most often means theory reduction and involves showing how one (eliminable) theory can be shown to be a special case of another, more fundamental, theory. For example, Newton demonstrated that Kepler’s laws were a special case of his own laws, and Einstein later showed that Newton’s laws were a special case of his own laws. Kincaid (1996: 145ff.) explains how theory reduction depends upon two fairly demanding requirements. First, we need to be able to translate the terms of the theory to be eliminated into the language of the more fundamental theory via “bridge laws” (a “connectability” requirement). Second, we need to be able to derive all the explanations in the theory to be eliminated from the more fundamental theory without at the same time presupposing any elements of the theory being eliminated. For example, were we to provide an individualist explanation of discrimination to replace social or holist explanations of discrimination, as does Becker (1971 [1957]), first, the language of a social practice would need to be framed in individualist terms—for Becker the idea that individuals have a “taste for discrimination”—and second, we would need to be certain that the factors involved in social explanations of discrimination were not being presupposed in our individualist explanation of discrimination, say, in accounting for how individuals acquire a “taste for discrimination” (cf. Kincaid 1996: 156–9). Becker effectively sidesteps this more difficult second requirement by treating a “taste for discrimination” as one of a number of factors that are exogenous to his explanation or model of discrimination. However, that he chooses to use models that neutralize
The atomistic individual
37
such factors does not imply that his individualist explanation of discrimination produces a satisfactory reduction of social accounts of discrimination. If anything, Kinkaid suggests, this strategy seems to signal that individualist reductions of social factors are not likely to be successful. As it turns out, in spite of the broad appeal of reductionism to mainstream economists, few have actually attempted systematically to reduce theories of supraindividual entities to the decisions and actions of collections of individuals. Rather, the more common strategy is to take an easy shortcut by first substituting individualist explanations for holistic ones, and then arguing that the latter can in principle be explained in terms of the latter. Perhaps the best illustration involves the project of developing microfoundations for macroeconomic relationships motivated by the belief that theorizing about aggregate relationships needs to be reinterpreted in terms of the behavior of individual economic agents (Weintraub 1979; Janssen 1993). Most notably, in the New Classical Economics “counter-revolution” against traditional Keynesian economics, Lucas, Sargent, Barro, and others attempted to create microfoundations for macroeconomics by developing Walrasian Arrow–Debreu– McKenzie type general equilibrium models based only on individual preferences, endowments, and technology so as to explain such aggregate phenomena as fluctuations in unemployment. Microfoundations models of this kind, however, are reductionistic in only a qualified sense, because the existence proofs that underlie general equilibrium theory, and on which these models accordingly depend, make use of a number of “global” assumptions that cannot be derived from individual behavior (Arrow 1968: 382; Rizvi 1991; Janssen 1993). These assumptions violate the second of Kinkaid’s two requirements for successful reduction, namely, that all nonindividualist factors are excluded from the analysis. However, the situation is even worse than this. A death blow was dealt to the entire microfoundations reductionist project with Sonnenschein’s impossibility theorem and the Sonnenschein–Mantel–Debreu (SMD) results, which show that Walrasian general equilibrium models impose almost no restrictions on observable excess demand functions, so that the well-behaved aggregate excess demand functions on which macroeconomic theories of employment depend are arbitrary in general equilibrium models (Sonnenschein 1972, 1973; Mantel 1974, 1976; Debreu 1974; also cf. Kirman 1989; Rizvi 1994b). The SMD results tell us that the first of Kinkaid’s requirements for successful theory reduction, that the theory to be eliminated can be translated into a more fundamental theory via bridge laws, cannot even in principle be fulfilled. This goes to the very heart of the shortcutsubstitution strategy for reducing aggregate phenomena to microfoundations, and indeed suggests that, to the extent that macroeconomics requires microfoundations, so also microeconomics appears to require macrofoundations. Thus, in taking stock of the SMD fallout, Grandmont calls for “alternative research strategies” that would “reverse the traditional Neoclassical research programme, and … obtain some form of aggregate rationality … by relying more on particular features of the distribution of
38
The atomistic individual
behavioral characteristics among the members of the system under consideration” (Grandmont 1992: 33). This view of the matter, in fact, might be said to have been quietly adopted from the outset by mainstream economics’ principal proponents of microfoundations for macroeconomics, the New Classical Economists, who forego using individual economic agents in their Walrasian general equilibrium models in favor of a single “individual” agent termed the “representative agent.” It is true that the representative agent in New Classical models has a utility function and budget constraint just as do individual consumers. It is also true that the representative agent engages in maximizing behavior just as individual consumers are said to do. However, the “individual” demand curve that is derived for the representative agent specifies an aggregate demand curve for all the individuals in the economy. Moreover, the representative agent represents all the individuals in the economy on the assumption that these individuals are homogeneous and identical. Realworld individuals, who of course are heterogeneous and different from one another, are thus eliminated from the analysis in favor of a supraindividual entity. However, the central premise for the microfoundations strategy for maintaining a place for real individuals in mainstream economics is that supraindividual entities are concepts that do not count as real existents! It is thus now generally accepted that the SMD results and the failed microfoundations project of New Classical Economics brought about the virtual abandonment of the Walrasian Arrow–Debreu–McKenzie general equilibrium research program. Its demise, however, led to an explosion of interest in game theory. What is significant about this in the current context is that game theory offered an entirely new, nonpsychology-based understanding of the individual as a bearer of strategies. At the outset this was not the intended goal. When economists first turned their attention to game theory some considerable time after its initial elaboration by von Neumann and Morgenstern (1944), their initial ambition was only the better explanation of strategic interaction between firms in oligopolistic and monopolistically competitive markets (Shubik 1953; Sonnenschein 1985; Rizvi 1994a). However, treating individuals as the bearers of strategies also gave new life to traditional individualist arguments by explaining supraindividual institutions, conventions, and social rules as products of individual choices. Indeed, adopting the Nash equilibrium concept over the rival von Neumann game solution concept was particularly appealing, because it relied upon the idea of the individual as a pure isolated strategic thinker in the tradition of Walrasian theory (Mirowski 2002: 346ff.).
2.5
Game theory to the rescue? 1
To see how the conception of the individual in game theory relates to the abstract individual conception, in this section I ask three questions: How, then, does game theory relate to the human individual? Does game theory give new life to the abstract individual conception and the mainstream reductionist project? And, with its
The atomistic individual
39
emphasis on interaction, might game theory actually represent the beginnings of a departure from atomistic individualism? 2.5.1
Game theory and the human individual
Strategies, as plans of action in different contexts, are associated with players of games rather than with characteristics of them as particular types of being. To treat individuals as “bearers” of strategies is to claim that strategies can be assigned to them. This is different from first specifying some type of agent, say, in terms of rationality properties, and then eliciting that agent’s strategies on the basis of its being the type of agent it is. Rather, being a “bearer” of something, like the relation of ownership, implies nonexclusive association. Thus, strategic action cannot be uniquely characteristic of human individuals as Locke believed subjective inwardness to be uniquely characteristic of human individuals. Indeed, that we can attribute strategic behavior to a variety of different types of agent—firms, human beings, computers, and animals—implies that strategic behavior does not tell us anything specifically about human individuals. Strategies, in fact, simply constitute a set of possible moves that any kind of player could make in light of moves that competing players could make. Players, moreover, are exhaustively described in terms of their strategies, and there is nothing to the concept of a player over and above a description of possible strategies. A game of strategies, finally, is only a particular kind of formalism that emphasizes systematic interdependence. Explaining a game does not require an explanation of why a player might pursue one type of strategy or another. Rather, players are assigned strategies in a completely conjectural manner. Game theorists do not say, because players have such-and-such type strategies, games will have certain outcomes. They ask, should players have such-and-such types of strategy, what outcomes might games have? 2.5.2
Game theory and reductionism
Could game theory, however, be used to say things about human individuals? Does it offer an alternative way of carrying out the stalled mainstream reductionist program in connection with the interpretation of human institutions, conventions, and social rules as the product of games? A barrier to this argument is that game theory, like standard equilibrium analysis, involves a hybrid atomist–holist rather than purely individualist type of explanation, since it relies on holist common knowledge of rationality (CKR) and “consistent alignment of beliefs” (CAB) assumptions (Hargreaves Heap and Varoufakis 1995: 23ff.). A proposition is common knowledge if each player has reason to believe it, has reason to believe that the other player has reason to believe it, and so on (Lewis 1969: 56). The CAB assumption expands on CKR— their combination often being termed the Harsanyi doctrine (Aumann 1976)—to tell us that two rational individuals with the same information (or common priors
40
The atomistic individual
over event probability distributions) will come up with the same conclusions. Can the Harsanyi doctrine, then, be explained in individualist terms? It is difficult to imagine how the idea of shared rationality inherent in CKR can be understood if not in terms of some sort of social foundation. Even more difficult to imagine without some account of social foundations is the CAB idea that individual beliefs possess a reasonable degree of alignment. Game theory is also a hybrid atomist–holist type of explanation in the way it addresses the problem of multiple equilibria. An early view of how games might settle on certain equilibria rather than others was that certain equilibria stand out by virtue of features whose salience is common knowledge between the players (Schelling 1960). These features act as focal points upon which players converge, and in the absence of communication between players the question becomes: how do players recognize salience? If one were to argue, as did Schelling, that players try to imagine their partners’ line of reasoning, then it seems clear that we must assume there exists a social and cultural framework that permits this to occur. Pre-existing norms regarding shared expectations would then underlie individualistic interaction in games. It is true that there are formal solutions to the problem of multiple equilibria that do not explicitly invoke holist considerations, for example, recursive algorithms for unique equilibria in noncooperative games (Harsanyi and Selten 1988), but it is not clear why the players of games should have recourse to one set of algorithms over another. Indeed, it would seem that they should rely on one set of algorithms only because they believe other players do, thus returning us to Schelling’s line of thinking regarding focal points. Thus, reductionism apparently has no more promise in connection with the game theory than it does in connection with the traditional microfoundations project. 2.5.3
Hollis’ converse proposal
Might game theory, then, actually involve the beginnings of a departure from atomistic individualist arguments in mainstream economics? If the history of neoclassical and mainstream economics involves a turning away from human individuals and a series of unsuccessful attempts to reduce multi- and supraindividual entities to individuals of some kind, then perhaps the proper conclusion is that individualism is no longer the mainstay of orthodox economics. Hollis takes a position much like this in his last book, Trust within Reason (Hollis 1998). The focus of the book is game theory and a version of the extensive-form Centipede Game in which backward induction leads rational players to the least desirable outcome.2 Hollis argues that, because rational choice and the pursuit of self-interest are often costly, individuals have reason to adopt social arrangements that involve cooperation. There are a number of different forms available. Beyond simple self-interest comes sympathy à la Collard (1978), followed by principles of impartiality in a Rawlsian sense, then by a Kantian sense of duty toward others, then by Wittgensteinian “forms of life” rule-
The atomistic individual
41
following behavior, and finally by social coordination through team thinking. Hollis believes that the moral of the Centipede Game is that learning how to achieve coordination in games teaches individuals to trust one another, and that the development of trust behavior requires that individuals give up thinking of reason in strictly instrumentally rational choice terms. Might this be the verdict of mainstream economists in general? One way of pursuing this proposal is through consideration of the concept of an indefinitely repeated game. Backward induction occurs in the Centipede Game, because players know that there is ultimately an end to the game, and thus that it pays to break agreements made during the course of play. Indefinitely repeated games do not have a last round of play, and consequently players have reason not to defect from possible cooperative agreements. Therefore, some early applications of indefinitely repeated games were represented as illustrations of how cooperative behavior could emerge from self-interested behavior, for example, as when interdependent oligopolistic firms seeking greater profits behave collusively (cf. Fudenberg and Tirole 1989). Unfortunately, indefinitely repeated games by their very nature possess relatively little structure, so that ultimately “anything goes” (Rizvi 1994a: 20ff.; Hargreaves Heap and Varoufakis 1995: 170ff.). In fact, this result is so well established that it is simply taken as popular wisdom, going by the name “the folk theorem,” that is, that “any outcome that one might imagine as sensible can turn out to be a Nash equilibrium” (Fisher 1989: 116). An alternative way of entertaining Hollis’ proposal involves the concept of evolutionarily stable strategies (ESS) developed by the biologist Maynard Smith (1982) and used influentially by Sugden (1986) to produce a game-theoretic explanation of the emergence of conventions. According to Sugden, in symmetrical games individuals play mixed strategies in a consistent manner until, through trial and error, they learn the proportion between strategies that maximizes their expected utility.3 The resulting Nash equilibrium is said to be stable in an evolutionary sense, and the players are said to play evolutionarily stable strategies. However, when more than one stable equilibrium is possible, were one set of individuals to begin playing asymmetrically (play one way in one set of circumstances and another way in other circumstances), it would then be in the interest of all players to adjust to this asymmetrical strategy of play. This establishes a convention. Conventions, therefore, are shared rules or norms regulating social life that arise from games having more than one possible ESS. They can be shown to be self-enforcing, and might accordingly be seen to be the basis of a variety of “social” institutions such as private property and a system of enforceable contracts. However, contra Hollis, nothing in this sort of explanation involves individuals acting in other than an instrumentally rational fashion. Nor do individuals cease to be atomistic beings in any obvious sense. Indeed, Sugden’s goal in developing the idea of an ESS is to show that atomistic behavior can explain cooperative behavior, and that one need not imagine that individuals must exercise a noninstrumental rationality in
42
The atomistic individual
order to escape a Hobbsean world. For Sugden, in fact, the moral of the Centipede Game is that instrumental rationality places significant boundaries on what sorts of cooperative arrangement are possible—not that the potential advantages of cooperation lead individuals to employ noninstrumental forms of reasoning (Sugden, 1999). In this respect, his views are probably representative of those of many mainstream economists for whom individualism remains fundamental to economics. Thus, while most would probably allow that the CKR and CAB assumptions rule out full-scale reductionist arguments of the order sought previously, few would take this to imply that the atomistic conception of the individual itself needs to be questioned. Indeed, that game theory now offers a variety of different accounts of institutions, conventions, and social rules that are individualist in spirit no doubt demonstrates to most mainstream economists that the atomistic individual remains a fundamental part of mainstream economics. This conclusion is reinforced by my arguments above in Section 2.3 which show how atomism can be successfully elaborated in connection with the abstract individual conception. Thus, it appears that mainstream economics has restored Schumpeter’s original definition of methodological individualism in a nonsubstantive, formal manner as a form of explanation deemed especially “informative” in economics. Though some may continue to believe that the stronger, prescriptivist meaning of methodological individualism with its associated reductionist project is still defensible, the weight of theoretical and methodological investigation in recent decades argues against this as a realistic program of research. Thus, mainstream economics remains individualist, though on fundamentally different terms than was neoclassical economics.
2.6
What kind of individual?
Locke’s conception of the individual, problematic though it is, leaves no doubt in our minds that we are dealing with human individuals. Not only does Locke define individuality as subjective inwardness and contrast it with a non-individual-occupied, law-governed world of nature, but he goes on to identify subjective inwardness with human psychology. The evolution of neoclassicism into mainstream economics, however, eliminated this role for human psychology and thereby inadvertently eliminated Locke’s basis for talking about human beings as individuals. Atomism can still be refounded on a formal basis, as discussed in Section 2.3, but this new basis for individuating agents has no essential connection to human individuals. This leaves us with two problems. We might worry about how human individuals are now to be defined and explained, and we might also worry about whether mainstream economics can be about human individuals. I postpone the first concern here to concentrate on the latter. One thing that we can say about our second concern is that mainstream economics has certainly been applied to human individuals. Properly specifying objective functions to apply the theory requires that one correctly identify real human characteristics, as, for example, in ruling out negative quantities.
The atomistic individual
43
However, specifications and restrictions of this sort must be made on an ad hoc basis, since there is no longer a theory of the individual in mainstream economics on which formal representation of the individual can be systematically grounded. Thus, the identification of characteristics of real-world individuals to be incorporated in objective functions is carried out in a casually empirical and often opportunistic manner. It may be inescapable that negative quantities must be disallowed, but it is not obvious that transitivity of choice is required. The latter, needed for formal representation, apparently trumps empirical considerations. This kind of practice, of course, is what psychologists and behavioral economists, who have made a careful empirical investigation of economic choice behavior, find so disturbing about rational choice theory. Mainstream economists, then, lack rigorous standards for application of the abstract individual conception to real-world individuals. This, in fact, is precisely what one would expect, since the same formal apparatus that implements the abstract individual conception can be applied equally well and with virtually no modifications to both many-person “individuals” and to single-person individuals. This tells us that there is a more serious problem here than that application of the theory is always ad hoc. The issue is not just that mainstream economists lack an understanding of real-world individuals to guide application. The abstract individual conception itself cannot differentiate between many-person and single-person types of individual. This clearly creates a paradox, as reflected in the oxymoronic nature of the term “many-person individual.” Why is such an entity—for example, a firm or a household—one individual rather than many? One motivation for the reductionist microfoundations project is to be able to dispel this problem. Most mainstream economists, however, simply ignore the issue and treat firms and households as if they were single individuals. Ironically, the single-person individual on which these many-person individuals are modeled is almost never discussed in economics except in the classroom in the most artificial “instructional” models! If one cannot systematically differentiate many-person individuals and single-person individuals, then one lacks—in the language introduced in the last chapter—individuation criteria. That is, one lacks one of the two key elements involved in being able to talk about what it means to be an individual, and one simply does not know the difference between one person and two. Thus, the question the abstract individual conception generates is, does it constitutes a conception of the individual at all? While it may be the case that atomism—the idea that individuals (whatever they are) are relatively autonomous from one another and supraindividual entities—can be defended on its own terms, this does not imply that the conception of the abstract individual fills the bill as an atomist conception. Reason to suspect that it cannot comes out of the history by which the neoclassical conception of the individual evolved into the mainstream conception. Since the neoclassical conception had origins in Locke’s view of individuals as unique in nature, elimination of that Lockean view removed individuals from a world that could only be characterized as nature.
44
The atomistic individual
In the following two chapters, I analyze whether the subjective individual can indeed be thought of as an individual by elaborating and applying identity criteria for individual economic agents to this conception. My hypothesis is that the ambiguity, ambivalence, and agnosticism in neoclassical economics over the subjective nature of individual is manifest in its inability to fulfill my two basic identity conditions for economic agents introduced in the first chapter. Like Shakespeare’s famous prince, then, neoclassical economics cannot decide whether individuals are to be or not to be. After these two chapters, I return in Chapter 5 to the abstract individual conception to offer an account of what sort of understanding of the individual first emerged post-neoclassicism.
3 Reidentification
Preferences and human capital
For it is by the consciousness it has of its present thoughts and actions, that it is a self to itself now, and so will be the same self, as far as the same consciousness can extend to actions past and to come. (Locke 1975 [1694]: 2.27.10) [S]ince it is not substance, but consciousness alone, which constitutes personality; which consciousness, being successive, cannot be the same in any two moments, nor consequently the personality constituted by it. (Butler 1896 [1736]: vol. 1, 392) [T]he economic approach is a comprehensive one that is applicable to all human behavior, be it behavior involving money prices or imputed shadow prices, repeated or infrequent decisions, emotional or mechanical ends, rich or poor persons, men or women, adults or children, brilliant or stupid persons, patients or therapists, businessmen or politicians, teachers or students. (Becker 1976b: 8)
The last chapter concluded that the human individual drops out of orthodox economics with the emergence of the abstract individual conception of mainstream economics which does not and cannot distinguish between many-person and singleperson types of “individuals.” This outcome, I argued, came about in an unintended way when a succession of efforts in the twentieth century to escape the problematic foundations of Locke’s subjectivist conception of the individual, adopted by neoclassical theory, left economics without any understanding of the individual. Having hitched its cart to subjectivism, once it abandoned that, neoclassicism left economics with a void where the individual had been. Of course, those who sought to minimize the subjectivist elements in the neoclassical conception believed that the main problem with utility theory, either in cardinal or ordinal forms, was that it was unscientific and lacking in rigor. However, this view, I believe, was largely beside the point, since there were deeper, unappreciated difficulties in the neoclassical conception that were ontological in nature, and that implied at bottom that the neoclassical view of the individual was indefensible in both its basic and more sophisticated
46
Reidentification: preferences and human capital
forms. To make this argument, I examine two versions of the later ordinalist neoclassical conception of the individual, taken here as representing the fullest development of the subjectivist view of the individual in economics prior to its wholesale abandonment (at least in new research and theory) in favor of the abstract conception of the individual of contemporary mainstream economics. The two versions are: (a) the standard “pure” preferences view and (b) the development and modification of that view in the Chicago School time allocation model which supplements preferences with embodied human capital to refashion the explanation of human subjectivity. My argument in this chapter is that the failure of these later neoclassical attempts to explain individuals subjectively signals the end-of-the-road for subjectivism in economics.1 This chapter thus initiates discussion of the ontological topic of individual identity in economics and the two problems of individuating and reidentifying economic agents. In this and the next chapter my thrust is primarily critical in that I seek to show that the preferences-based account of the individual economic agent does not satisfy the two identity criteria for economic agents introduced in Chapter 1. The objective of this chapter is to demonstrate that this account does not fulfill the second of these two criteria, namely, that individual economic agents can be reidentified through change. The objective of the next chapter is to show that this account does not fulfill the first of these two criteria, namely, that economic agents can be individuated or distinguished from one another. As noted in Chapter 1, I consider the reidentification requirement first, because neoclassical economics developed around the assumption that the atomistic individual is a distinguishable human being. It accordingly seems reasonable to begin by granting the theory the benefit of the doubt on this score, and first determine whether the individual as conventionally understood is reidentifiable through change. Concluding here that this is not the case, I turn to the problem of individuation in Chapter 4. In Section 3.1, I first identify the different possible ways of representing the individual in the standard pure preferences view of the individual, and conclude that preferences are the only available basis on which they might be reidentified through change. Section 3.2 introduces Locke’s explanation of personal identity to show that reidentifying individuals in the neoclassical pure preferences approach to the individual operates in much the same way as it does for Locke. Section 3.3 reviews the arguments of Locke’s main eighteenth century critics on the subject of personal identity, namely, Bishop Butler and Hume. By most accounts, these arguments are regarded as having been completely successful against Locke. Section 3.4 then applies the Butler–Hume critique to the pure preferences view of the individual to show that this critique is equally successful against the standard subjectivist view of the individual in ordinalist neoclassical economics. In Section 3.5, I turn to the second version of the preferencesbased account of the individual in the form of the Chicago School time allocation model. The new picture that emerges here is that the individual is a hybrid of preferences and embodied human capital endowments. In some respects, this modification and development improve upon the pure preferences account. However, Section 3.6
Reidentification: preferences and human capital
47
looks at two new problems arising in connection with social influences on ability and education in human capital accumulation that undermine any attempt at reidentifying the subjectivist individual understood in this new way. Finally, in light of the failure of both the pure preferences view and its time allocation model extension to provide an account of individuals as reidentifiable through change, Section 3.7 asks whether neoclassical atomistic individuals can even be distinguished and individuated from one another. This introduces the topic of the following chapter.
3.1
The standard pure preferences view of the individual
In order to ask whether individual economic agents are reidentifiable through change, we must first ask what it is that is conventionally thought to make them distinct and independent, and then see whether individuals thus understood can be tracked through change. In what different ways, then, can individuals be represented as distinct and independent in the pure preferences account of the individual? Basically, there are three possible ways, though we will see that only one offers a means of explaining how individuals might be reidentified through change. First, of course, individuals can be represented as distinct and independent by virtue of having only their own preferences. It may sound odd or redundant to say that individuals can only have their “own” preferences, or that individuals cannot have others’ preferences. However, this proposition has been contested, most notably by Veblen (1899) and more recently by Packard (1957) and Galbraith (1969), who all argued that taste is socially determined, and that preferences are in an important sense shared (cf. Hodgson 2003). That individuals can only have their own preferences, then, immediately tells us that preferences are one basis on which neoclassical individuals can be distinguished from one another, since preferences in this regard are defined, as it were, as bearer-identified. Given this assumption, the development of choice theory from early neoclassicism to its twentieth century axiomatic reformulation has involved increasingly refined treatment of the properties that “own” preferences must have. Thus, in a nonprobabilistic world, individuals have preferences over different consumption bundles containing different combinations of goods and services, and, should these preferences satisfy the properties of reflexivity, completeness, and transitivity, then individuals may be represented as possessing single preference orderings. Should preferences also satisfy the property of continuity, these preference orderings may then be represented by single unique utility functions (up to a positive monotonic transformation). In a probabilistic world, if reflexivity, completeness, and transitivity still apply, a second sense of continuity applies, strong independence applies, and the standard rules of probability apply, individuals may again be represented by a single unique utility function (up to a linear transformation). One representation of the individual in mainstream theory as a separate and independent being, then, is in terms of having “well-defined preferences” that can be summarized by a single unique utility function. That individuals’
48
Reidentification: preferences and human capital
various different preferences are their “own” preferences, that is, is something that can only be shown if these preferences can be demonstrated to be organized in such a way as to produce a single (utility) function for that individual. Second, individuals can also be represented as separate and independent in terms of their endowments, i.e., the resources in their possession. Seen specifically as scarce resources valuable in consumption and production, endowments determine what individuals can acquire and offer in markets, and therefore what they can become if they are identified with their possessions. Seen specifically as private property, endowments give individuals separate and independent legal identities in the form of exclusive rights to themselves and the things that they own by law. Since Coase’s transaction cost analysis (1937), however, the scarce resource conception of endowments has become more fundamental than this private property conception. Coase showed that, under appropriate conditions, irrespective of the distribution of property rights, differentially endowed individuals can reach agreements that are efficient in redistributing access to goods and resources between them. By this token, then, individuals are separate and distinct from one another according to the opportunities their endowments create and the powers their endowments permit them to wield in market transactions. Third, economic agents can also be represented as separate and independent in terms of the information they possess. Originally, in connection with the model of perfect competition, information did not distinguish individuals. Individuals were said to know with certainty all the conceivable, potential consequences of their actions, or possible states of nature, how every possible state of nature would affect their utility given each possible action they might choose, and the probability distributions or sets of likelihood weights attached to each of the possible states of nature. After Allais (1953) and Simon (1957), economists began to investigate the consequences of placing various limitations on this conception, but these limitations applied in the same way to all individuals and had no significance for the question of what distinguished individuals from one another. However, with Akerlof (1970) and others’ subsequent emphasis on asymmetric information, information became another type of endowment differentially possessed by individuals, and individuals could then also be distinguished according to what information they possessed. Some might argue that we should also consider a capacity to make choices a fourth possible representation of individuals as separate and independent. Indeed, many regard choice as the defining characteristic of orthodox economics; for example, “Economics is all about how people make choices. Sociology is all about why they don’t have any choices to make” (Duesenberry 1960: 233). In spite of this, it is difficult to say that choice in the ordinary sense of the term is what is really involved in choice theory. Sen emphasizes that choice requires not only that one be able to consider options other than those one ultimately selects, but that one also be able to select options that are not the most preferred if we are indeed to say that real choice is involved (Sen 1993). Or, as Lawson puts it, “if real choice means anything it is that any individual could always have acted otherwise” (Lawson 1997: 9). Real choice,
Reidentification: preferences and human capital
49
that is, requires that individuals have “counterfactual” opportunities. However, in the standard analysis of choice, individuals can never choose anything other than their most preferred option as determined by their preferences and feasible set. Preferences, then, determine choice. As Reder sums up the matter in his review of the Chicago School, “‘society’s’ freedom of choice is seen as illusory. Freedom appears to consist not in power of choice, but (pace Hegel) in recognition of necessity” (Reder 1982: 35). But if choice is illusory, then it cannot be a means of representing individuals as separate and independent. There are only three ways, then, of representing individuals as separate and independent in the pure preferences account of the individual. However, neither individuals’ endowments nor the information they possess can serve as a means of reidentifying individuals. Endowments and the information that individuals possess change as a result of the choices they make, that is, they are endogenous to choice. Were an individual to be distinguished at one point in time according to possession of certain endowments or information sets, a change in those endowments or information sets would prevent our identifying this individual as the same individual at a later point in time. Thus, individuals cannot be reidentified in either of these ways. In contrast, preferences offer a clear basis for reidentifying individuals through change. When individuals’ preferences are well ordered, they can be represented by a single unique utility function. According to the standard view, however, neither individuals’ utility functions nor their preferences are changed by choices they make. That is, they are exogenous to choice. Thus, once an individual is distinguished at one point in time in terms of a particular set of preferences and accompanying utility function, having those same preferences and utility function at a later point in time should in principle distinguish that selfsame individual. Accordingly, the basis for reidentifying individuals through change in the neoclassical pure preferences view of the individual is that individuals have unchanging preferences.
3.2 Reidentification à la Locke and in the pure preferences account Seeing the neoclassical individual as distinct and reidentifiable by virtue of having unchanging preferences is the latest chapter in the long history of subjectivist thinking about individuals and personal identity which traces back to Locke. What characterizes the entire tradition from Locke to the late twentieth century neoclassical conception of the individual in economics is the view that internal mental states not only distinguish individuals from one another but also are the basis on which individuals can be reidentified through change. Though many things about individuals change over time, the one thing that does not change about them is their having a single inner mental life all their own. Locke’s particular version of this personal identity argument combined the assumption that we can identify ourselves with consciousness with the claim that we remember ourselves in our memories of the past.2 His most famous statement of this view is as follows: “For it is by the
50
Reidentification: preferences and human capital
consciousness it has of its present thoughts and actions, that it is a self to itself now, and so will be the same self, as far as the same consciousness can extend to actions past and to come” (Locke 1975 [1694]: 2.27.10). In effect, the self is immediately revealed in consciousness, whether in memory or current experience. Locke’s argument, then, can be reconstructed in essentially the following way: 1 2 3 4
It is by our consciousness that we are selves to ourselves. We remember or are conscious of ourselves being in our memories of the past. The consciousness we have of ourselves in the present in our current experiences is the same consciousness we have of ourselves in the past. Therefore, we are the same individuals in the present as we were in the past.
The version of this argument in neoclassical economics is similar, but also simpler by virtue of an amendment it makes to Locke’s view of consciousness. Whereas Locke distinguished consciousness of the past from consciousness of the present, in the pure preferences account individuals simply have timeless preferences over possible objects. This eliminates the need for making claims about the nature of memory—as we will see, a weak point in Locke’s argument—and makes the pure preferences reidentification argument quite straightforward. This argument can be set forth as follows: 1 2 3
We can be represented in terms of our preferences. Our preferences and their associated utility functions are exogenous, and are accordingly free from the effects of change. Therefore, we are the same people irrespective of when or on what objects we exercise our preferences.
Stigler and Becker (1977) popularized the second step here which simplified Locke’s view and made recourse to claims about memory unnecessary. If preferences never change, then we need not be concerned about whether our preferences in the past are the same as those we have in the present. With unchanging preferences, we are conscious of what we prefer at all times in precisely the same way. Thus, we are always distinct and unchanging individuals through whatever change we undergo in other respects (for example, in terms of the things in our possession). Preferences, then, constitute the basis on which individual identity can be explained in the pure preferences version of ordinalist neoclassical economics.
3.3
Locke’s critics
However, Locke’s understanding of personal identity was shown to be problematic almost as soon as it came to wider attention, first by Bishop Butler who charged that Locke’s view was circular, and then even more devastatingly by Hume who argued that Locke had presupposed a metaphysical concept of substance to make the argu-
Reidentification: preferences and human capital
51
ment work. Today, philosophers not only regard Locke’s argument as unsatisfactory but also have serious doubts about whether consciousness or individuals’ mental states can even in principle explain personal identity (cf. Noonan 1993: xi-xx). Thus, just as Locke’s subjectivism provides problematic foundations for the neoclassical theory of the individual, so his specific account of personal identity offers a questionable basis for understanding reidentification and the identity of the individual in neoclassical economics. 3.3.1
Butler
Locke had argued that the self we remember as our own in our past experience is the same self we find before us in our present experience. This sameness of self, exhibited through our having a single consciousness, was the basis on which he argued for our each having a unique individual spiritual substance—something Locke believed needed to be demonstrated for both religious and political reasons. The two key pieces in this argument—(a) that we recall ourselves as having had our past memories, and (b) that having a unitary consciousness justifies our supposing we each possess a unique individual spiritual substance—were the targets respectively of Butler and Hume. Butler (1896 [1736]) detected the obvious flaw in Locke’s claim that we remember ourselves as having had our past memories when he asked how Locke knew it was he himself who had had his past memories. The point of this peculiar question is not that one might remember someone else as having had one’s own memories. Rather, Butler was simply pointing out that Locke presupposed it was he himself who had had his past memories. But this made Locke’s explanation of personal identity in terms of memory questionable or circular. If we see ourselves as having our past memories, we cannot then argue that a sameness of consciousness across the past and the present demonstrates that we are the same persons today that we were in the past. Butler’s critique has stood the test of time, and Locke’s argument is generally regarded as circular. However, there is another possible interpretation of the argument that is worth considering. Rather than say Locke was attempting to demonstrate the existence of a single self, his goal might simply have been to exhibit the criterion by which we recognize it, since he was already convinced for religious reasons that we each did in fact possess a soul which could not itself serve as a criterion of identity on account of its immateriality.3 3.3.2
Hume
It was this latter conviction, then, that Hume attacked, in the process creating the fundamental challenge to all subsequent theories of personal identity (1888 [1739]). As a thorough-going skeptic and empiricist, Hume saw no need for a concept of self or a concept of individual spiritual substance which he labeled metaphysical in the worst sense of the term. He thus aimed his skeptical attack on Locke directly at the
52
Reidentification: preferences and human capital
idea that there could even be such a thing as a self or individual spiritual substance that somehow inhered in each of us. Locke had supposed that each person could detect his or her own spiritual substance by a sort of inner self-inspection of current or past experience, and used this assumption to pull off his memory argument. Hume replied that he was unable to detect any such substance in himself, and doubted that anyone else was able to do so either. A supporting and even more powerful argument comes from Hume the empiricist. If the idea of spiritual substance is nonsense, then each of us is “nothing but a bundle or collection of different perceptions” (Hume 1888 [1739]: 252). However, because these perceptions are constantly changing, they cannot constitute a single unity. This was a new version of the classical philosophical problem of trying to find the “one” in the “many.” For Hume, once the individual was defined in terms of the varied contents of consciousness, this problem was insoluble. Having rejected the concept of personal identity and the idea of a self, Hume then turned to diagnosing our attachment to the idea. If it was an obvious mistake to “suppose ourselves possest of an invariable and uninterrupted existence thro’ the whole course of our lives” (Hume 1888 [1739]: 253), why was it that so many people believed in the existence of the self? The root of the problem, he believed, lay in a conceptual confusion. When people think of the self or indeed any other sort of single thing, they encounter and conflate two ideas that need to be distinguished, namely, the idea of an object enduring through change and the idea of a succession of closely related objects. However, being mentally lazy, most people conflate the latter idea with the former, and see a collection of closely related objects as one single thing. This creates a habit of thinking that encourages overlooking gradual changes and small differences in things, and, since this habit is pervasive among people, it gradually acquires the sanction of custom. This is bad enough, but the real problem in Hume’s view arises when philosophers legitimate custom and begin to speculate about the nature of personal identity. Noticing the paradox involved in ordinary thinking, they take the easy way out and invent a metaphysical ground for the self, such as Locke’s concept of individual spiritual substance. Yet philosophical devices such as these can never justify customary ways of thinking, and the only reasonable conclusion available to us, Hume concludes, is to reject the entire concept of the self. Hume’s argument, then, undercuts Locke’s strategy for explaining personal identity by directly eliminating the basis on which individuals might be reidentified through change. Butler’s point about Locke’s argument being circular creates doubts about that argument, but still leaves open the possibility that an individual having the same consciousness over time might somehow be reidentified as a single individual in terms of consciousness. Thus, if individuals are to be thought reidentifiable by virtue of always having the same preferences, we need to ask whether Hume’s antisubstance argument has an application to the neoclassical model. I do this in the following section, and then also take a look at whether Butler’s circularity charge also has an application to this model.
Reidentification: preferences and human capital
53
3.4 The Butler–Hume critique applied to the neoclassical view On the surface, it might seem that Hume’s arguments are not really relevant to the neoclassical conception of the individual, since they have Locke’s concept of an individual spiritual substance as their primary target, and that notion seems far removed from the idea of the individual having a single utility function. However, if we put aside the unfamiliar philosophical term “substance,” Locke’s idea that we can be identified with something over and above our individual states of consciousness is basically just what is involved in saying that individuals can be identified with utility functions over and above their preferences. In both cases, the question at hand is whether a collection of things—states of consciousness for Locke and discrete preferences in neoclassical economics—can be treated as a single unity over and above that collection. To see how this problem arises in the neoclassical account, notice the difference between describing a set of preferences, even well-ordered ones that can be formally represented by a single utility function, and saying that a set of preferences belongs to a single individual. That a set of preferences is well ordered and can be formally represented by a single utility function in itself says nothing about the ownership of those preferences or utility function. All that is involved in saying that a set of preferences is well ordered and representable in terms of a single utility function is that a description can be imposed on these preferences that gives them certain properties as a set. Also, one can only say that a set of preferences belongs to an individual on the basis of the claim that each and every preference an individual has is an “own” preference. However, this assumption is precisely the one that Hume thought unsupportable. Thus, from his perspective, saying that a set of preferences belongs to a single individual is to invoke an independently existing thing. Hume, presumably, would not have objected to neoclassical economists referring to preferences as “a bundle or collection of different perceptions” (though he might have differed with them over how “bundling” was to be understood). But he would probably have seen the neoclassical claim that preferences belong to individuals as an unfounded claim about individual substances. Furthermore, if asked to speculate about why sets of preferences are regarded as belonging to individuals, Hume might well have said that economists follow ordinary opinion in confusing two different ideas: the idea of a unitary thing or object (here, the individual as the owner of a set of preferences) and the idea of a collection of closely related things (namely, those preferences customarily ascribed to individuals). This conflation, he might have added, is promoted in arguments such as Stigler and Becker’s (1977) that preferences do not change, since in that regard the small differences people tend to overlook are further minimized, and it is only a small step to supposing that an individual’s preferences may be identified with the individual to whom they belong. Indeed, from Hume’s perspective it might be said that Stigler and Becker’s paper, devoted as it is to the proposition that preferences do not change,
54
Reidentification: preferences and human capital
constitutes the supreme defensive posture for those attempting to infer an idea of individual substance from a collection of mental experiences. However, by itself a collection of preferences, Hume would say, cannot pick out anything. Preferences are only a form of experience, and, when we inspect our preferences, we do not detect among them any substantial self or individual. Do neoclassical economists have a reply to Hume? It might be argued that the specific understanding of preferences in neoclassical economics presupposes the existence of a self to whom preferences belong. Preferences can have many different kinds of properties, but the ones that economists assume (such as reflexivity, completeness, etc.) are those that only (rational) individuals possess. Thus, it might be replied, one does not begin with sets of preferences and then attempt to infer ownership by individuals. Rather, one begins with (rational) individuals and asks what kinds of preference must they have. For example, consider the logic behind a popular argument for why preferences must be transitive: 1 2 3
An individual with intransitive preferences would serve as a “money pump” in that other individuals could carry out transactions with this individual until the individual was depleted of all resources. However, such an individual would ultimately cease to exist. Therefore, preferences must be transitive.
Here, we conclude that preferences are transitive because we are dealing with a certain type of individual, namely a rational one. Thus, the individual must be presupposed in our characterization of transitive preferences, and, contrary to the argument above, we would not have Hume’s problem of trying to elicit ownership from preferences. While there is something to this argument as a response to Hume, notice that this is just the kind of reasoning to which Butler objected as circular. Locke wanted to show that a single self exists, but, in supposing his past memories were his own, he ended up assuming what he wanted to explain. This same problem arises in different form in the response above. If the answer to Hume is that neoclassical economics only considers the kinds of preference that must belong to (rational) individuals, then the problem is solved by simply assuming what is to be proven. That is, restricting attention to a particular kind of preference does not explain the relation between preferences and individuals; it simply classifies preferences as individual preferences. Indeed, that this does not explain that relationship is evident from the fact that transitive preferences can be attributed to all kinds of subjects including individuals who are single persons and “individuals” made up of many people. Thus, Hume’s problem remains. For reasons entirely independent of the Butler–Hume critique, however, the neoclassical preferences-based conception of the individual changed significantly as a result of the development of the Chicago School time allocation model of decision
Reidentification: preferences and human capital
55
making and the attendant elaboration of human capital theory. Human capital theory has had a tremendous impact on contemporary society as reflected in the number of ways that it has influenced the promotion of education and training. At a deeper level it has also changed the way that individuals think of themselves, since it has encouraged them to think of themselves as products of human capital investments. However, my principal interest in this modification of the pure preferences account of the individual concerns its standing as the last major attempt to make the preferences-based view of the individual work as a theory of the individual. I thus take the time allocation model to be the last significant substantive conception of the individual in the transition from neoclassical economics to mainstream economics. How, then, did the time allocation model and human capital theory change the neoclassical concept of the individual and the basis on which individuals are distinguished from one another? Once we determine this, we can then ask whether this new view provides an alternative way of reidentifying individuals through change.
3.5
The time allocation conception of the individual
The time allocation model underlies human capital theory, because investments in human capital require that individuals allocate resources at one point in time in order to bring about certain outcomes at another point in time. The model thus implicitly requires that we see individuals as sustaining a single identity through time, since individuals would only carry out such investments if they expected to be the ones who later received the benefits from those investments. Hume’s critique has at its core the problem of how something that changes can still be called one thing. To say, then, that individuals are reidentifiable as the selfsame individuals through change, human capital theory must consequently be able to say that, at some level, gradual changes and small differences in individuals as they invest in themselves do not change these individuals. Failing this, we face Hume’s charge that some idea of substance has to be invoked as a substitute for demonstrating sameness through change. As we will see, human capital theory does indeed offer a basis for saying that there is sameness in the individual through change in a human capital investment process. This basis lies in a changed conception of how the individual is to be represented from being purely a set of preferences to being a set of preferences plus embodied human capital. What primarily distinguishes the time allocation model from the earlier neoclassical thinking about individual decision making is the idea that individuals are simultaneously involved in both production and consumption activities (Becker 1965; Lancaster 1966). One thing that this implies is that individual demand behavior has to be explained both in terms of the characteristics or qualities of market commodities and in terms of the goods that consumers themselves produce using market commodities as inputs. Individuals regard the commodities available to them in the market as intermediate rather than final goods, and purchase them according to how
56
Reidentification: preferences and human capital
their characteristics can be put to use in producing their desired final goods. One demands market commodities because one has a demand for final, self-produced goods. Of course, not all final, self-produced goods are the result of human capital investments, since individuals demand market commodities to produce household capital as well. However, many of the most important final goods that individuals produce are the result of human capital investments. For example, taking health as an object of final consumption and health care as a market commodity, the demand for health care is explained as a derived demand for health when consumers combine health care with non-market inputs such as their time, labor, and the skills they possess to produce better health for themselves (Grossman 1972). How, then, does theory account for the identity of the individual through time? The answer lies in how the pure preferences conception of the individual is expanded into a conception of the individual as a combination of preferences and self-embodied human capital stocks. The theory explains how individuals use a budget of time and market inputs at one point in time essentially to reconfigure their utility functions. Consider, for example, Stigler and Becker’s paradigmatic music appreciation case, in which utility depends upon two produced final goods: U = U(M, Z)
(1)
where M and Z respectively are “music appreciation” and “other goods” produced and consumed (Stigler and Becker 1977). Music appreciation itself is an output of the following production function: M = M(t, S)
(2)
where t is the time spent listening to music and S is the amount of “music capital” one already possesses. Putting this in period analysis terms, the amount of music capital one possesses at time j is generated in a dated “learning-by-doing” process via accumulation of the effects of earlier music appreciation on later stocks of music capital: Sj = h(Mj–1, Ej)
(3)
where Ej measures the effect of education in music on one’s music appreciation skill. Music capital Sj thus increases with earlier consumption and production of music appreciation at Mj–1. The new stocks of S are combined with time t in equation (2) to produce higher values of M and thus higher levels of utility in terms of equation (1). Music listening accordingly has positive consumption and production effects over time, as one’s enjoyment of music is heightened through accumulation of a “consumption capital” in music appreciation occurring through time.4 Notice, then, how the introduction of time and investment in human capital into the model of the individual suggests a way of possibly circumventing the application
Reidentification: preferences and human capital
57
of Hume’s critique to the pure preferences version of the individual by providing an alternative representation of the individual. The main thrust of Hume’s critique rests on the idea that gradual changes and small differences in an individual’s preferences are incompatible with treating the individual as a unity of contents of consciousness. The time allocation model, however, effectively does away with these gradual changes and small differences by first assuming constant tastes, and then tracking gradual changes and small differences in the individual in terms of accumulations in human capital where, except early in the process, the capital accumulated is large relative to the additions to it. Thus, though there are indeed gradual changes and small differences in the individual, the combination of constant tastes and a sizable stock of relatively permanent human capital embodied in the individual provides a plausible basis for speaking of continuity in the individual. However, the model also offers a response to a second theme in Hume’s critique. In applying Hume’s critique of Locke’s concept of individual substance to the preferences-based view of the individual above, we saw the underlying problem that there was no essential connection between a set of preferences and the particular individual to whom those preferences belonged. This was the “ownership” problem. For the human capital model of the individual, then, the parallel question would be whether the set of changes in the individual associated with additions to human capital have any stronger connection to the individual. Note that human capital investments are initiated by and carried out by the individual to whom they accrue, and that additions to an individual’s human capital stock are endogenous to the individual’s decision to invest in human capital. Sets of preferences do not possess any intrinsic connection to particular individuals, because in themselves they do not display any “belonging” or signs of ownership. A set of human capital additions, however, seems to possess precisely this characteristic, since they exist only because they have been brought about by a particular individual. In broader terms, what the human capital model does is introduce an unarticulated and undefined principle of action in the idea of carrying out a campaign of human capital investments. The pure preferences view, in contrast, treats the individual as passively responding to changing prices given a structure of tastes. Thus, human capital theory’s implicit principle of action is arguably attended by an equally implicit idea of there being an agent guiding a series of investments. Hume criticized Locke’s idea of an individual substance not only because it had no essential connection to states of consciousness but also because the idea of something subsistent was by itself at odds with his thorough-going empiricism. In the idea of the individual originating human capital investments we certainly do have the idea of something subsistent, and the idea of an agent of this kind, even if implicit, was no more to Hume’s liking than was the concept of causality. However, today most prefer to reject Hume’s empiricism to preserve cause-and-effect accounts of the world. Thus, agency and a capacity to originate events cannot be easily dismissed, and, while Locke’s idea of individual substance remains objectionable, some other
58
Reidentification: preferences and human capital
idea of the individual as subsisting through change might yet be reasonably pursued. The time allocation model provides a way of thinking of the individual in these terms in its characterization of individuals as agents capable of originating sequences of events. Let us pull together, then, how the time allocation model strengthens the conception of the individual in neoclassical economics from an identity perspective. Of the three different ways in which individuals can be represented in the pure preferences approach, an individual’s endowments and information were ruled out as a basis for reidentifying individuals through change, because these could change with choice. In the time allocation model, however, endowments are not simply possessed and therefore separable from the individual, but are embodied in the individual and therefore inseparable from the individual (putting aside depreciation). Moreover, in the human capital framework, the distinction between information and endowments breaks down. We might thus say that the time allocation model shifts the representation of the individual from simply a set of preferences to a hybrid structure of preferences and embodied informational endowments. It is specifically a hybrid structure because individuals’ original preferences guide their accumulation of human capital and thus remain implicit in the individual further embodied with human capital. Alternatively, it is a hybrid structure because individuals’ preferences are heightened or enhanced by acquisition of human capital stocks that enable them to enjoy higher levels of utility than when they were without those stocks. In this new representation, then, the individual evolves in an orderly manner through change. This suggests a means of reidentifying the individual through gradual changes and small differences, laying a foundation for an explanation of the identity of the individual.
3.6
New problems
Unfortunately, the richer framework of the time allocation model also creates two new problems associated with attempting to reidentify the individual as a hybrid structure of preferences and embodied human capital. They concern (a) the concept of ability that this new conception employs and (b) the role of education in heightening an individual’s preferences. Both problems are associated with the theory’s half-hearted introduction of a concept of agency. First, the model’s idea of the individual as an agent able to initiate sequences of events invites us to ask how the model’s implicit concept of ability is understood. In equation (3), the model’s predictions regarding the accumulation of stocks of music capital depend on the assumption that the partial derivative of Sj with respect to Ej is positive. Why should the partial derivative of Sj be positive? When individuals acquire a stock of capital, as in the music appreciation case, and then make use of that stock of capital, they exercise certain skills, namely, the skill to learn about music and the skill to put what one has learned to work in enjoying music. Acquiring a skill, however, depends upon having certain abilities, since not all people can acquire the
Reidentification: preferences and human capital
59
same skills or acquire them to the same degree. In the music appreciation case, for example, an individual might be tone deaf. For such an individual the partial derivative of Sj with respect to Ej might be zero. The time allocation model has little to say about how abilities might vary across individuals. That they vary is not by itself a serious problem, since Becker would presumably be happy to allow for cases such as the tone deaf individual. More troubling is the issue of what might determine a variation in abilities across individuals, since, while some variation no doubt has to do with natural differences between individuals, surely some variation in abilities comes from interactions with other individuals. However, if individuals’ abilities are socially influenced, then human capital investments heighten individuals’ preferences and reflect social factors determining the distribution of abilities. In this case, it would not be possible to reidentify an individual as a selfsame individual through a sequence of gradual changes and small differences in the human capital embodied in that individual, because these embodiments could not be strictly identified with the individual alone. The individual enlarged in terms of embodied human capital is in important respects a socialized individual. A second problem concerns the role of education in heightening an individual’s preferences. In the music appreciation case, when individuals’ human capital stocks improve their music appreciation skills, this also increases the utility they derive from listening to music. Their preferences for listening to music are thus effectively heightened relative to their preferences for other things in which no investment has been made. In terms of the model, this all depends on the Ej term in equation (3), which measures the effect of education on music appreciation skill. But how does the Ej term operate on individuals? If we suppose that education is consumed as valueneutral information, we would say that individuals expend effort and time to embody information in human capital stocks in themselves that then heighten their existing preferences. It would be naive, however, to suppose that the information people acquire in education is value neutral. In learning to appreciate music, for example, individuals must learn what others believe constitutes good music. They may well have their own opinions about what good music is, but these opinions usually emerge after one has learned what others have classified as good music. Thus, the individual’s later values of M bear the imprint of other individuals’ preferences, and the hybrid structure of preferences and embodied endowments that characterizes the educated individual picks out an individual who has also been enlarged socially. This is not to say that individual preferences are determined by society, but only that they are influenced by society. The upshot again, however, is that we cannot reidentify the individual simply as a hybrid structure of preferences and embodied endowments through change when other individuals’ values are incorporated in the individual’s heightened preferences. Of course, a response to this might be that the language of “heightened preferences” is not consistent with the formalization of the analysis, since equation (1) remains unchanged through the process of human capital investment. The model
60
Reidentification: preferences and human capital
itself does not speak of “heightened” preferences, and thus there is no formal mechanism by which social influences become incorporated in individuals’ utility functions. The problem with this reply is that it takes a formal representation of time allocation behavior as a definitive representation of that behavior—a now common and increasingly widespread error in formalist work in all of economics. Formal representation, like other types of representation, involves the theoretical isolation of certain relationships in order to facilitate a particular explanation (Mäki 1992, 1994; Runde 1998). This means that modes of representation must always be evaluated according to whether or not they help produce good explanations. However, the explanation in question here is arguably one designed to maintain the assumption that preferences do not change. Indeed, the formalization disallows the very issue in question when we consider the impact of investments in human capital on individuals. Consequently, it is hard to say that this is a “good” explanation, unless the measure of a “good” explanation is that it sustains prior assumptions meant to comport with a particular vision of the individual. Should we, however, be open to the idea that individuals are not fully described in terms of their private mental lives, and suppose that some social components go into the make-up of individuals, then we are free to reconsider what might constitute a “good” explanation of individuals’ investment in human capital. In the music appreciation example, one thing about which most people would probably agree is that individuals who invest in music appreciation enjoy music more after learning more about it. One could explain this by saying that their taste was unchanged, and that having more music appreciation human capital “leverages” those tastes in a way not possible before investment. However, one could equally explain this as a change in taste. In my view, most people would adopt the latter explanation. But it does not matter whether I am right, since the upshot of there being different interpretations of what happens in cases such as this one is that there are disputes over the nature of taste. Contrary to Stigler and Becker, de gustibus est disputandum. From this perspective, Becker’s more recent efforts in his Accounting for Tastes, which seek to introduce a variety of nonindividual factors into individual utility functions, do not really constitute a departure from his and Stigler’s original view. For example, Becker now allows that, since “a person’s personal and social capital form part of his total stock of human capital,” these additional social factors ought to be included in a new “extended utility function” (Becker 1996: 5). Social capital can include a variety of matters that reflect society’s influence on individuals, such as individuals’ social standing and reputation, and consequently we can imagine that individuals seek to produce these final goods with market commodities just as they seek to produce more ordinary objects of consumption to raise their utility. However, as in the original time allocation model, “the utility function itself is independent of time, so that it is a stable function over time of the goods consumed and also of the capital goods” (Becker 1996: 5). That is, while social influences enter into individuals’ lives as unchanging arguments in their utility functions, they do not
Reidentification: preferences and human capital
61
“heighten” those arguments. Why not? The answer rests with Becker’s preference for a modeling approach that gives determinacy to choice analysis, emphasizes equilibrium market mechanisms, etc. These are reasonable goals, but they are not necessarily the only goals one ought to consider in providing a “good” explanation of human capital investment. Moreover, they can have a particular influence on public policy debate about how to support human capital investment. Thus, if individuals could be identified and reidentified in terms of their internal mental states alone, it might make sense to argue that public support for education and human capital investment ought to focus on making resources directly available to individuals, for example through cash outlays to be used for education as individuals themselves would determine. However, the conclusions above regarding how human capital investments “heighten” preferences suggests a different emphasis. If social factors play an important role in determining both the amounts and kinds of human capital individuals acquire, then it would make better sense to channel public resources towards building up the social educational infrastructure supporting individuals’ human capital investments—a system of public education being but one example. Furthermore, since variations in human ability are influenced by society, on equity grounds changes in the distribution of ability might also be a proper focus of public policy. A variety of public programs aimed at rectifying unequal access to education could then be called for. The narrow view of ability in the time allocation model, however, supports neither of these conclusions. Of course, many already accept social policy views such as those based on empirical evidence and an understanding of the roles played by education and the distribution of ability in human capital accumulation. What the arguments here do is point to the role played by theories of the individual in economics in these debates, and how particular types of “good” explanation support certain policy conclusions.
3.7
Individuation reconsidered
While the time allocation conception of the individual departs from the traditional pure preferences conception in explaining the individual as a hybrid of preferences and embodied human capital, individuals are still ultimately seen as distinct and independent in terms of their own mental states. Preferences are effectively heightened by human capital investments, but, thus understood, preferences are still the basis on which individuals are distinguished from one another. Because of this, the time allocation model brings out fundamental weaknesses in the basic subjectivist conception of the individual. The two problems that the time allocation version encounters regarding social influences on ability and in education are only problems if the goal is to distinguish and reidentify individuals strictly in terms of inner mental states. Social influences on the distribution of ability and in education on individual preferences concern factors external to individuals’ inner mental lives. Thus, recharacterizing individuals in terms of preferences heightened by human capital
62
Reidentification: preferences and human capital
investments makes external social factors relevant to how individuals are reidentified in terms of inner mental states. Individuals cannot be reidentified merely in terms of their own preferences since it is not merely their own preferences that are involved in their make-up after human capital investments. The objective of this chapter was to determine whether individuals in the two neoclassical preferences-based accounts of the individual can be reidentified through change. To investigate this issue, I began by supposing that individuals could be at least provisionally distinguished from one another, since asking whether an individual can be reidentified through change requires that we first begin with some understanding of the individual as a distinct being, and then ask whether the individual understood in this way persists through change. Looking at the issue of reidentification before the issue of individuation also made sense because neoclassical economics is typically thought to offer an account of individuals as distinct and independent. However, that the neoclassical account does not succeed in explaining the reidentification of individual economic agents through change in either of the two versions examined here invites us to ask whether understanding individuals in terms of preferences is sufficient for distinguishing them as separate and independent in the first place. In fact, the issue of whether individuals can be distinguished from one another in terms of their preferences is closely connected to the issue of whether they can be reidentified as the same individuals after change. When reidentification fails, as when we take into account social effects on ability and in education on individual preferences in human capital investments, something other than just the individual’s own preferences goes into the subsequent make-up of the individual. What distinguishes the individual after human capital investment is not what distinguishes the individual before human capital investment, and thus the failure to reidentify the individual indicates a problem in the way the individual is distinguished in the first place. Hume, in fact, reasoned precisely in this way. Since, as an empiricist, he rejected Locke’s idea of an individual spiritual substance, he saw the question regarding whether the self existed to be a matter of whether the “bundle or collection of different perceptions” we observe when we look into ourselves can be treated as one single thing. Is there a unity to each of us as a “bundle or collection of different perceptions”? If not, then we cannot each be distinguished as separate and independent things. That is, we cannot distinguish ourselves from one another, and consequently cannot say that we are individuals. This more fundamental issue is the subject of the next chapter’s investigation of whether the neoclassical conception of the individual satisfies the second identity criterion required for a successful conception of the individual.
4 Individuation Multiple selves
Two souls abide, alas, within my breast, And each one seeks for riddance from the other. (Goethe, Faust, lines 1112–13) Strawson: How can one know that there is just one soul and not, say, twenty inhabiting one’s body? Swinburne: I cannot know for certain, but I have good reasons for believing in only one soul, because postulating just one soul connected to this body is enough to explain all behaviour of this body. It is not necessary to believe in twenty souls unless there are twenty different patterns of behaviour which can be explained in terms of twenty different systems of belief, desires, and conscious lives manifested by this body. (Peacocke and Gillett 1987: 50)
Hume’s chief intuition regarding the problem of personal identity was that one cannot explain the self as a collection of states of consciousness, because the idea of a collection of different things immediately contradicts the idea of a single thing. Once one has defined the self as a collection, the idea that that collection also possesses an essential unity can always be questioned by some sort of skeptical argument. In fact, Hume was not entirely consistent in advancing this critique of personal identity, since he began by saying that he could look into himself when surveying his states of consciousness in search of a self. Thus, he inadvertently acknowledged a unity of sorts to his own states of consciousness in the very process of trying to deny there could be one. However, this was not a serious flaw in his argument, because he could equally have said that he was surveying “states of consciousness” (without any indication of ownership), and that there was no self anywhere to be found. This latter way of putting his initial move has the advantage of bringing out more clearly how the problem of finding a unity to states of consciousness is the same as the problem of individuation. If one could survey states of consciousness potentially belonging to anyone, or if states of consciousness were not immediately “own” states, then one would have the problem of knowing where one person ended and another began, or where the boundaries were between people. Thus, finding a unity to one’s own or
64
Individuation: multiple selves
anyone else’s states of consciousness would involve the same task as trying to individuate one person from another. This chapter investigates this problem of individuation in connection with the pure preferences version of the individual in later neoclassical economics. It asks whether the individual understood as a collection of well-ordered preferences can be said to constitute a unity separate and distinct from other individuals. The basic idea that preferences are well ordered and representable in terms of a single utility function has its origins in Bentham’s assumption that all our pleasures and pains are commensurable. Bentham believed that reason “approves or disapproves of every action whatsoever, according to the tendency which it appears to have to augment or diminish the happiness of the party whose interest is in question” (Bentham 1970 [1789]: 2). However, reason is only able to do this if all an individual’s pleasures and desires are commensurable and can thus be compared with one another according to whether they tend to “augment or diminish” overall happiness (Hollis and Sugden 1993: 4–5, 29). The form of the commensurability assumption in the axiomatic interpretation is that our preferences are complete, meaning that for any two bundles the individual either prefers one to the other or is indifferent between the two. This assumption is arguably the most fundamental of all the assumptions made about preferences, since noncomparable preferences would be directly at odds not only with the idea that individuals engage in rational choice but also with the idea that the individual can be understood as a set of preferences. Before Bentham, however, Hume had rejected the commensurability assumption and claimed that this implied reason must be “a slave of the passions.” His argument was that, when our pleasures and pains pull us in different directions, in the absence of a common metric, reason is powerless to judge the competing value of these different options. Hume’s reluctance to suppose our pleasures and pains commensurable stemmed from his empiricism. Were reason able to evaluate experience, it would be invested with a power reminiscent of rationalist theories of knowledge which Hume rejected. It would also provide a unity to individual experience that would be tantamount to assuming the existence of a self. Thus, for Hume the assumption that there was no common metric for pleasures and pains was equivalent to denying that there might be a unity of consciousness that would identify the self. Given that popular opinion was wedded to the idea of personal identity, the implication of Hume’s argument was that, if one insists on using the language of the self, then the multiplicity of an individual’s different and conflicting pleasures and pains implies that each “self” is made up of many selves—a contradiction. This way of addressing the problem of individuation shifts the focus from whether the individual is a unity of states of consciousness to whether the individual is a unity of selves. It is the appropriate way to approach the problem of individuation in connection with later neoclassical economics, because the issue of multiple selves in the form of multiple utility functions has received considerable attention in the standard litera-
Individuation: multiple selves
65
ture. Not surprisingly, this multiple selves literature begins with careful scrutiny of the commensurability or completeness assumption. In Section 4.1 of this chapter I survey the main contributions to the standard literature on the subject of whether individuals have more than one utility function. The general conclusion of this literature is that, in the absence of special assumptions about the nature of preferences, intrapersonal collective choice is really no different from interpersonal collective choice, so that Arrow impossibility results and other collective choice paradoxes apply no less to the single individual than to collections of individuals. Ironically, then, methodological individualism requires that one be a methodological collectivist, at least with respect to the individual (and then why not across individuals?). Section 4.2 looks at three unsuccessful responses to this dilemma that seek to explain how various asymmetrical relationships between noncomparable subsets of preferences offer interesting opportunities for explaining the unity of the individual. These three responses are: 1 2 3
Principal–agent models. Weakness of will and self-control arguments. Metapreferences or hierarchical organizations of preference analyses.
Section 4.3 looks at one last approach to understanding individuals in terms of preferences—one in fact that bears a resemblance to human capital theory’s introduction of a concept of agency. This approach investigates nonstatic and endogenously changing preferences by asking whether individuals adopt strategies by which they might manage the evolution of their preferences. I conclude that this general approach is also unsuccessful. Finally, Section 4.4 takes stock of the preferencesbased view of the individual in terms of the problem of individuation, and points towards the investigation of Part II.
4.1
Single versus multiple utility functions
In his famous “Rational fools” paper that did much to generate initial interest in the topic of multiple selves, Sen argued that commitment (in contrast to sympathy) is essentially nonegoistic in nature, and that recognizing its relevance to economic choice drives a wedge between individual choice and welfare which most economists had assumed were indissolubly linked (Sen 1977). Harsanyi had earlier considered the possibility that individuals might have both “subjective” and “ethical” preferences, where the latter expressed what the individual prefers “only in those possibly rare moments when he forces a special impartial and impersonal attitude upon himself” (Harsanyi 1955: 315–16; cf. Fontaine 2001). Sen went further, however, and supposed that individuals possessed distinct, incommensurable or noncomparable sets of preference rankings, with some preferences so organized as to express an individual’s moral judgments and others organized to express an individual’s self-
66
Individuation: multiple selves
interest. Margolis subsequently further developed Sen’s idea explicitly in terms of the notion that the individual was a “divided self” (Margolis 1982, 1990). Individuals, he proposed, have two utility functions, each of which possesses the usual utility function characteristics, a self-interested utility function and a group-interested utility function, with the latter displaying both “goods altruism”, i.e., concern about others’ utility, and “participation altruism”, i.e., utility from “giving resources away for the benefit of others” (Margolis 1982: 21). However, neither Sen nor Margolis ventured to explain how a single individual could have different preference rankings or different utility functions. Moreover, their respective ways of looking at the matter—that individuals are not utility maximizers when fulfilling their commitments, and that individuals are self-interested and also group-interested—suggested that commitment and group interest were exceptional types of behavior, or perhaps something that mostly occurred outside the domain of the market. From the perspective of a discipline that conceived itself as being a “separate” science (Hausman 1992), and one especially that drew a hard line between positive and normative analysis, economists might accordingly simply disregard the Sen and Margolis proposals that individuals possessed competing preference rankings or multiple selves. Yet while this may begin to explain why most neoclassical economists disregarded the matter, it does not address the fact that there is no obvious reason why one should not suppose that individuals have more than one utility function, indeed conceivably an indefinite number of utility functions, and thus an indefinite number of different selves. For example, one might simply assume, as Steedman and Krause (1986) subsequently did, that individuals have a variety of different roles in life, and associate with each of these roles a separate utility function. “Individuals,” then, would have as many selves as they had roles and utility functions. The Steedman and Krause paper is important because it formally demonstrates how in the language of utility maximization one can model individuals as possessing different subsets of incommensurable or noncomparable preferences. Drawing on the ordinary intuition that individuals occupy a variety of roles in life, Steedman and Krause suggest that these roles produce different points of view from which individuals assess their choice opportunities. Preferences organized from different perspectives would then need to be represented by separate utility functions—just as if one were speaking of separate individuals—and, for purposes of comparison with single individual–single utility models, each of these utility functions could be said to possess the usual properties that utility functions possess in the standard analysis. Steedman and Krause did not claim that individuals have many sides to themselves or possess multiple utility functions. They only show that it is possible to conceive individuals in this way, thus shifting the burden of argument onto those who reject such an account. Their paper also exposed the lack of foundation to Bentham’s psychological hypothesis regarding the commensurability of our pleasures and desires. Rather than give an argument for commensurability, Bentham simply assumed it, because he had
Individuation: multiple selves
67
decided in advance that individuals could be understood as happiness maximizers. However, as Hume argued, this assumed precisely what needed to be shown. The Steedman and Krause paper did not contain any formal innovations, and its basic argument could have been advanced earlier. Why, then, has the single individual–single utility model gone unquestioned for so long? Clearly, its advantages are the predictability of single-motive behavior and the attendant formal tractability. The disadvantages are a less realistic account of motivation and the detachment of economic analysis from what ordinary discourse reports about the nature of behavior in economic life. Schelling has emphasized the latter issue in a succession of papers on multiple selves that showed how traditional utility maximization models fail to explain the phenomenon of individuals attempting to exercise self-control or selfcommand (Schelling 1978, 1980, 1984). Looking at the issue of competing and conflicting sets of preferences from the perspective of ordinary experience, Schelling notes that people in everyday life often feel themselves torn between different values and different ways of looking at the same opportunities, so that they often feel they are inhabited by alternative selves (the study of which he labeled “egonomics”). As he sees it, the usual complaint people make is that they feel unable to determine which self among their different, competing selves is their “authentic” self. From this perspective, the single utility approach is not only a less rich framework but, by suppressing individuals’ sense of inner conflict and their strategies for confronting it, the single utility approach also distorts our representation of ordinary experience. Often, of course, individuals speak, as it were, with a single voice. However, Schelling does not tell us how to distinguish between those cases where the individual’s desires are incommensurable and the individual is inescapably conflicted and those cases where the individual’s desires are commensurable and the single utility case applies. How are we to investigate, then, where and when the individual is able to speak with one voice? Kavka (1991) argued that the logical way to approach the question is to liken multiple selves individual choice problems to multiple individual social choice problems. That is, intrapersonal collective choice problems are essentially the same as interpersonal collective choice problems.1 Kavka then frames the desire commensurability question in terms of three information requirements that agents would need to fulfill in order to produce single utility rankings over different subrankings (or “suborderings”): According to the conventional account, decisions are reached by an agent whose information about the desirability of alternative outcomes is of three sorts: (1) orderings of possible outcomes along various dimensions or according to various desires (what I have called “suborderings”), (2) information about the relative strength (or intensities) of the preferences within each suborder, and (3) information about the relative importance of the different dimensions or desires that determine the relevant suborders. (Kavka 1991: 146)
68
Individuation: multiple selves
The first information requirement entails that the individual be able to distinguish between different suborderings. The second and third information requirements concern what is required for agents to make intrapersonal comparisons of desires or preferences across their different suborderings. According to the conventional view, Kavka argues, agents would have to be able to assign sets of weights to different preferences in different suborderings, so as to determine overall their strongest preferences irrespective of the suborderings to which they belong, and thereby produce the single utility ranking needed for utility maximization. However, there are two difficulties with using the information requirements strategy in Kavka’s view. First, there is very little evidence that individuals have the information needed to assign preference weights to different preferences in different suborderings. Introspection, in fact, seems to indicate just the opposite. In ordinary experience, as Schelling emphasized, individuals often find themselves subject to competing and seemingly irreconcilable desires, and certainly on these occasions there is no inner cognitive apparatus that they may rely on to produce the needed weights to compare these desires. Second, and more seriously, since the purpose of assigning weights to preferences is to be able to say how individuals carry out what amounts to an aggregation procedure across their different preference orderings, it is open to us to consider any available aggregation functions as a means of explaining individuals’ organization of their different suborderings. Then, by analogy to the wide range of collective choice problems studied by economists and decision theorists—Kavka himself only considers a small sample: “internal” prisoner’s dilemmas, majority voting paradoxes, and Arrow’s impossibility analysis—it becomes apparent that the idea of there being a single utility ranking for the individual represents a special case, and that the general theory of individuals with different suborderings typically not aggregated into a single utility function would be a multiple selves theory. In their paper, Steedman and Krause did investigate the consequences of using Arrow’s impossibility theorem framework to explain individuals as having multiple utility functions. Taking each of an individual’s different points of view as representing a different utility function ui (.), an individual can be said to have a set of orderings (R1, R2, …, Rn) on the set of all alternatives A, where each Ri is an ordering for assessing alternatives from one perspective or point of view. The single utility function individual would then require some overall evaluation R on A, such that R is related to (R1, R2, …, Rn)—call it a character formation rule. A character formation rule F is a mapping from Ri to R onto the set of alternatives A that would model the individual as a single utility-maximizing agent. Steedman and Krause note, however, that an individual’s overall evaluation R may not be an ordering of the same sort as Ri, since Arrow’s theorem implies that individuals represented by standard ordinal utility functions may lack complete, transitive, all-purpose orderings of their alternatives. Indeed, one might imagine a variety of character formation rules depicting a whole range of different possible character types, only some of which would we be able to represent along the lines of the traditional model. Steedman and Krause
Individuation: multiple selves
69
accordingly develop and prove four propositions for what they term “harmonic” preferences that would need to be satisfied to regard an individual made up of many selves as a single utility-maximizing individual. However, that there is no reason generally to expect these four propositions to be satisfied tells us again that the single utility function account of the individual represents a special case. Kavka’s general comparison of intrapersonal collective choice and interpersonal collective choice invites ironic comment upon the label often taken as synonymous with the single utility function starting point: methodological individualism. Given the multiple selves problem, insisting that we always begin with single individuals makes methodological individualism tantamount to its historical antithesis, methodological collectivism. The latter stance is associated with the idea that individual members of social collectives such as classes are subsumed within the larger individuality constituted by the group. However, since the burden of proof rests upon defenders of Bentham’s psychological hypothesis, methodological individualism amounts to a proposal to subsume an individual’s multiple selves under a single individuality. Just as a social collective subsumes its individual members, so the single utility function agent subsumes an individual’s different points of view and the subutility functions they imply. Thus, if the principlal objection against methodological collectivism is that classes do not adequately represent the diversity and individuality of their individual members, a fair complaint against methodological individualism is that the methodologically collectivist single utility function agent does not adequately represent the multiple selves that comprise it. In effect, both the class hypothesis and Bentham’s psychological hypothesis are cut from the same cloth. Both assume questionable commensurabilities to justify operating with single agents that arguably fragment on closer inspection.2
4.2
Internal preference structures
The papers by Kavka and Steedman and Krause share the idea that individual choice faces intrapersonal Arrow-type impossibility results when subsets of preferences have the same standard ordinal utility function properties that are ascribed to the utility functions of different individuals. However, neither paper concludes that individuals are necessarily fragmented Faustian beings without any principle of internal unity whatsoever. Rather, their arguments have two objectives: a critical one in raising questions about the orthodox characterization of the individual, and a constructive one in pointing to the need for an analysis of how individual preferences might be organized. To this end they call for identifying “the aggregation rule that integrates the various dimensional evaluations into an overall decision” (Kavka 1991: 158) and “the agent’s integrating process or [character] formation rule” (Steedman and Krause 1986: 212). Yet, while we may be sympathetic to the idea that individuals are not fragmented Faustian beings, it is not clear that an aggregation or character formation rule for combining an individual’s different preferences can provide an adequate understanding of the unity of the individual.
70
Individuation: multiple selves
First, the notion of an aggregation or character formation rule does not by itself tell us much about how different preferences might play different roles in creating cohesiveness in character. Were different preferences to operate in some internal division of labor, then emphasis would better be placed upon the functional roles that different preferences play in structuring the behavior of the individual as a whole. This approach to the unity of the individual is the subject of the multiple selves literature surveyed in the remainder of this chapter. Second, aggregation and character formation rules may employ too narrow a scope for explaining the unity of the individual if it turns out that we must somehow understand individuality in terms of its embeddedness in a social context. This alternative strategy goes well beyond the neoclassical Lockean orientation, which represents individuals in terms of states of mind, and is taken up in Part II in connection with the heterodox economics approach to the individual. Those contributors to the multiple selves literature who seek to explain the unity of the individual in terms of a structure or organization of preferences reject Bentham’s hypothesis that all our desires or preferences are commensurable, and try to show how various asymmetrical relationships between subsets of preferences offer interesting opportunities for explaining the unity of the individual. Three types of strategy of analysis that employ these ideas are examined below: 1 2 3
Principal–agent models. Weakness of will and self-control arguments. Metapreferences or hierarchical organizations of preference analyses.
4.2.1
Principal–agent views of the unity of the individual
An early view of the internal structure of the individual was based on the idea that the individual is like a business firm in acting as a unitary organization whose subunits or divisions, like different selves, possess some measure of independence but are not completely autonomous from one another. Thaler and Shefrin used this idea to suggest that principal–agent relationships in firms exhibiting conflicts and bargaining between owners and managers might serve as a model for individual efforts at selfcontrol, which Schelling had argued could be understood as a process of negotiation between an individual’s different selves (Thaler and Shefrin 1981). One difficulty with this idea is that there is an element of circularity in the development of the analogy between individuals and firms. One reason that firms have been treated as single individuals is that doing so facilitates a symmetrical supply-and-demand analysis that places firms on the supply side and individuals on the demand side of the market. However, if firms are treated as single agents by analogy to individuals, then one cannot say that the firm as a single agent is a model for the unity of the individual. Nonetheless, the underlying conceptual structure of principal–agent relationships is still helpful for thinking about the unity of the individual, because it shows how mutual
Individuation: multiple selves
71
advantage may enable a set of different interests to unite in a common enterprise. The basic idea behind the principal–agent relationship, which dates back to Smith’s (1976 [1776]) thinking about how to realize gains from specialization in production, is that, by making different specific investments in production, individuals find it advantageous to enter into long-term relationships with one another, in that the combination of their different unique technologies and capacities enables them to realize higher returns on their investments than when outside these relationships. Thus, because productionspecific assets possess lower value outside the activities for which they are created, those participating in such relationships become dependent upon one another in a way that cements their participation in joint activities and gives unity to a common enterprise. Working against these unifying forces is the fact that individuals’ specialization in different specific assets also provides them (as agents) with unique information about the production process unavailable to others (their principals). This encourages opportunism in principal–agent relationships, and there is consequently also a tendency in shared enterprises for individuals to engage in “self-interest with guile … . More generally … opportunism refers to the incomplete or distorted disclosure of information, especially to calculated efforts to mislead, distort, disguise, obfuscate, or otherwise confuse” (Williamson 1985: 47). When business firms do succeed in developing on a principal–agent basis, then, the advantages of cooperation must exceed its disadvantages for all parties involved. Similarly, we might thus argue, individuals might act as single agents when on balance it pays for their different selves to operate in unison. In Schelling’s language, individuals succeed in practicing self-control when the advantages from cooperation between their multiple selves exceed the disadvantages from cooperation. The main difficulty with pursuing the analogy between the unity of firms and individuals much further than this, however, is that firms and individuals are different in too many ways. In the first place, it is unrealistic to apply to the multiple selves of an individual the idea that potential participants in a business enterprise bargain and negotiate with one another, since the different selves of an individual can really only be said to bargain and negotiate with one another in a metaphorical sense. Second, the basic idea underlying principal–agent relationships as applied to the individual, namely that the individual’s selves may find it to their advantage to cooperate when these selves have different specializations, provides only the beginnings of an understanding of how asymmetrical relationships between an individual’s selves might support their organization as a single entity. What more is there to this organization beyond the idea of mutual advantage? The analogy between the firm and the individual is unlikely to offer much more to go on, because standard theory takes mutual advantage to be a sufficient explanation of how independent individuals interact. It is difficult to believe, however, that human psychology is equally simple. Thus, though modeling the unity of the individual on principal–agent relationships introduces the idea of functional roles as a basis for understanding the internal organization of the individual, it does not take us very far beyond this idea itself.
72
Individuation: multiple selves
4.2.2
Weakness of the will and self-control
Perhaps the most studied and intensively analyzed multiple selves problem dates back to the philosophy of Plato and Aristotle, and is known as the problem of weakness of will or akrasia. Weakness of will is when an individual, desiring one thing more than another, somehow chooses the less desired of the two. In the Nichomachean Ethics, Aristotle termed such an individual incontinent, that is, impulsive and deficient in the virtue of self-control owing to poor character development (Aristotle 1962: 1145a–52a). One possible response to the difficulties this creates for a theory of the single individual is simply to deny the phenomenon that Aristotle describes. Why not just say, using the more modern idea of revealed preference, that the individual really acts on the stronger of two desires, and thus deny that weakness of will exists in the first place? A stronger response is to allow that weakness of will exists, but to say that the individual choosing the lesser of two desires lacks information and would have chosen differently with adequate information. The main application of this in the economics literature is to the myopic or shortsighted individual who favors the present over the future, but upon gaining more information about the world with the passage of time revises earlier consumption plans. Elster (1979: 65–77) thus explains myopia in terms of time preference and distinguishes it from weakness of will proper which he explains in terms of time-inconsistent preferences. Myopia is a matter of individuals exhibiting time preference when they allocate more consumption to the present than to the future. In contrast, weakness of will occurs when myopic individuals revise their consumption plans with the passage of time. As time passes and the future becomes the present, those myopic individuals who find themselves unable to live with what have become their past decisions, and who alter their previous consumption plans to favor the new present, exhibit weakness of will. In contrast, myopic individuals who do not revise their past consumption plans, or have time-consistent preferences, do not exhibit weakness of will. Elster characterizes these latter individuals’ preferences as consistently irrational on the grounds that myopia and a continual preference for the present is irrational (Elster 1979: 70). However, these individuals are at least not inconsistently irrational, as are those individuals who are both myopic and weak of will. From a multiple selves perspective, individuals who are myopic but not weak of will, and who are able to stick by their plans, always have their shortsighted selves in control of their future selves. Elster argues that these individuals adopt what he characterizes, following Schelling, as precommitment strategies, whereby, much like Homer’s Ulysses who had himself tied to his ship’s mast so he could hear the Sirens, individuals bind themselves through a variety of intrapsychic and extrapsychic means to their irrational time preference in the interest of consistency and in the face of subsequent temptation to abandon and revise their plans. Elster believes the study of such precommitment strategies to be both part of a theory of imperfect rationality (and the theory of second-best solutions) and also the basis of an explanation of how
Individuation: multiple selves
73
individuals solve the problem of having multiple selves. These strategies are classified as ways of manipulating the feasible set (restricting the set of physically possible actions and changing one’s reward structure by public side bets), ways of manipulating one’s character, and ways of manipulating available information (Elster 1979: 103). Elster compares his treatment to that of Strotz (1955–6) and Pollak (1968), who describe individuals with time-inconsistent preferences as being engaged in a strategy of consistent planning. Individuals who regularly revise their consumption plans with the passage of time may do so because they perceive sticking to their past decisions to be worse or less rational than abandoning their past decisions. This might be understood as a way of putting a “rational” planning self in control over a succession of myopic selves rather than as evidence of weakness of will, if it is perceived that the main “issue is how to cope with my predictable lack of ability to act upon my present intentions, and not to cope with the irrational component of these intentions themselves” (Elster 1979: 73). Elster’s view, however, is that the Ulysses allocation represents a better response to the multiple selves problem, because the idea of adopting a precommitment strategy seems more clearly to express Schelling’s notion of self-control. He nonetheless allows that “the choice between the Ulysses allocation and the inconsistently irrational allocation … depends upon which desire is the strongest, the desire for consistency (the wish to impose my decision upon my later selves) [or] the desire for rationality (my concern with the welfare of my later selves)” (Elster 1979: 76). That is, the individual can be represented either according to whether priority is given to present perspective, the self in the present, or to future perspective, the self in the future. The Ulysses allocation favors the present self, while the strategy of consistent planning favors the future self. We should ask, then, whether one of these conceptions is more plausible than the other as a solution to the problem of multiple selves. As Elster makes clear, there is nothing that really justifies one of these perspectives or selves over the other. Thus, there is no clear way of ruling out that individuals remain divided by competing perspectives or selves. Elster’s account, however, does differ from the Strotz–Pollack perspective in one important way. Precommitment behavior of the extrapsychic sort, which Elster explains in terms of the idea of individuals making public side bets (Elster 1979: 104), requires that individuals be understood in a social context. Public side bets, that is, only constrain the individual because others may monitor the individual’s commitments. Thus, one might go on to argue that an individual is a single self by virtue of being a member of a community of individuals who recognize the individual as one able to honor commitments. Unfortunately, Elster stays within the subjectivist framework and does not develop the idea of the individual being a member of a community of individuals, treating others as little more than a means of monitoring the individual’s side bets. However, his recourse to public side bets—his strongest example of a precommitment strategy—clearly raises the question as to whether it will ultimately be necessary to consider individuality in a social context.
74
Individuation: multiple selves
In this light, it is interesting to compare how Davidson (1980) believes weakness of the will occurs, and what it takes for it to be overcome. His analysis also rests on the idea that individuals possess multiple selves that may produce the incompatible judgments we associate with moral and other forms of inner conflict. The reasoning of our different selves, Davidson suggests, can be represented in terms of the idea of our having competing practical syllogisms, the conclusion in each case being—in keeping with the traditional Aristotelian view of the matter—an action that the individual ought to perform. How is weakness of will then possible? If we suppose, as Davidson suggests following Hempel (1965), that a central methodological principle of good inductive reasoning is that one ought to begin by assembling all relevant available evidence when deciding among a set of mutually exclusive hypotheses— Davidson relabels this the principle of continence—then weakness of will or incontinence derives from a cognitive error he terms “weakness of warrant” [Davidson 1986: 81; also cf. Schick (1991: 110–15) on “weakness of reasons”]. Thus, the incontinent individual, who suffers from “weakness of warrant,” is that individual who relies only on evidence bearing on one particular hypothesis. This is irrational and against the individual’s own best judgment, and is an example of how individuals often supply the wrong reasons for their conclusions. Here, they offer wrong reasons because they possess different, competing (cognitive) selves that are each ready to provide justification for different courses of action. Because individuals always have these competing selves, the virtue of continence and the capacity to assemble and review all the evidence relative to all competing hypotheses is something that is rare. Davidson’s argument brings out an important aspect of the weakness of will problem in making use of Hempel’s requirement of total evidence for inductive reasoning. Hempel originally advanced the total evidence requirement in order to establish standards for scientific investigation. Davidson adapts the requirement to the problem of how individuals might avoid weakness of will. Each, however, supposes that there is a community of individuals who will entertain these requirements in shared, mutually reinforcing social practices, suggesting that individuals generally need to rely on such requirements to sort out individual problems, personal in Davidson’s case and scientific in Hempel’s case. If we apply this conclusion to Elster’s treatment of weakness of will, emphasis on the making of public side bets shifts from the actions of the individual to the practices of a community of individuals. How weakness of will is overcome, then, is a matter of explaining how individuals participate in social practices, and how this enables them to achieve a unity of their multiple selves. 4.2.3
Metapreferences and hierarchical organizations of preferences
In his “Rational fools” paper on commitment, Sen also suggested that preferences might be hierarchically organized in terms of metapreferences or second-order preferences over first-order preferences to capture the idea that, in addition to being able
Individuation: multiple selves
75
to rank bundles of goods, individuals are also able to rank preference rankings themselves (Sen 1977: 102ff.). Some authors have questioned whether metapreference thinking should be thought of as multiple utility analysis (George 2001), but that Sen was working in terms of both these sets of ideas seems to suggest that metapreference analysis may be regarded as one response to the multiple selves problem. Indeed, metapreference analysis potentially offers a straightforward solution to the problem of the unity of an individual of many selves. On the one hand, metapreference rankings function like Kavka’s aggregation rules or Steedman and Krause’s character formation rules in coordinating an individual’s different sets of rankings, but on the other hand, unlike these kinds of rules, they provide a clear view of how different sets of preferences play different roles in structuring the cohesiveness of individual character. Specifically, since metapreferences are rankings of sets of preferences, they associate the importance of different preferences to the individual with the (lexical) order in which the different preferences will be acted upon. For example, should an individual have one set of preferences that can be grouped to reflect a general taste for self-indulgence and another set of preferences that can be grouped to reflect a general taste for self-restraint, the ranking of these two sets of preferences tells us in principle how the two sets stand in relation to one another, and thus how the individual acts as a single self. That the metapreference approach offers a possible solution to the problem of multiple selves is argued somewhat differently by Frankfurt (1971) and Dennett (1976). Frankfurt and Dennett link the identity of the individual to the individual’s being an intentional agent. This is not just a matter of the individual having beliefs and desires, that is, first-order intentions. The aspect of intentionality that distinguishes us as individuals, they argue, is our having intentions about our intentions, or second-order intentions. Essentially, our ability to detach ourselves from our primary intentions enables us to see those intentions as belonging to us. This capacity in turn enables us to see other individuals’ intentions as belonging to them. In this way we then come to see ourselves as individuals who are like other individuals in being members of a community of intentional agents. The metapreference framework also contains a way of responding to the concern that Kavka and Steedman and Krause have regarding the appearance of social choice types of problem within the single individual–multiple selves framework. An important barrier to social choice in Arrow’s collective choice framework was the independence of irrelevant alternatives condition that Arrow thought necessary to any complete, transitive, overall social choice ordering based on multiple individual utility functions (Arrow 1963 [1951]). Metapreference analysis, however, requires that we give up the irrelevance of independent alternatives assumption, because having second-order preferences over sets of first-order preferences means preference alternatives are no longer independent. Thus, just as social choice is possible once the independence of irrelevant alternatives condition is given up, so single individual–multiple selves choice should also be possible with preferences over
76
Individuation: multiple selves
preferences. Moreover, since there may be many types of metapreferences or rules for how first-order preferences are structured, it is reasonable to suppose that the other types of social choice problems contemplated by Kavka and others are addressable in the metapreference framework.3 There are, however, three important difficulties with the metapreference approach as a solution to the multiple selves problem, one of which can be dismissed, but two of which appear more serious. First, it might be said that the approach faces an infinite regress problem in allowing there to be metapreferences over preferences, since one might then say that there must also be metapreferences over metapreferences and so on without end. However, while it makes sense to speak of a hierarchy of successive metapreferences over metapreferences, this hardly implies that one never reaches a highest level of metapreferences representing the individual’s chief tastes. Some argument is needed to say that the formal possibility of an infinite regress corresponds to a real possibility, and it hardly seems plausible to say individuals should always have ever-higher metapreference rankings. A second, more serious question concerns whether it makes sense to say that there is one single set of highest-level metapreferences. If the hierarchy of metapreference rankings has a ceiling, it might nonetheless be composed of multiple metapreference rankings so that the individual still possesses multiple selves. Then, all the multiple selves problems encountered in the standard framework (Arrow-type difficulties, majority voting paradoxes, and internal prisoner’s dilemmas in Kavka’s selection) reassert themselves in the metapreference setting. In the final analysis, sets of metapreferences are still just sets of preferences.4 This leaves a third difficulty with the metapreference approach as a means of addressing multiple selves. It was suggested, following Frankfurt and Dennett, that an important advantage of the metapreference framework is that second-order intentionality introduces the concept of a community of intentional agents and thereby provides elements of an understanding of the distinctness of the individual. However, the argument of Frankfurt and Dennett seems to include a notion not easily elicited from the metapreference framework. Their basic idea is that an individual is distinguished from other individuals when able to see them as like intentional agents. But the full idea here seems to be that individuals who see themselves as being like other intentional agents are also seen by others as being intentional agents. That is, intentionality presupposes reciprocal relationships between members of a community of intentional agents. In this broader setting, however, it becomes unclear in what respects individuals can be said to have subjective rather than shared preferences, whether first-order or second-order, as the metapreference framework requires. Though there is certainly a sense in which preferences are specific to individuals and are accordingly subjective, what is not clear is the degree to which individuals’ preferences depend upon their interpretation of those preferences in a public language shared by the community of intentional agents. If individuals learn to recognize their preferences in a shared language, determining the dividing line
Individuation: multiple selves
77
between public and subjective becomes difficult.5 Lacking a way of articulating the subjective-public nature of preferences, then, the metapreference framework seems to raise more questions than it solves regarding the status of individuals. Perhaps it should be thought instructive that Sen abandoned the metapreference framework in his later work to focus on individual capabilities when discussing interpersonal comparisons between individuals.
4.3
Endogenously changing preferences
Thus far, attention has been focused on stable and unchanging internal preference structures. However, one might also investigate the idea of endogenously changing preferences on the assumption that understanding the dependence of choice on taste also involves an understanding of the dependence of taste on choice.6 This might allow treating the individual as a unity of multiple selves if individuals could be thought to deliberately influence the development of their future selves—a main theme of Schelling’s egonomics—in terms of how the policies they adopt influence their future preferences. Imputing some notion of agency to individuals in this way was an implicit feature of Becker’s time allocation–human capital model. Elster is more explicit about there needing to be a role for agency in understanding the preferences-based conception of the individual. In his view, those economists originally interested in preference change failed to raise these issues because: very rarely is it assumed that an agent can take an active or strategic attitude towards his own preferences. Rather he is seen as the passive vehicle of preferences that change according to some logic that he does not himself understand. (Elster 1979: 77–78)
His own contribution, then, involves supposing that individuals actively adopt precommitment strategies that bring about intended change in their characters through change in their preferences that enable them to subsequently control their future behavior (Elster 1979: 104). Speaking of preference change as an adaptive preference formation, Elster distinguishes between two sorts of cases (Elster 1983: 109ff.). On the one hand, we might imagine an intentional or engineered adjustment of preferences when an individual’s metapreferences justify a planned change in character, and on the other hand we can imagine a causal process of adaptation that is explained in terms of how an individual’s drives operate behind the back, as it were, of the individual. An example of the latter is the “sour grapes” phenomenon whereby a causal mechanism for dissonance reduction within the individual’s system of preferences leads to a change in preferences that then accommodates a set of unpreferred changes in the individual’s opportunities.7 The former sort is closely linked to the metapreference framework in that a planned change in tastes requires that the individual possess some higher
78
Individuation: multiple selves
standpoint or set of “metadesires” that substitute for causal adjustment of tastes by drives (Elster 1983: 111). Essentially, should an individual possess higher-order and lower-order desires, strategic behavior is a matter of higher-order desires altering the nature or operation of lower-order desires. In this respect, the idea is much like what in the social welfare literature Cowen terms the “cleansed preference” standard, that is, the attempt to establish what an individual hypothetically would want were he or she fully informed (Cowen 1993: 262ff.).8 Just as social planners hoping to devise an adequate social welfare function may postulate fully informed preferences, so Elster’s strategic individual decides at the metadesire level what lowerorder desires should be encouraged to develop. Problems, however, that Cowen finds in this social welfare approach to preferences also apply to Elster’s approach to individual welfare and preferences. Different sets of preferences can be formally treated as either belonging to different individuals or as belonging to single individuals whose preferences change (cf. Malinvaud 1972). In effect, an individual whose preferences change from one point in time to another could equally be regarded as two different individuals with different preferences. Cowen notes this in order to argue that exclusive reliance on preference information in policy formulation concerning present and future generations raises questions about the ideal of consumer sovereignty. For Elster and those who argue that individuals may strategically change their preferences, the difficulties concern the identity of the single individual. We saw above in connection with Elster’s discussion of weakness of will that the preference framework lacks sufficient structure to show how an individual can be both weak of will and able to devise precommitment strategies. This point is reinforced here when we see that, just as for Cowen it is difficult to distinguish our future preferences and the preferences of future generations (Cowen 1993: 258–61), so in the changing individual preference framework it is difficult to distinguish the case of the changed individual from the case of different individuals. The problem in thinking about how individuals strategically alter their preferences, it can be argued, derives from our initially being unable to individuate them as single individuals, that is, pick them out as separate and distinct from one another. Different order preferences are given as if they belonged to single individuals, but nothing in the nature of these preferences imprints them with the identity of any particular individual as opposed to any other. Of course, an individual A may say, “I prefer x to y,” and thus appear to link self and preferences. However, x may also be preferred to y by individual B who says the same thing, and in fact nothing in the simple registering of x being preferred to y specifically distinguishes A’s ownership from B’s ownership of this preference. Thus, though the analysis Elster gives purports to say that a single individual can change his or her preferences and remains the selfsame individual, there is nothing beyond the assertion that the same individual is involved after such a change in preferences to rule out that we are not dealing with another individual. Thus, failure to individuate and pick out individuals whose preferences change in the first place
Individuation: multiple selves
79
precludes reidentifying these selfsame individuals across change. In the changing preference framework as currently understood, we cannot trace an individual through a set of changes if we are unable to pick out that single individual both before and after those changes. This conclusion essentially repeats that regarding Becker’s time allocation–human capital analysis in the previous chapter. Accordingly, just as significant problems concerning personal identity arise in multiple selves frameworks that assume stable internal structures of preferences examined above, so such problems also arise in connection with changing preferences.
4.4
A unity of selves or a plurality of selves?
In all the approaches to the multiple selves problems discussed above, the strategy is always to show that an individual made up of many selves somehow remains a unity.9 All of the contributors to this literature basically agree on the need to produce a better theory of the individual, and it is probably fair to say that on some level they all agree that the need for a better theory derives from inherent weaknesses in the traditional preferences-based conception of the individual. Consequently, in some instances the strategy adopted is to remodel the analysis of preferences, such as in the early views of different kinds of preferences and in the metapreference framework. In other instances the strategy adopted is to introduce new, non-preferences-based principles that refer to social context, such as Elster’s agency-based precommitment argument or the Davidson–Hempel requirement of a community of reasoners, in order to supplement the basic account of the individual in terms of preferences. However, in all cases it is assumed that the human individual is nonetheless a unity, even if we are unable to say how this is the case. Are we entitled to accept this assumption? The practice of economics should give us pause in this regard. Single individuals, in fact, rarely appear in economic models and in the analysis of markets. Rather, what we generally find is that collective individuals, such as households and firms, count as individuals, subsuming the human individuals who make them up. Thaler and Shefrin used this fact to justify thinking of human individuals as unities, but the circularity in their argument opens the door to reversing their argument and to saying that, if human individuals are unities just like households and firms, then, as the latter must thus be made up of individuals, so human individuals would also be made up of individuals, or are multiple selves. That is, the ordinary practice in orthodox economics of treating collections of individuals as individuals basically tells us that there is no obvious reason for thinking of the human individual as a unity, at least within the traditional neoclassical framework. Here, then, we are reminded of why the problem of individuation is important in economics. Philosophers may set aside this issue in focusing on the problem of reidentification as central to understanding personal identity, but in economics we have strong grounds for not treating human individuals as unities, in spite of economists’
80
Individuation: multiple selves
longstanding conviction that orthodox economics is primarily about human individuals. This chapter demonstrates that this conviction is not well supported. However, the goal of this book is not to eliminate the individual from economics, or even merely to show that neoclassical economics does not live up to its most wellknown self-promotion. Rather, the goal of the book is to determine the basis for an adequate conception of the individual in economics, where adequacy of a conception is understood in terms of whether it describes individuals that are both distinguishable and reidentifiable. The preferences-based conception of the individual in its simple and more developed forms is not adequate in this regard. However, the discussion in the last two chapters has also left us with two new threads that might be pursued in search of an alternative account of the individual. One is the idea of adding agency notions into one’s thinking about the individual, implicit in Becker and recommended by Elster. The second thread is the idea that social context somehow enables us to individuate and reidentify individuals. Becker’s model raises questions about social context in relation to ability and education, Davidson’s reference to a community of reasoners makes social context indispensable to overcoming weakness of will (or warrant), and Frankfurt and Dennett contemplate communities of intentional agents. In Part II, I seek to explain the individual in a social context by explicating the idea of the individual being socially embedded. However, before turning to that investigation, I devote one further chapter to the subjectivist tradition of the individual in economics. Actually, Chapter 5 looks at this tradition “after the fall,” since it examines the abstract individual conception in early mainstream economics after the neoclassical subjectivist view of the individual had been largely abandoned in orthodox economics. Chapters 3 and 4 looked at the fullest development of the subjectivist conception of the individual in neoclassical economics, and found that conception wanting. Where does a tradition go once it has given up a central pillar? Again, my argument is that, having begun with the Enlightenment dualism of subjectivity and objectivity, loss of the former leaves the field to the latter. This does not mean that individuals cease to appear in mainstream economics. However, they are now characterized as abstract individuals who bear little resemblance to human individuals.
5 After the fall
The machinery of choice
Indeed we can consider the individual—with his given indifference map and initial endowment—to be a utility computer into which we “feed” a sequence of market prices and from whom we obtain a corresponding sequence of “solutions” in the form of optimum positions. (Patinkin 1965: 7) A cyborg is a cybernetic organism, a hybrid of machine and organism, a creature of social reality as well as fiction … . In a sense, the cyborg has no origin story in the Western sense—a final irony, since the cyborg is also the awful apocalyptic telos of the West’s escalating dominations of abstract individuation, the ultimate self untied at last from all dependency. (Haraway 1991: 151) Strawson: Would a computer have a soul? Swinburne: The answer to that is certainly that modern computers do not, but that in the next century, computers might. (Peacocke and Gillett 1987: 50) In a phrase, methodological individualism is being slowly and inexorably displaced by methodological cyborgism. (Mirowski 2002: 441)
Contemporary mainstream economics contrasts with the relative hegemony of past neoclassical economics in being a multidimensional, pluralistic endeavor made up of a variety of different and competing currents of thought (Morgan and Rutherford 1998). Game theory, bounded rationality, experimental economics, behavioral economics, evolutionary economics, new institutional economics, and a number of other recent strategies of investigation are pursued alongside a still active, if now less ambitious, neoclassical research program. Thus, there are a variety of incipient, new ideas about individuals in mainstream economics, some of which still draw on neoclassical assumptions and others that introduce altogether new themes about the nature of individuals. However, in my view these new currents have to date generated only different collections of loosely related ideas that may or may not subsequently
82
After the fall: the machinery of choice
be woven into new theories of the individual. For example, bounded rationality and evolutionary game theorists often characterize individuals roughly as “rule followers.” However, this sort of characterization is less a theory about the nature of individuals and more an eclectic set of ad hoc claims about individual behavior driven first and foremost by broad goals involved in elaborating alternative theoretical strategies for economics after the demise of general equilibrium theory brought about by the SMD results.1 Thus, I suggest that, for the present at least, the only alternative theorization of the individual in mainstream economics is the default abstract individual conception, more the inadvertent outcome of the history of critique of the subjectivist neoclassical view than an intended destination. What is the character of this conception? In this chapter I offer only one interpretation of this abstract individual conception by arguing that it shares much the same philosophy of mind underlying an important early strand of cognitive science, namely, computational functionalism, or the view that mind is a computer and the individual a symbol-processing system. While there is no longer a single philosophy of mind underlying contemporary cognitive science (van Gelder 1995), computer functionalism did form its early core, and is also the basis for the classical view of artificial intelligence. My argument, then, is that this vision of the individual and mind is the one that best unifies the different ideas that came together in the immediate postwar period to explain individuals in abstract terms. Let me, however, register three caveats regarding my argument. First, as cognitive science is no longer dominated by a single philosophy of mind, and is not the only source of ideas about mind, new views of mind and the individual are likely to be adopted in the future by those working in the increasingly heterogeneous landscape of mainstream economics. Consequently, the view that I argue can be thought to unify the abstract individual conception in immediate postwar mainstream economics may well have a relatively short half-life. It is nonetheless interesting as the conclusion of a long history of thinking about the individual in economics that traces its roots to the Enlightenment. Second, I do not attempt to offer a historical argument regarding how cognitive science and computational functionalism may have directly or indirectly influenced the development of early mainstream economics. Thus, my methodology here departs from that in Chapter 2 where I did seek to provide an outline of how, historically, neoclassicism progressively eliminated its inherited subjectivism. Rather, my goal in this chapter is to offer an analytical/philosophical argument regarding how the metaphor of the mind as a computer helps unify our understanding of early mainstream economics on the subject of the individual as it emerged from neoclassicism. Third, I also do not claim that the mainstream figures I discuss below thought themselves attached to computational functionalism and the idea that the brain is a computer. Indeed, economists often believe themselves attached to a variety of philosophical ideas irrespective of whether these ideas have a genuine basis in their thinking about economics. My view, rather, is that the development of ideas about individuals in early mainstream
After the fall: the machinery of choice
83
economics can be given a clear focus in connection with the notion that individuals’ mental processes can be represented naturalistically as a set of highly determinate mechanisms understood as computational processes. Cognitive science itself is a wide-ranging, cross-disciplinary research program originally influenced by the work of Turing on artificial intelligence, which investigates cognitive processes and mechanisms that happen to be implemented in the brain but that could equally be implemented in an indefinite number of other hardware systems. My primary interest is in two key conceptions of these cognitive processes and mechanisms: that they are (a) unconscious, a paradigm example being Chomsky’s universal rules of grammar (Chomsky 1986), and (b) algorithmic or computational in nature, as instanced by digital computer processing. However, even cognitive science narrowed in these ways is far from being a monolithic and homogeneous research program, and as a multifaceted research program it finds different expression in psychology, linguistics, neuroscience, philosophy, and the theory of artificial intelligence. One broad strategy, however, is to simulate thought and cognition as information and algorithmic procedures for its manipulation. Computer simulations of human thought thus treat the brain as a digital computer and mental processes as computational processes. A second broad strategy is to study various types of “cognitive” subjects, and then to infer their rules of operation. For example, the behavior of one’s thermostat, or one’s neighbor, is said to involve cognitive mechanisms in which the respective agents respond “cognitively” to changes in their immediate environment. A third broad strategy is to analyze puzzlesolving and game-playing in the way that it employs heuristic principles of search through a problem space. These heuristic principles represent generic thinking devices or tools for problem solution that might be adopted by a variety of types of cognitive being. Early mainstream economics, I believe, was comfortable with all of these strategies, though it worked out none of them particularly carefully. In Section 5.1, I begin with a brief history of how one tradition in postwar philosophy of mind produced cognitive science and the view of the mind as a computer. This survey covers identity theories of the mind, functionalist views of mental states, and early philosophical thinking about artificial intelligence. In Section 5.2, I show evidence of attachment to this same philosophy of mind in the thinking of five key contributors to early postwar economics—Arrow, Samuelson, Friedman, Simon, and Lucas—arguing that each provided pieces to a puzzle that can be assembled to produce a cognitive science vision of mind and the individual for economics. My argument is not that these individuals (with the exception of Simon and perhaps Lucas) saw themselves as engaged in reconstituting orthodox economics along cognitive science lines. Since cognitive science’s emergence as a self-conscious research program largely postdates the main works of these individuals, I rather argue that their work shares common lines of development with what emerged as cognitive science. In Section 5.3, then, I briefly take stock of the theory of the individual in orthodox and mainstream economics, and then introduce Searle’s views regarding
84
After the fall: the machinery of choice
mind and consciousness to explain what I believe to be mistaken about the philosophy of mind underlying cognitive science and early mainstream economics. Finally, I link the philosophy of mind to the issue of producing an adequate theory of the individual, emphasize the importance of the concept of individual position, and point towards the themes of the second half of the book.
5.1
Mind as a computer
The philosophy of mind is about the nature of mental phenomena, such things as our beliefs, thoughts, wishes, feelings, desires, and so on, and how they are to be understood relative to the causal structure of reality. The classic problem in philosophy of mind is the mind–body problem, or whether mental phenomena are physical phenomena, and, if not, how might they be related to physical phenomena. We saw in Chapter 1 that Descartes gave expression to the mind–body problem when he characterized mind and body as two entirely different substances. In postwar Anglophone philosophy, one main response to Cartesian dualism has been a variety of single-substance theories that assert the identity of the mental and the physical, and treat mental states as brain states. However, if brain states are identifiable as computer states, then this implies that the mind itself is a computer. This inference emerges in the succession of views from identity theories of the mind to functionalist views of mental states to the theory of artificial intelligence. 5.1.1
Identity theories of the mind
Early proponents of the identity theory were the philosopher Smart (1959) and the psychologist Place (1956) who argued that sensations are identical to brain processes. To avoid the typically mentalistic language of our ordinary talk about sensation, they recommended that terms normally used to describe sensations be rephrased in a “topic-neutral” vocabulary, so that, for example, reports of seeing an orange would rather be reported as, “There is something going on in me that is like what goes on when I see an orange.” According to this view, later termed central state materialism, all mental states are contingently identical to states of the brain or central nervous system. As Smart put it, the identity of the mind and the brain was true as a matter of fact, just as it had turned out to be a matter of fact that lightning is nothing but electrical discharge. By this token, mental states have causal effects in that the exercise of mental capacities consists of brain processes or neurophysiological activities that produce observable behavior. The ambition behind the identity theory, inspired by the ideal of the unity of science, was to provide a reductionist account of mental states as brain states that relied on bridge laws that asserted the equivalence between the two. However, this required that types of mental states be strictly identified with types of brain states, and it seemed quite unlikely that for every type of mental state there was one, and only
After the fall: the machinery of choice
85
one, type of brain state to which it was identical in each and every person (Block and Fodor 1972). Why should two people with the same belief about something or other have identical neurophysiological configurations in their brains? However, a weaker, nonreductionist form of the identity theory held that for every token instance of a mental state there was some token neurophysiological event to which it was in fact identical, thus suggesting a token–token identity theory rather than a type–type identity theory. Though it could not be said that some belief that p was identical to a type of firing of neurons, r, in the brain, it could still be said that Smith’s particular belief that p was identical to a particular firing of neurons in Smith’s brain at time t. This weaker thesis seemed to preserve the materialist unity of science motivation behind the original identity theory and still avoid Cartesian dualism, though it gave up the goal of finding a single set of psychophysical bridge laws that accommodated mind–brain to the causal structure of reality. The token–token identity theory, however, faced its own difficulties. If two people can be in the same mental state, for example, believing the same thing, but can be in different neurophysiological states when doing so, then what is it about those different neurophysiological states that makes them the same mental state? Given the goal of producing a materialist theory of mental states, it would hardly do to say that the two neurophysiological states were associated with the same sort of mental state, since this would mean explaining brain states in terms of mental states rather than the reverse. Faced with this conundrum, one response by those attached to the project of a physical science understanding of the world and mind was to reject ordinary views about the mind and mental states as mere “folk psychology,” a primitive form of thinking that was likely to turn out to be false as a more mature cognitive science emerged. Termed eliminative materialism, and best represented by Churchland (e.g., 1984), proponents argued that, were our folk psychology convictions about beliefs, desires, hopes, fears, and so on, to turn out false, then the entities named by these expressions—the mental states of believing, hoping, and so on— could not really exist. One of the most sophisticated arguments in this regard was that of Quine (1960) who argued that the holistic character of mental or intentional concepts allowed there to be a variety of competing hypotheses about an individual’s mental states that were all compatible with the same set of behavioral facts about that individual. Good science depended upon developing extensionalist semantics that permitted clear assignment of truth values to sentences. Since this was not possible when employing the intentional idiom, that is, the language of mental states, it was best to move beyond ordinary ways of thinking about the mind, eliminate propositional attitudes such as “believe that,” “wish that,” and so on, and describe the world extensionally in terms of truth conditions for physical events. Rather than reduce mental states to brain states, Quine’s proposal was to discard the intentional idiom altogether in favor of a fully natural science treatment of observable behavior. Behaviorism was thus the consequence of eliminative materialism.
86
After the fall: the machinery of choice
5.1.2
Functionalist views of mental states
An alternative response to the question of how two brain states could be tokens of the same type of mental state—one that has acquired a wide following among philosophers—was that different brain states identified the same mental state if they performed the same function for the individual. Functionalism is the view that mental states should be defined in terms of their causal relations to input stimuli, output behavior, and also other mental states (Lewis 1972; Grice 1975). If, for an individual, a particular sensory event is conjoined to a specific behavior, then we may distinguish and identify a mental state of that individual in terms of that sensory event and resulting behavior. Mental states, then, are functional states of the individual in the sense of being ways in which the individual mentally functions in relation to the causal relations that he or she occupies. As such, functionalism goes beyond eliminative materialism and behaviorism in allowing that there are mental states that are internal states, albeit ones that always have some generally external physical realization. Functionalism also escapes old Cartesian dualism, since it denies that the mental involves a nonphysical substance. Finally, functionalism avoids the simplistic reductionism of early type–type identity theories that sought to translate the language of mental experience into a neurophysiological language, since it gives an account of how the same mental state can be implemented in two different brain states. Indeed, an important conclusion of functionalist thinking is that, because two different brains—or physical systems of any kind for that matter—could share the same functional or mental states in that they might involve the same causal relationships, mental states can be implemented in any kinds of hardware. Functional states, that is, exhibit what may be termed multiple realizability in different kinds of physical system, whether they be humans, computers, or Martians, since the same functional/ mental state could be observed in different sorts of occupant. One might, then, even drop the language of “mental states” altogether in favor of the language of “functional states,” since the former, because of its traditional associations, tends to suggest only one type of occupant, namely, human beings, whereas the theory applies more generally to functional states that could be implemented in a range of different possible occupants.2 Lewis (1972), in fact, suggested a way of doing away with any reference to the mental through a technical device called the Ramsey sentence. In defining mental states in terms of their causal relations, were one to suppose, say, that some belief of Smith’s that p stood in relation to Smith’s perception r, and want to say that, together with Smith’s desire that q, this belief that p caused an action, a, on Smith’s part, then, rather than refer to this belief in psychological terms, one might replace such language by saying that there is something, x, that stands in a causal relation to Smith’s perception and action, or, more formally, using the existential quantifier (Ex) (Smith has x & x is caused by the perception that r & x, together with a desire that q causes a)
After the fall: the machinery of choice
87
Additional Ramsey sentences would then similarly replace occurrences of the remaining mental terms (“perception” and “desire”), thus demonstrating that defining mental states in terms of neutrally specified causal relations makes all reference to the mental unnecessary. In relation to this, Lewis’s account also makes it clear that we need not understand how mental processes operate within the mind, since by his account they are fully represented in terms of our specification of their causal connections in the world. For this reason, functionalism has often been called “black box functionalism.” Essentially, whether there are internal mental processes is irrelevant to our account of the individual’s functional states. One early, influential version of functionalism by Putnam, was termed machine state functionalism or Turing machine functionalism (Putnam 1967). Turing machines are defined in terms of tables that indicate transitions from input states to output states (Turing 1950). Turing machine functionalism is an algorithmic or computational type of theory that treats mental states as Turing machine table states. Another early version of functionalism was conceptualized by Armstrong in terms of a causal theory of mind. Armstrong, an early proponent of central state materialism (1968), may be counted as a functionalist in arguing that the concept of a mental state is the concept of a state apt to bring about certain kinds of behavior (1980). Putnam’s computational emphasis and Armstrong’s causal emphasis are complementary, and both have played a role in providing philosophical foundations for development of the theory of artificial intelligence as the most sophisticated and influential form of functionalist philosophy of mind. 5.1.3
The theory of artificial intelligence
The theory of artificial intelligence further extends functionalist thinking in providing an answer to a question raised by the idea of the “mind” as a kind of black box. With the actual working of the “mind” taken to be a black box, we are unable to say in the language of science how different material phenomena give rise to the same causal relations. How is it that different “brains” or different kinds of physical systems operate in the same cause-and-effect ways when in science generally different physical structures exhibit different causal relations? Moving beyond the idea of the mind as simply a black box, the theory of artificial intelligence answers that different physical structures can be causally equivalent if they are different hardware implementations of the same computer program. The advent of computers created a distinction between the computer and its program or between a system’s hardware and its software. Given this distinction, the mind is simply a computer program that might be implemented in a human or computer hardware, or, as many have put it, “the mind is to the brain as the program is to the hardware” (Pylyshyn 1984; Johnson-Laird 1988). Thus, different material systems or structures can give rise to the same causal relations if those structures are operated by the same programs. Moreover, because computer programs are perfectly determinate, the different patterns of action they
88
After the fall: the machinery of choice
give rise to must be perfectly determinate as well. Computer functionalism, as it was called by some (Dennett 1978), actually extends cause-and-effect explanation fully into the “mind,” provided that the “mind” is understood essentially to be a computer program. It is now possible to see clearly how functionalism is the underlying philosophical basis for cognitive science. Though a wide-ranging, multidisciplinary research program, cognitive science as a whole may be defined as the investigation of computational processes that may be implemented in different types of physical structures. Computational processes are sequences of well-defined purely formal symbols or syntactical objects, the content or meaning of which “is irrelevant to the way they get processed” (Segal 1991: 463). Moreover, as purely formal mechanisms computational processes neither need be understood nor consciously apprehended to constitute an organized system of connection. Indeed, computational processes might be regarded as a silent “language of thought” (Fodor 1975) or as sets of unread “sentences in the mind” (Field 1978; Harman 1977) that may be assigned various semantic contents or interpretations according to the uses to which we wish to put them. More generally, functionalism provides the necessary theorization for cognitive science in that it offers an account of mental phenomena as computational processes compatible with a causal analysis of reality. This same conception, and this same understanding of mental phenomena, I argue in the next section, underlies the early development of much of mainstream economics in the postwar period. From this perspective, this economics is one branch of cognitive science theory.
5.2
Economics as cognitive science
One way to see how mainstream economics, as it emerged from neoclassicism, produced a view of the individual understandable in terms of cognitive science’s computational functionalism is to consider its relationship to formalism. When formalism in economics is criticized, economists usually reply that new mathematical and statistical tools are necessary because they enable economists to sort through complex phenomena, and solve longstanding problems in the discipline, notably those investigated by Smith and Walras regarding the functioning of an equilibrium price system (e.g., Krugman 1996: 135–6; Varian 1996: 243). At the same time, they allow that an unfortunate but unavoidable byproduct of the introduction of these tools is that, because economists compete with one another for reputation, prestige, and earnings, economics has at times become absorbed in what might be termed a mathematical tools “arms race” that gives rise to the charge of excessive formalism. The implication of this is that a sufficient corrective for the situation would be for economists simply to make more judicious use of mathematical tools. What this response overlooks, of course, is whether mainstream economics has been transformed by its tools. More forceful critics of contemporary mainstream economics consequently argue that economics today is less about solving problems
After the fall: the machinery of choice
89
regarding how markets work and more about solving problems in formal economic theory (e.g., Backhouse 1996: 8; Hodgson 1996: 106). Thus, if economists’ conceptions have been changed by the mathematical technologies they have adopted, and if they have become fully reliant on those tools in their theorizing, it might well be argued that economists have lost the resources necessary for seeing individuals as other than mathematical objects. However, the impact of mathematics on mainstream economics may go further than this. If economics is changed by the tools it uses, then it might also be argued that economists themselves have been transformed by the tools they have adopted. Pickering has argued that scientists and their technologies become so entangled with one another that they come to constitute one inseparably intertwined human– machine, or, as he terms it using Haraway’s metaphor (Haraway 1991), a “supercyborg, harnessing … material and human performances” in a single process (Pickering 1995: 145). Rejecting both the idea that tools are neutral and the representational concept of science as an independent body of knowledge to which detached scientists contribute, Pickering argues that scientists operate in practice-oriented communities that produce a complex mangle of tools, theories, and scientists. Scientists are continually engaged in adjusting their goals and conceptions of what their investigations are about according to what their tools allow them to achieve in what he calls a “dialectic of material resistance and scientific accommodation.” They thus learn from how their technologies work what their theories are actually about, and do not stand disinterestedly apart from their investigations, sifting through alternative hypotheses to see which best correspond to the facts or are falsified by the evidence. Rather, scientists, their theories, and their technologies are all in a process of continually evolving together. In this regard, economists do not use mathematics as a ready-to-hand tool to work out new solutions to old problems about how individuals interact in markets. Rather, from the point of view of the mathematical economist individuals can only be mathematical objects and human rationality can only be a machinery of choice. Thus, it is largely beside the point to complain that there is “too much” mathematics in economics, and it is also not enough to say that excessive formalism in economics implies that mainstream economics is no longer about real-world problems that arise in our attempts to understand the way that markets work. Rather, from the point of view of the mathematical economist, since the real world is a Pythagorean one governed by number, real individuals must be abstract individuals understandable in terms of their mathematical properties. By this token, the theory of rational choice is not a normative theory but rather a descriptive one. What, then, are the chief features of this mathematical abstract individual that has replaced neoclassicism’s subjective atomistic individual? The key underlying philosophical commitments of mainstream mathematical economics, I suggest, are those of early cognitive science which models individuals as mathematical or computational machines. There are four key respects, introduced in the last section in terms
90
After the fall: the machinery of choice
of this conception’s underlying philosophy of mind, that give structure to our understanding of the abstract individual in early mainstream mathematical economics as it emerged in the wake of a ruined neoclassicism. They are: 1 2 3 4
Mental processes of individuals are treated as computational processes. Mental processes are treated as unconscious processes. Mental processes may be implemented in a variety of types of hardware. Mental processes are governed by a deterministic cause-and-effect logic.
Premises for these themes have been developed to varying degrees by many of the contributors to early postwar mainstream economics. Here, I only illustrate this in a selective way by describing how one or more of these themes emerged in the ideas of five postwar Nobel Laureates in economics: Arrow, Samuelson, Friedman, Simon, and Lucas. I do not attempt to offer a comprehensive account of these themes in their thought, nor argue that they saw themselves as contributing to the functionalist paradigm. My argument is simply that each of these individuals in different ways contributed important pieces to the puzzle that made computational functionalism the implicit philosophy of mind underlying thinking about the individual in early mainstream economics. 5.2.1 Arrow: “mental” processes as computational processes implemented in different hardwares Formalist economics replaced an earlier common-sense type approach to economics that relied on ordinary language, and did not require axiomatic-deductive organization of assumptions and conclusions (Coates 1996). Why was this earlier, less formal economics jettisoned to make way for formalist economics? And how did the move to formal representation of individuals effectively transform them into mathematical objects? To answer these questions, and to begin describing the links between formalist economics and functionalist cognitive science, it helps to look briefly at a parallel formalist project, namely, the project of creating formal semantics for science, especially as popularized in the postwar period in the work of Quine. As noted, Quine’s goal in developing formal semantics was to produce a logical language in which the truth claims of science might be unequivocally expressed. The same sort of project was taken up by early formalists of mainstream economics in their turn towards axiomatic explanation in economics, the exemplar being the Arrow–Debreu model of general economic equilibrium. My view is that this model set the postwar standard for what constituted “good” economics, in the process demonstrating the advantages of a narrow construal of preferences. This narrow construal of preferences was itself a step in the direction of a functionalist view of mind. First, however, a little background about Quine.
After the fall: the machinery of choice
91
Following the prewar work of Tarski, Quine’s idea of formal semantics is of what is called a fully extensionalist language, or one that restricts meaning and truth to settheoretic relations between names and predicates that have spatiotemporal objects and abstract entities respectively as their extensions (Quine 1960). The extension of a name is the particular object denoted by that name; the extension of a predicate is the set of objects having the property represented by that predicate. An extensionalist language is one in which any expression in a sentence can be replaced by another that has the same extension without change in the truth value of the sentence. The advantage of such a language for science is that it preserves the truth value of sentences across the use of different expressions for the same things. Expressions with the same extensions may be exchanged salva veritate. However, another thing that an extensionalist language eliminates is all language expressing intentional life, that is, beliefs, desires, fears, and so on. When a sentence expresses what a person believes or desires, Quine made use of what were called propositional attitudes, logical expressions such as “believes that –––” or “desires that –––,” where the blank is filled in by the relevant belief or desire. The problem with these expressions, termed intensional expressions,3 is that they are vague and referentially opaque, so that the replacement of what appears in the blank by expressions with the same extension often changes the truth value of the sentence in which they appear. It may be true that Smith believes that the evening star has risen, but not true that Smith believes Venus has risen, though “evening star” and “Venus” have the same extension. Quine’s solution to this problem was simply to exclude propositional attitudes, intensional expressions, and the phenomena of our intentional life from formal semantics, so as to concentrate exclusively on observable behavior where replacement of one expression by another does not cause truth value fluctuation. The difficulty with our intentional idiom, he argues, is that these notions can only be understood in terms of one another: “there is no breaking out of the intentional vocabulary by explaining its members in other terms” (Quine 1960: 220), so that we are better off setting this language aside altogether if “we are venturing to formulate the fundamental laws of a branch of science” (Quine 1960: 221). Arrow’s initial interest in his undergraduate studies before turning to economics was in mathematical logic—an interest reinforced by an early course he took from Tarski. Later, Arrow reconnected with Tarski when he read the proofs for a translation of Tarski’s logic textbook while he and Tarski (and perhaps Quine) were at Rand together (Mirowski 2002: 297, 422). Thus, it should come as no surprise that, throughout his career, Arrow has remained committed to formalizing economic argument. Indeed, his own interest in mathematical logic connects to the prior genesis of the Arrow–Debreu general economic equilibrium model itself, the antecedents of which are to be found in papers and discussions in Menger’s Mathematical Colloquium of prewar Vienna where participants included Tarski and others interested in the development of formal logic (Weintraub 1985: ch. 6). It is interesting, then, that Arrow began his influential social choice possibility theorem argument by
92
After the fall: the machinery of choice
noting that there were serious obstacles to making sense of social choice that derived from the fact that economists made different value judgments in their representation of economic agents’ tastes and desires (Arrow 1963 [1951]: 3–5). Economists’ different value judgments, Quine might have said for Arrow, are intentional phenomena whose expression in intensional contexts resists formal analysis. This suggests that we might do better to put aside economists’ judgments in welfare analysis and specify a set of conditions under which economic agents themselves might make social choices, thus eliminating ambiguities in economists’ ordinary ways of speaking about social choice. It is also interesting that one of Arrow’s “reasonable” conditions for social choice was his independence of irrelevant alternatives assumption which eliminates intensity of individual preferences. Preference intensity, or how strongly we feel about a preference, adds mental depth, as it were, to a preference, since, in addition to the preference itself, we exhibit a subjective stance towards that preference. It seems fair to say, then, that Arrow, like Quine, implicitly recognized that allowing subjective or mental depth, either at the meta level or within the object language of analysis, made one’s results less precise. Like Quine, Arrow placed a high value on being able to generate clear and uncontroversial propositions for science, and this could be facilitated by eliminating that range of intentional experience that resisted formal regimentation. In this respect, Arrow put a new slant on the old ordinalist conception of a preference. Mathematically speaking, a preference is simply a formal relation between an agent and the objects to which the agent’s preference applies. However, in excluding information about preference intensity, this formal conception of a preference also eliminates the need for saying anything about an individual actually experiencing having a preference. Indeed, this conception of preference makes no necessary reference to the idea of preferences being part of a “mental” apparatus, thus allowing “mental” processes to be modeled as computational processes. Thus, whereas in earlier ordinalism preferences always belonged to individuals, with Arrow’s formal conception preferences in principle can be implemented in any sort of hardware. In terms of Quine’s formal semantics, the formal or syntactical relations Arrow specifies are compatible with a range of different semantic interpretations. This is precisely the position underlying cognitive science: not that cognitive processes cannot be implemented in human beings, but that it is a matter of indifference whether they are. It is true that, alongside Arrow’s formal conception of preference, mainstream economics in this early postwar period also revisited the cardinal conception of preferences which does accommodate preference intensity. The von Neumann– Morgenstern expected utility analysis and Amartya Sen’s social choice analysis (which rejected Arrow’s independence of irrelevant alternatives assumption) both departed from Arrow’s path. In spite of this, it was the Arrow–Debreu general equilibrium model that constituted the dominant research program in mainstream economics until the SMD results. Consider, then, the first of the four themes above, namely, that the mental processes of economic agents are computational. According
After the fall: the machinery of choice
93
to this conception the brain is a digital computer that performs various algorithmic operations on syntactical structures that underlie a “language of thought.” Arrow’s formalist view of preferences, as part of his larger project of creating formal or extensionalist semantics for economics, contributes directly to this conception. The exercise of individual preferences is computational in the sense that there is one formal structure of preferences across all individuals that algorithmically determines how preference information is processed for each and every “individual.” Furthermore, that Arrow’s formal system employs a fully extensionalist language (one that is preference preserving rather than, as with Quine, truth preserving) guarantees that, for any exercise of preferences, there is a well-defined set of operations that produce a determinate set of outcomes. With Arrow, then, we see the foundations of a functionalist view of economic agents in restricting intentional life to a decision process that can be captured as computational processes suitable to any form of intelligence. 5.2.2
Samuelson: “mental” processes as unconscious processes
Samuelson, we saw in Chapter 2, also believed that minimizing the role of psychology in economics held important advantages for the progress of economics as a science, and sought to contribute to this advance by separating out the mathematical theory of consumer behavior from the traditional psychology of utility and Hicksian indifference curve analysis he believed to be primitive and objectionable (Samuelson 1938, 1948, 1950). His 1948 revealed preference theory and weak axiom of revealed preference (WARP) make modest assumptions about consumer utility or preferences, and work solely with the idea of successive choices between bundles of goods to derive the equivalent (negative semidefiniteness of the Slutsky matrix) of the inverse relation between price and quantity known as the law of demand. While it has been shown that Samuelson was not consistent in his attitude towards traditional indifference curve analysis across his different papers on consumer behavior (Wong 1978), he did consistently demonstrate that he thought economics would be better off were it to avoid making reference to the life of the mind. Methodologists of economics (e.g., Hands 2002) have emphasized that Samuelson’s aim was to follow Bridgman (1927) in sketching an operationalist program for economics that would replace theories with representations of their empirical implications. Concepts had to be observable by virtue of the operations used to measure them, and meaningful theorems in economics—Samuelson’s main objective—are refutable hypotheses about empirical data (Samuelson 1947: 4). This view, however, went hand-in-hand with Samuelson’s behaviorism in that, since we cannot observe the workings of the mind or collect data regarding utilities, an economics based on claims about the workings of the mind cannot be expected to produce meaningful theorems. However, being a behaviorist did not imply for Samuelson that mental processes do not exist. First, WARP presumably concerns a mental process because it is about preferences. Second, in the paper in which he introduced the concept of
94
After the fall: the machinery of choice
revealed preference (Samuelson 1948), he backed off his original eliminativist position towards preferences to argue for their reinterpretation in revealed preference behavioral terms (Hands 2002), thus indicating that mental processes underlie (if not explain) preference behavior. What this apparently implies, then, is that WARP concerns an unconscious mental process, namely, one that has none of the notions associated with it—awareness, experience, etc.—that usually distinguish conscious from unconscious mental processes. Samuelson, it should be noted, does not actually say that individuals for whom one bundle of goods is revealed preferred to another are unaware of their preferences. However, consider the character of the revealed preference concept. That individuals find their preferences revealed to them in their choices essentially tells us that they do not know or are unaware of their preferences before making their choices. Preferences, that is, are “revealed” to individuals because they do not know them in advance of their choices. One could try to avoid this interpretation by taking a strong behaviorist position, and by arguing that there is nothing to be revealed over and above the act of revealing itself. That is, preference emerges when revealed. However, not only is this view contrary to the usual sense of the term “reveal,” it also contradicts Samuelson’s 1948 position that treats revealed preference as the best way of understanding our preferences. In addition, it borders on making the explanation of consumer behavior incoherent, since it would remove the anchor against which prices and incomes operate. Goods being revealed preferred consequently still needs to be regarded as a mental process, with the object of that revealing—the individual’s preference—being different from the act of revealing itself. Thus, revealed preference needs to be characterized as an unconscious mental process, or at least one that operates perfectly well without individual awareness. In this respect, Samuelson went further than Arrow in his treatment of the “language of thought.” As with Arrow, individuals for Samuelson are simply preference calculators. However, whereas Arrow narrowed the domain of intentionality by replacing ordinary language expression in economics with extensionalist or formal semantics, Samuelson rather reworked the existing Hicksian preference framework to make it more conformable to formal expression. The upshot of this was that he came up with an alternative understanding of preferences as mental processes—one it turned out that essentially made them unconscious—where Arrow had simply set aside concern over the nature of mental processes. 5.2.3
Friedman: “as if” intentionality
Friedman’s influential essay “The methodology of positive economics” (1953) is a landmark in the development of mainstream economic methodology, and famous for its irrelevance of assumptions thesis. It was written largely as a response to apparent empirical disconfirmations of marginalist theory shown in survey results compiled by Lester (1946, 1947) which implied that business firms do not attempt
After the fall: the machinery of choice
95
to maximize expected returns. Friedman’s main claim was that the test of good theory was not the realism of its assumptions but rather whether a theory produced reliable predictions. Economics, he claimed, only aims at prediction, not at explanation, so the quality of its assumptions was really irrelevant to its appraisal. Theories are merely instruments for making predictions (thus the methodology of instrumentalism), and are not meant to provide understanding or explanations of the behavior they appear to describe. In fact, Friedman originally exaggerated his view by saying that “to be important … a hypothesis must be descriptively false in its assumptions” (Friedman 1953: 14). By this he meant that we should treat economic actors “as if” they were guided perfectly by those motives that would produce the results that the theory successfully predicted, just as expert billiard players made their shots “as if” they knew the complicated mathematical formulas required for doing so. This followed, Friedman argued in a Darwinian vein (Vromen 1995: 27ff.), from the fact that, whatever businessmen believed motivated their actions, the forces of competition ultimately compelled “individual firms behave as if they were seeking rationally to maximize their expected returns” (Friedman 1953: 21). Lester’s survey results, then, were basically beside the point. Friedman’s argument has many problems (cf. Lawson 1992; Mäki 1992), but I will concentrate on how it contributed to a functionalist philosophy of mind in early mainstream economics. First, since the irrelevance of assumptions thesis meant that realistic explanation of individual behavior is unimportant in economics, it also meant that realistic explanation of individual mental processes is unimportant in economics. Second, Friedman’s “as if” view implies that whatever we choose to identify as mental processes—Friedman’s exemplar was maximizing behavior—can be instantiated in any kind of economic agent, human or otherwise. This latter implication follows from the fact that, in likening the forces operating on economic actors to the forces operating on expert billiard players, Friedman made the logic of the system determinative of the logic of the agent. That is, rather than begin with some account of agent behavior, and then go on to ask how economic relationships between agents needed to be explained, Friedman took the system of competitive forces as his starting point, and inferred the type of behavior that economic agents had to have to act in conformity with that system of competitive forces. Thus, maximizing behavior as a mental process is defined as that behavior involving perfect response to changes in prices and incomes, just as expert billiard play is defined as that behavior involving perfect response to the system of physical forces operating in billiards. By this token, then, maximizing behavior could be instantiated in any sort of agent as long as that agent operated successfully in a system of competitive market forces. The idea that mental processes, or agent software, as it were, can be implemented in a variety of different types of hardware is computational functionalism’s multiple realizability idea. Consider household demand for goods and services. If a household utility function is a good predictor of demand, then it is irrelevant whether the household is taken to be a single unit, decomposed into individual family member
96
After the fall: the machinery of choice
demands and then reaggregated, taken to be the demand of the head of the household, or represented in terms of any other conception of where household utilitymaximizing behavior might reside. Any of these hardware specifications is acceptable if it accommodates the maximizing software necessary for perfect response to price and income information. Basically, if a theory produces good predictions, one works backwards from those predictions to postulate a behavior consistent with those predictions “as if” the relevant agent had such a behavior. However, notice that the main idea here is not just that mental processes have multiple realizations. That they do means more importantly that their particular instantiation is a consideration secondary to their understanding. Thus, we ought really to drop the term “mental” when referring to such processes, and rather refer to them as cognitive or functional processes. This broader notion extends to software of all different kinds, with “mental” software as but one subcategory in the whole, indeed one associated primarily with individual human beings. This is essentially the way the well-known functionalist philosopher Dennett explains mind. Dennett holds that when we talk about the cognitive processes of human beings, computers, and thermostats (e.g., Dennett 1987: 29ff.), our humanderived vocabulary of “belief,” “desire,” “hope,” etc., does not actually stand for “mental” phenomena but merely represents our traditional and customary way of speaking when we refer to any cognitive entity’s “intentional” stance towards the world. We can indeed say that the thermostat “perceives” the change in temperature, or that the computer has a “memory,” or even that a person “believes” something, but in each case the use of this language involves a metaphorical way of explaining how the relevant mechanism is oriented towards its object. That is, we do not literally ascribe intentional states to thermostats, computers, or human beings; rather we speak as if they behaved intentionally. Talk about intentional states always involves an as if intentionality. Like Friedman, Dennett is not interested in the realism of our assumptions about the mind, but only about their predictive value. And like Friedman, that these assumptions are serviceable for a variety of types of agents—human, animal, natural, artifact—means for Dennett that we are primarily interested in the “intentional” system itself, irrespective of where it might be instantiated. Dennett’s work antedates Friedman’s, but it has not gone unnoticed in economics. Cottrell thus suggests that mainstream economics might be seen as an applied intentional systems theory along Dennett’s lines (Cottrell 1995: 170). This seems right on the mark in that it has been customary since Friedman’s as if argument to ascribe maximizing behavior to any and all sorts of agents: firms, individuals, government agencies, unions, organizations, clubs, and so on. 5.2.4
Simon: rationality as symbol processing
Simon is an especially interesting and important contributor to the functionalist vision of the individual of early mainstream economics because he was knowledge-
After the fall: the machinery of choice
97
able about cognitive science and a student of artificial intelligence while also explicit in recommending that a cognitive science approach to the analysis of decisionmaking processes be adopted in economics. Equally important is the fact that Simon developed his early cognitivist views by criticizing the rationality postulate of neoclassical economics, so in his own mind a cognitivist view of the individual was desirable precisely because it addressed the deficiencies in the traditional view of the individual. Thus, a brief look at Simon’s thinking gives us a check of sorts on the above cognitive science interpretation of Arrow, Samuelson, and Friedman. If what appears implicit in their views turns out to be explicit in Simon’s work, then this is evidence that the thinking of these figures was part of a broader trend in early mainstream economics. Simon began by arguing, contra traditional choice theory, that human rationality is bounded by our limitations in knowledge and information processing capacities. As a student of psychology and management science, he first investigated decisionmaking processes in organizations, and concluded that individuals generally employ heuristics, or simple rules of thumb, to achieve solutions that could be regarded as satisfactory according to minimal criteria. This “satisficing” behavior enabled individuals to break down complex decision problems into series of subproblems, each of which could also be addressed in terms of minimal criteria for their respective solutions, leading to further subproblem tasks with their respective criteria, etc. (cf. Sent 2000). Thus, whereas in traditional choice theory every single decision involved a relatively unique evaluation of constraints and desires, for Simon decisions fell into categories with associated evaluation rules. This meant that decision-making processes, such as took place in organizations, could be described in a systematic manner, both so as to identify proper procedures within organizational units and so as to account for linkages and connections between units to explain organizations taken as a whole. Standard choice theory was unable to provide explanations of this degree of complexity, and consequently remained reliant on metaphors such as “spontaneous order.” With this understanding of decision making, Simon then turned to computers to simulate rule-driven problem-solving strategies in complex contexts. If one could delineate well-defined sequences of operations, or algorithms, that human beings implemented in large, complex organizations, then these might be modeled with machine information-processing tools, substituting for human organizations a general model of hierarchical systems decomposable into interrelated but independent subsystems (Simon 1969). Moreover, given the capacity of computers, simulating decision-making processes offered clear advantages to human decision makers. Simon, then, adopted a symbol-processing view of artificial intelligence in digital computers, and devoted the greater part of his later work to investigating what this implied about how to apply cognitive psychology to economics, a field that has since come to be known as behavioral economics. The central conclusion in this regard is that individuals do not exercise perfect rationality or indeed anything approaching it, the “most convincing demonstration of [which] has been obtained from simulated markets in which the agents are computer programs that make their
98
After the fall: the machinery of choice
decisions on the basis of very primitive rules that depart widely from the criterion of profit maximization and perfect knowledge” (Simon 1998: 396). Of course there are different ways of looking at what using computers to model human cognition means. Several things, however, stand out. First, for digital computers the “language of thought” is naturally modeled as algorithmic symbol processing. Thus, being able to use computers to address decision-making in complex contexts first requires that one develop something like Quine’s formal semantics or Arrow’s formal/syntactical relations compatible with a range of different semantic interpretations. The direction of development offered by Arrow’s axiomatic approach, then, is one Simon needed to pursue in order to investigate machine information processing. Second, a “language of thought” modeled as algorithmic symbol processing is one meant to generate output rather than exhibit other intentional phenomena such as awareness or experience. That is, symbol-processing systems are computational in nature, solely designed to transfer information efficiently. This, however, is essentially what Samuelson’s revealed preference framework allows, in that it translates preference into something revealed, and therefore determinate and objective. Samuelson was interested in operationalizing consumer behavior, and Simon had a similar goal in simulating decision-making processes in computers. Third, Simon’s simulation strategies show how decision processes can be implemented in different kinds of hardware. This does not imply that Simon thought human thought and machine thought identical. Rather, he proceeded “as if” the former could be simulated in the latter. Thus, he may or may not have shared Friedman’s assumption that the realism of our assumptions is irrelevant in economics. However, he certainly shared with Friedman a strong regard for predictive power in theory. If computer simulations of human cognition generated good predictions about complex decision making, then they were instrumentally valuable. Simon, consequently, drew explicitly on the cognitive science themes that are implicit in Arrow, Samuelson, and Friedman. His work on artificial intelligence, however, influenced the work of Lucas, who, as a proponent of Walrasian general equilibrium theory, also falls within the ambit of early mainstream economics. 5.2.5
Lucas: artificial economies
Boumans (1998) explores the connections between Simon and Lucas in regard to the role the theory of artificial intelligence played in Lucas’s understanding of model building in economics.4 Lucas’s basic view was that a model imitates actual economic behavior by being “a fully articulated artificial economy which behaves through time so as to imitate closely the time series behavior of actual economies” (Lucas 1977: 11). Traditional Keynesian macroeconometric models, Lucas had argued (the Lucas critique), were not good models, because their parameters changed when there were changes in economic policy.5 Thus, Lucas recommended a traditional general equilibrium approach to modeling the economy, which he believed involved “structural
After the fall: the machinery of choice
99
parameters that are invariant under the kinds of policy changes one is interested in,” specifically those describing individual tastes and technology (Lucas 1981: 11–12). In principle, such models would circumvent the Lucas critique problem, because they would be “constructed so as to predict how agents with stable tastes and technology will choose to respond to a new situation” (Lucas 1977: 12). Yet, as Boumans points out, such models exhibit such a high degree of complexity that Lucas’ ambition required considerable optimism regarding “purely technical developments that [would] enlarge our abilities to construct analogue economies” in the future (Lucas 1980: 697). Lucas justified his optimism by citing Simon’s The Sciences of the Artificial (1969) as the “immediate ancestor” of the idea that constructing “a mechanical imitation economy” to model the actual economy was something within the realm of future possibility (Lucas 1980: 697). That is, we could expect to model complex systems in the future by taking advantage of new computational techniques made possible by artificial intelligence systems. At the same time, to take advantage of this expected “advance,” one also had to rethink how one conceptualized economic agents in models. As he later put it in regard to the new kinds of modesl—“artificial worlds”—he foresaw as likely to emerge: “This is what I mean by the ‘mechanics’ of economic development—the construction of a mechanical artificial world, populated by the interacting robots that economics typically studies, that is capable of exhibiting behavior the gross features of which resemble those of the actual world that I have just described” (Lucas 1988: 5). Models, that is, could imitate actual economic behavior if it were assumed that real-world agents understood at the level of tastes and technologies behaved as “interacting robots.” Short of that advance in “purely technical developments” expected to “enlarge our abilities,” however, Lucas rather chose a shortcut on modeling the world’s complexity by adopting a representative agent strategy in which one or a few types of agents stand in for the behavior of the many real-world agents occupying real economies. The idea was that modeling the decision making of many economic agents in terms of the decision making of a single representative agent adequately captured the role of taste and technology parameters in the functioning of the economy. Against this, one might well suppose that actual economies operate far less predictably and deterministically on account of the heterogeneity of individual economic agents. From this perspective, “the representative agent approach [is] fatally flawed because it attempts to impose order on the economy through the concept of an omniscient individual” (Kirman 1992: 132). However, Lucas’s view was that a good economic model involved the “construction of a mechanical artificial world, populated by the interacting robots.” Thus, the differences between individuals were superficial. As he put it, “I think it is exactly this superficiality that gives economics much of the power it has: its ability to predict human behavior without knowing very much about the make-up and the lives of the people whose behavior we are trying to understand” (Lucas 1986: 425). Lucas’s vision of an economy, then, is one of individuals as computational mechanisms or interacting robots and the economy as a mechanical artificial world. Similar
100
After the fall: the machinery of choice
to this is Turing’s conception of the relationship between individual Turing machines and the universal Turing machine. An individual Turing machine is a device that translates input symbols into output symbols through a deterministic sequence of steps defined by a state table. The universal Turing machine can simulate every individual Turing machine by taking the state table of any particular machine and performing the same input–output computations that the individual machine performs. First, then, individual economic agents are like individual Turing machines because they deterministically translate input symbols (prices, incomes, and preferences) into output symbols (demands and supplies). Second, since the representative agent by definition can accomplish all the same computations as the economy’s individual agents, the representative agent is like a universal Turing machine. Third, since the representative agent exhibits rational expectations, and knows the true model of the economy, it is like the universal Turing machine which can simulate not only every individual Turing machine but also itself. Lucas, then, gave particularly sharp expression to Simon’s artificial intelligence based approach to explanation in economics, albeit in a general equilibrium framework, and albeit that he substituted the representative agent strategy for Simon’s modeling of complexity in terms of subsystems. Simon, however, possessed an understanding of individuals as symbol-processing agents that developed associated cognitivist themes in Arrow, Samuelson, and Friedman. Thus, though it would be too strong to say that all of early mainstream economics was exclusively or even consciously devoted to pursuing a cognitivist program for economics, it can be said that much of it bore the imprint of cognitivism. As it turned out, however, the momentum for this reconceptualization of individuals and economics was weakened by the subsequent collapse of general equilibrium theory as the chief project of early mainstream economics.
5.3
After the fall
The break-up of the Walrasian general equilibrium theory program—the centerpiece of early mainstream economics—brought about by the SMD results has stimulated the development of a variety of new mainstream approaches with limited connections to one another. Because of this, I suggest, the abstract individuals conception has failed to emerge as a new dominant conception of the individual, though its influence is still considerable (as indeed is the traditional neoclassical subjectivist conception). Rather, what seems better to describe the current situation is that the fragmentation of mainstream economics into different approaches has generated a variety of ad hoc characterizations of individuals, none of which is either very well developed or seems to have any special advantages over what one finds in other approaches. To take one example, evolutionary game theory treats individuals as rule followers. However, with a number of competing accounts of evolutionary processes, it has yet to produce a theory of the individual, much less demonstrate
After the fall: the machinery of choice
101
that it can accommodate other ideas about individuals found in other mainstream approaches. But, putting aside the effects of fragmentation, it can also be argued that the legacy of atomism itself constitutes a significant obstacle to the emergence of a new theory of the individual in mainstream economics. Thus, a number of recent mainstream approaches (for example, new institutional economics) investigate possible social-institutional influences on individuals, but then stop short of attempting systematically to explain these influences on account of a continuing commitment to the general methodological idea that individuals must always be given the points of entry (epistemologically or ontologically). Indeed, mainstream economics can still be loosely characterized as methodological individualist, in spite of the fact that it no longer possesses a theoretical account of this position (besides the old subjectivist one that continues to have pedagogical value but that its leading contributors no longer accept). Thus, fragmented in direction and still attached to ideas that have more an ideological rather than a scientific function, mainstream economics appears unlikely to offer a new understanding of individuals in the foreseeable future. Where might one begin, then, to look for alternative foundations for a theory of the individual in economics? In light of the discussion above, it seems reasonable first to ask what might be mistaken about a cognitivist philosophy of mind. Here, I draw on Searle as one of the best-known critics of computer functionalism and cognitive science. Searle takes consciousness to be our “central mental notion” (Searle 1992: 84) and argues that being conscious of something is fundamentally different from exhibiting a computational process. A computational process is a sequence of formal or syntactical symbols that is independent of different possible interpretations that can be given to it by virtue of different meanings or content that may be assigned to its symbols. Being conscious of something, however, always involves an awareness of meaning. Moreover, for Searle, awareness of meaning is always from an individual perspective, so that our consciousness always has an “aspectual shape,” and meanings are only partly but not exhaustively characterizable in third person terms (Searle 1992: 157). Computational processes, then, as by definition uninterpreted sequences of formal symbols, lack “aspectual shape” and cannot be conscious mental processes. In Searle’s view we consequently ought not to model our mental states on computational processes. However, might there be two distinct domains of the mind, say, the “computational mind” and the “phenomenological mind,” as some have asserted (e.g., Jackendoff 1987)? There clearly do exist such things as computational processes, as the evidence of digital computers clearly demonstrates, but is there a domain of the human mind constituted solely out of unconscious computational processes, a “silent language of thought,” that in being distinct from the conscious domain must in principle be inaccessible to consciousness? Searle’s view is that all mental states must be at least potentially conscious (Searle 1992: 152ff.). There are indeed many things that we believe or feel but are not aware of, such as that the Eiffel Tower is in Paris or that we might not get somewhere on time. Nonetheless, we could become conscious
102
After the fall: the machinery of choice
of these beliefs and feelings, and thus for Searle these unconscious states should be classified as mental states. Contrast these examples with functions of the brain that are always unconscious, such as the firing of neurons and involuntary responses to various stimuli. One might be tempted to label these unconscious mental states since they are associated with the brain. However, in Searle’s view, since these brain states can never become conscious, it makes no sense to consider them as mental states. Thus, a fully “silent language of thought” is not a language of thought at all, and there is only one domain of the human mind—that describable in the language of consciousness. This does not imply that that the mind does not engage in symbol processing or exercise computational processes. Obviously, human beings exercise logical abilities. However, we can only speak of these abilities as a language of thought when they can be represented in terms of consciousness. That is, since consciousness is always from some perspective or involves an aspectual shape, any computational process we suppose constitutes a language of thought must be interpretable from some perspective or involve an implicit aspectual shape. As Searle put it in connection with his famous Chinese room argument, which showed the inseparability of syntax and semantics or interpretation, syntax is not a physical feature of the world but is always an observer-relative notion that depends upon interpretation coming from outside that syntax (Searle 1980: 208). Applying this idea to the axioms of rational choice theory as a possible language of thought implies that we should thus investigate—as have those recently concerned with what psychology offers economics—how perspective or interpretation influences the operation of the axioms of the theory, such as transitivity.6 However, in what way does all this bear on the theory of the individual in economics? If we suppose that consciousness is always perspectival, then individuals would need to be understood as always occupying some vantage point, that is, a position that creates a perspective or from which things take on an aspectual shape. This is different from the basic idea behind neoclassical subjectivism, namely, that individuals are understood in terms of their internal private states of mind, since there is nothing in the idea of subjectivity as pure internality that gives the individual a location in the world.7 Subjectivity and perspective understood in terms of position and vantage point, however, embed the individual in a context and thus explain individuality in terms of an external framework. This helps motivate the distinction introduced in Chapter 1 between internalist and externalist types of theory of the individual. There, I simply differentiated between two broad approaches, ignoring the role of subjectivity in the latter. However, the more accurate way to look at the matter is to see externalist theories as embedding individual subjectivity in an external context, as, for example, when we consider how preferences can be endogenous rather than exogenous. Of course, early mainstream economics, as discussed above, is not subjectivist. Indeed, its rejection of neoclassical psychology opened the door to the multiple realization of computational “mental” processes in different types of hardware.
After the fall: the machinery of choice
103
However, early mainstream economics still involves an essentially internalist view of the individual in that, when it does claim to be discussing individuals and their “mental” states, the states it attributes to them are unconscious in nature, and therefore lack perspective or aspectual shape. Thus, these abstract individuals are no more in the world in the sense of occupying a position than are neoclassicism’s individuals. In this respect, neoclassicism and early mainstream formalism combine to summarize the history of internalist views of the individual in the atomist tradition of orthodox economics. The verdict both have delivered—neoclassicism in a history of self-critique and the early mainstream in its turn towards computational processes—is that neoclassical subjectivism is a failed project. But so also, I argue, is the brand of cognitivism picked up by Arrow and others. Thus, the second half of this book attempts a new point of entry on the theory of the individual in the form of an explicitly externalist strategy of investigation. If we suppose that thought always involves perspective, and that this implies a position or vantage point that individuals occupy, then we need to begin by examining how individuals can be defined in terms of their being embedded in some context, specifically a social context. The aim of the next chapter is to explicate in broad terms the main ideas involved in this notion of individual embeddedness as a means of presenting a general account of the heterodox conception of the individual.
Part II
Heterodox economics
6 The embedded individual
By the imagination we place ourselves in his situation [our brother … upon the rack], we conceive ourselves enduring all the same torments, we enter as it were into his body, and become in some measure the same person with him … . His agonies, when they are thus brought home to ourselves, when we have thus adopted and made them our own, begin at last to affect us, and we then tremble and shudder at the thought of what he feels. (Smith 1976 [1759]: 9)
The neoclassical conception of the atomistic individual originates in Locke’s subjectivist view that was itself the product of Enlightenment dualism. An alternative tradition of thinking rejects the idea that individuals are defined atomistically apart from their relations to one another, rejects the idea that individuals are defined subjectively in terms of their own mental states, and rejects the idea that the world should be seen dualistically in terms of inner subjective and outer objective realms. Rather, individuals are to be understood and explained in terms of how they are embedded in historical social relationships. However, like the neoclassical conception, this alternative conception has also undergone an evolution, one, I argue here, that was similarly driven by its own problematical aspects. Recall the evolution of the subjectivist conception described in Chapter 2. Since that which is truly subjective cannot be understood in the language of science, the cumulative effect of successive attempts within neoclassical economics to scientifically explain the unique inner worlds of individuals ultimately eroded their characterization as subjective beings. Cognitive science was one outcome of this evolution. Compare, then, how the alternative embedded individual conception involves a dilemma of its own. To say that individuals are embedded in historical social relationships is quite close to saying that individuals disappear into those relationships. Indeed, many would argue that the embedded individual conception is not a conception of individuals at all, but rather a proposal to ignore individuals, in order to focus on groups, classes, movements, historical forces, history, and so on. According to this interpretation, in fact, rather than there being two traditions of thinking about the nature of the individual, there
108
The embedded individual
are really just two great traditions of thinking about society—one that is individualist and includes individuals as agents, and one that is collectivist in which it is not individuals that are agents but instead groups, classes, movements, etc. My view, however, is that, contrary to this interpretation, there are nonetheless two distinct traditions of thinking about the nature of the individual. That the embedded individual conception is subject to collapse in the direction of collectivism (a conclusion actually welcomed by some of its proponents) does not imply that there is no understanding of the individual involved, but rather means that there is something incomplete and problematical about the idea of individuals being socially and historically embedded. That is, just as pure subjectivism is an unsustainable conception, so the idea of pure embedding is an unsustainable conception. Thus, it seems that what deserves our attention is how we are to understand the respective evolutions of these conceptions. The evolution of the subjectivist conception, I argue, ends up eliminating the very basis on which individuals are defined. In other words, an atomistic conception is ultimately self-defeating (something that took the better part of twentieth century neoclassicism to demonstrate). But ironically, the embedded individual view, which many believe eliminates individuals from the outset, appears to exhibit an evolution that increasingly emphasizes individuals as active beings able to act upon and change social frameworks. Hence, my argument in this chapter is that what is problematical in the embedded individual conception, namely, that embedding borders on eliminating individuals, has led contributors to this alternative tradition to explain how embedded individuals are still in fact individuals. Thus, there is a converse symmetry in the respective developments of these two main traditions of thinking about the nature of the individual. In addressing its weakness—pure subjectivism—neoclassicism eliminates its only account of the individual. However, in addressing the weakness in their view—pure embedding—those in the alternative tradition of thinking about individuals have been driven to create a stronger account of the individual. Thus, historically, the defenders of individualism have left the stage (though they are still in the credits), while those on whom suspicions have long fallen as enemies of the individual work largely behind the scenes to defend a realistic place for individuals in an increasingly complex society. This chapter describes the evolution of the embedded individual conception. Section 6.1 discusses both Marx’s thinking about individuals and class membership and Durkheim’s explicit methodological holist position. I suggest that ambiguities in Marx’s thinking when combined with Durkheim’s position enabled methodological individualists to portray Marx and Durkheim as cut from the same cloth, thus saddling them (and all others who favored the embedded individual conception) with the view that individuals are reducible to being the bearers of social identities, and consequently reinforcing the view that since the nineteenth century the two main opposed traditions in social thinking are methodological individualism and methodological collectivism (or holism). In Section 6.2, I argue that one important set of critics of methodological individualism since the 1970s have responded to this
The embedded individual
109
situation by turning to structure–agency models of individual and society, both to escape being characterized as methodological collectivists and because of the inherent problems involved in methodological collectivism. These models treat socially embedded individuals as agents able to act upon and change society, thus answering the standard methodological individualist criticism that socially embedded individuals are simply bearers of social identities. Yet proponents of structure–agent models have devoted only limited attention to the concept of self implied by their treatment of individuals as agents able to act upon and transform social relationships in which they are embedded. In Section 6.3, then, I look at a related and parallel development, also dating from the 1970s, in the form of social psychology’s individual selfconcept. The individual self-concept treats individuals as beings who reflexively distinguish themselves in the process of being acted upon by and acting upon social structures. Section 6.4 then briefly surveys recent heterodox economic contributions to institutionalism, social economics, critical realism, feminism, intersubjectivity theory, and reflected rationality theory in terms of how they combine the logic of the structure–agent model with the individual self-concept to provide new views of the individual as socially embedded. Finally, Section 6.5 summarizes the chapter’s discussion and outlines the argument of the following two chapters.
6.1
The legacy of Marx and Durkheim
Marx’s view of the individual is generally associated with his theory of history, known (but not so labeled by Marx) as historical materialism. There are two principal ideas in the theory (cf. Reuten 2003). The first is that a society’s “base,” namely, the character of its production and associated economic relationships, explains the development of its “superstructure” made up of all political, legal, and cultural relationships. The second is that historically all societies have possessed two main classes: one that supports both classes through its work, and a second that depends upon and directs the labor of the first. According to this view, then, the relationship between classes determines the pathway history takes. This suggests that individuals do not have an important role in history, or, to the extent that they do, only as bearers of class identities. Put in terms of the concept of agency and the question of where power to bring about events resides, classes are historical agents, but individuals are subsumed in classes, and accordingly are not independent agents themselves. This familiar interpretation, however, needs to be scrutinized, since it is conceivable that history is best explained in terms of forces that operate above the level of the individual, and yet that individuals are still independent agents. Whether this is the case depends upon how one understands what it means to be a member of a class. Thus, a case can be made for saying that there are basically two different ways of understanding Marx on the subject of class membership. According to the view suggested above that treats individuals as simply the bearers of class identities, individuals and their behavior are constituted for them and determined by their membership
110
The embedded individual
in classes. A second view is that individuals are members of classes, because people in similar circumstances act in similar ways, so that similarity in circumstances defines class. By this token, class location constrains and limits individual nature and behavior but does not determine it. Individuals occupy positions, and this biases their behavior but does not constitute it. Individuals may consequently still be thought to be independent agents, but the scope of independent action is highly circumscribed. We might thus say that it is contingently the case that individuals act as bearers of class identities, whereas in the first view individuals necessarily act as they have been determined to do by their class identities. Many commentators on Marx seem to interpret Marx’s view of class membership in the determinative sense which reduces individuals to bearers of class identities. This is certainly understandable in light of the longstanding popular conviction that individualism and collectivism are the two great traditions in thinking about the nature of society. However, the nonreductionist “positions” view of membership is probably the more reasonable interpretation of Marx’s own thinking, particularly if we consider his early Paris writings on alienation. The simple idea behind alienation is that the individual activity, sociality, and nature are distorted (“estranged” in the original German) by labor’s being treated as a commodity. But if this is the case, then individuals cannot be reduced to their class identities, since that would preclude there being a tension between being in a certain position and being alienated in it. It is true that some modern Marxists believe that Marx abandoned his early alienation thinking (e.g., Althusser 1969). But there is also evidence in his later work that Marx adopted the more flexible view, for example, in his 1852 “The eighteenth Brumaire of Louis Bonaparte” where he states, “Men make their own history, but not of their own free will; not under circumstances they themselves have chosen but under the given and inherited circumstances with which they are directly confronted” (Marx 1973 [1852]). Moreover, whether or not Marx gave up some of his early views does not change the fact that one can argue that a nonreductionist, “positions” view of individuals and classes is a possible interpretation of his ideas. In any event, Marx is often seen as the principal fount of a tradition in thinking beginning in the nineteenth century that treats individuals as being determined by society and social relationships. The work of Durkheim, writing later at the turn of the century, fits this view quite well, and thus helps create the impression of a single tradition in ideas. Durkheim criticized the classical economists for what he perceived to be their individualist orientation, and argued that human beings are social animals whose nature is created by society. Methodologically, he began by arguing that society is ontologically an autonomous entity, which implies that a scientific sociology needs to concern itself with social phenomena external to individuals: “The group thinks, feels, and acts quite differently from the way in which its members would were they isolated. If, then, we begin with the individual, we shall be able to understand nothing of what takes place in the group” (Durkheim 1938 [1901]: 103). This view, methodological holism, was subsequently posed as the alternative to
The embedded individual
111
methodological individualism, most influentially by Popper, who gave it a long prehistory sweeping up Plato and Marx and a more modern posthistory extending through twentieth century totalitarianism (Popper 1945). The antithetical character of methodological holism and methodological individualism and Durkheim’s explicit defense of holism has helped cement his view of individuals as the understanding of an entire tradition. Arguably, this situation has created considerable ambivalence in methodological individualism’s later opponents, who, in rejecting atomism as naive, emphasize the social influences that operate on individuals, only to then see themselves as branded with the view that individuals are determined by social forces. Denials that this equally naive view is not intended are generally either ignored or dismissed in a climate that understands the history of thinking about society in terms of a polar opposition between methodological individualism and methodological holism. Indeed, the difference between saying individuals are influenced or determined by society, though philosophically sharp, is still a subtle one within social explanation itself. To claim that individuals are constrained and limited by their social positions, and accordingly still exercise choice and discretion in some respects, is hardly a strong answer to the methodological individualist who equates this answer with the proposition that, basically, individuals are constituted in nature and behavior by the social relationships they occupy. It should be added that there is considerable disagreement among self-declared methodological holists over the extent of individual discretion and freedom in social relationships. The upshot from all this, I suggest, is that those who reject atomism have largely given up on trying to persuade the critics of holism that class membership does not imply that individuals are constituted by society. Rather, they now increasingly argue that individuals, who as socially embedded beings are acted upon by society, are also themselves agents who act upon and change society. Note, then, that supposing individuals are also agents who act upon and change society makes irrelevant the question of whether individuals are determined by society. It also renders irrelevant the old methodological atomism/methodological holism debate, which was largely a debate over where agency resides in society. Finally, supposing that embedded individuals also act upon and change society brings to the fore the question of the nature of the individual, supplanting the old debate between methodological individualism and methodological holism with one between two rival views of the individual, the atomistic and socially embedded.
6.2 Social theory’s structure–agency models of the individual and society Social theory is a large multidisciplinary research strategy that broadly investigates power, social organization, agency, rationality, identity, culture, modernism/ postmodernism, technology, and politics. It draws from many of the social sciences (sociology, anthropology, political science, geography), history, critical studies,
112
The embedded individual
philosophy, and feminist studies, representatives of which are organized in the International Society Theory Consortium founded in 1998. In spite of its diffuse and many-sided character, those in social theory appear generally to share a rejection of reductionist reasoning such as is involved in the methodological individualism/methodological holism debate, and a commitment to structure–agency frameworks for investigating relationships between individuals and society. From this perspective, Archer provides an excellent history of the postwar debate and stalemate between methodological individualism and methodological holism which culminated in a new set of arguments by the late 1970s and early 1980s regarding the mutual effects of individuals and society—or agency and structure—upon one another (Archer 1995: 31– 64). An early leading view in this latter regard was Giddens’ “structuration” theory. Giddens’ “structuration” theory treats individuals and social structures as interdependent and “inseparable,” or as a “duality of structure,” in that each is understood to constitute and determine the other through recursive social practices (Giddens 1976, 1984).1 As one commentator puts it, structure and agency, “instead of separate and opposing things in the world or as mutually exclusive ways of thinking about the world … are simply two sides of the same coin. If we look at social practices in one way, we can see actors and actions; if we look at them in another way we can see structures” (Craib 1992: 3–4). For Giddens, in their ordinary everyday experience, individuals continually give meaning to their lives and the world around them, and yet at the same time find their actions are structured by how meanings generated in the past become routine and established. This process is constantly ongoing, and this leads Giddens to reject the concept of “social roles” as unnecessarily static in favor of the idea of a continuous “positioning” of meanings through social practices in which individuals produce and reproduce social meanings. Social practices thus understood are the basis for institutions, which are accordingly characterized as “sedimented” or regularized practices. Institutions, that is, are not something fixed and permanent, but rather fluid and evolving structures tied to social practices. Bhaskar and Archer subsequently criticized and revised Giddens’ thinking, arguing that reality is stratified and multilayered with emergent properties that differentiate one layer or level from another in historical time (cf. Hodgson 2002). One of their principal complaints against Giddens’ emphasis on social practice as a continuous process is that it seems to rule out historically emergent properties. Put in terms of the idea that structure is “inseparable” from agency, the charge is that one still needs to be able analytically to differentiate structure and agency in order to avoid seeing them as “sinking” into one another (Archer 1995: 65). Thus, one needs to “distinguish sharply … between the genesis of human actions, lying in the reasons and plans of human beings … and the structures governing the reproduction and transformation of social activities” in order to examine the interplay between them over time (Bhaskar 1989 [1979]: 79). This interplay over time reveals a temporal asymmetry of agent and structure. Individuals are born into a world of pre-existing institutions, and institutions evolve as a consequence of the past actions of individuals. Moreover,
The embedded individual
113
seeing structure and agency in interaction, rather than as “inseparable,” makes the concept of causality central to the investigation of structure–agent models. Thus, for Bhaskar and Archer, structure and agency are mutually constitutive of one another, though the way in which this is understood has important implications for our understanding of how the two influence and affect one another. How does this all affect our understanding of the concept of the individual? In a later work, Modernity and Self-identity, Giddens argued in an explicitly anti-Cartesian manner that the self is fully constituted in social practices (society is also constituted in these same social practices), and cannot consequently be explained either in terms of individual psychological characteristics or in terms of any “transcendental philosophy of the ego [that] terminates in an irremediable solipsism” (Giddens 1991: 51). This “oversocial view” of the person, as he favorably termed it, is a consequence of his duality approach, and is based on what Giddens takes to be a Wittgensteinian view that the “I/me” concept is internal to language. That is, since language is itself a set of social practices, if the “I/me” concept is entirely internal to language, it can have no origin other than in social practice. However, this implies that all an individual’s experiences are socially mediated, and recalls Durkheim’s holist claim that individuals are fully subsumed within social groups. It also essentially gives up Giddens’ earlier view that individuals are agents who influence social structure, since it builds the concept of the person entirely on the “I/me” language phenomenon, thereby failing to include in the concept of the person the agency idea that individuals possess a capacity to act upon and influence social structures, including language. A better way of seeing individuals in a structure–agency framework, then, arguably comes in the form of Granovetter’s influential two-sided characterization of socially embedded individuals as neither “atoms outside a social context” nor beings who “adhere slavishly to a script written for them by the particular intersection of social categories that they happen to occupy” (Granovetter 1985: 487). Granovetter distances himself from both the “oversocialized” and “undersocialized” views of the individual, and invites us to investigate how the individual is both influenced by and helps influence social structures. However, the question that all contributors to this more recent line of investigation confront is how to capture the latter half of this proposition. As the thinking of both Durkheim and Giddens suggests, ease in developing explanations of how individuals are acted upon by social structures is matched by a difficulty of the same magnitude in developing explanations of how they might act upon social structures. The key to such an explanation is being able somehow to account for how individuals possess agency or a power to act upon their social environment, since this would directly express the idea that individuals also act upon social structures. However, the problem with simply attributing agency to individuals is that it involves postulating a power that has no apparent basis in their characterization as socially embedded beings. Like the old, equally unsatisfactory stratagem of assuming that individuals possess free will, simply attributing agency to individuals introduces a deus ex machina, whose motivation (but not justification) is to find an
114
The embedded individual
escape from the dilemma of seeing socially embedded individuals acted upon but not obviously themselves acting upon social structures. This problem, nonetheless, can be solved by coupling the concept of the individual as an agent with the concept of the individual as a reflexive being. The latter concept of the individual has a long and even ancient history in literature and philosophy, but has recently been investigated in specifically social science terms in the interdisciplinary field of social psychology in connection with the question of how individuals develop “self-concepts” when engaging in “self-referent behavior.” I briefly survey this literature in the following section. However, its importance in the context of structure–agent models is that, by describing how individuals actively form selfconceptions—precisely because of social influences operating upon them—it succeeds in introducing agency into the conception of the individual as socially embedded in a nonarbitrary manner. In effect, the argument is that it is only when individuals form self-conceptions that they are capable of organizing the social influences operating upon them, rather than simply serving as passive repositories of those influences. The argument allows (as it must) that individuals may also fail actively to form selfconceptions (or have their self-conceptions determined for them by society). However, that individuals may manifest this behavior, as the social psychology literature can reasonably be said to demonstrate, justifies our regarding socially embedded individuals as capable in principle of acting on social structures as they themselves are acted upon by them. That is, it justifies our seeing individuals as agents in a structure–agent framework, when regarded as reflexive beings.
6.3 Social psychology’s self-referent behavior and the individual self-concept Kaplan defines self-referent behavior as a class of social behaviors in which people have themselves rather than others as the objects of their behavior (Kaplan1986: vii). Individuals are thus both subject and object, and “[t]he structure of self-referent behavior is such that a person can be conceptualized as two separate individuals— one who acts and one who reacts to the behavior of the actor” (Kaplan 1986: 2). Here, the terminology of “act” and “react” is used differently from above, with the “one who reacts” being the subject who examines or evaluates the behavior of the “one who acts” as the object. For example, “[a] person anticipates how he will behave in a given situation and responds to this self-perception with a particular selffeeling, which in turn stimulates further self-awareness” (Kaplan 1986: 2).2 Despite the difference in terminology, Kaplan’s aim is to explain how in self-referent behavior individuals examine their own behavior, and, in forming judgments about it, create and re-create their self-conceptions. Note the difference this involves from simply saying that individuals’ self-conceptions are passively absorbed from society. Suppose that society continually imposes views on individuals about themselves. That individuals take these views and themselves as objects means that they act upon
The embedded individual
115
these externally imposed self-conceptions—evaluating, investigating, and judging them—and thereby transform them. The self-conceptions that emerge from this self-referent behavior must thus differ from the ones externally imposed by society. Consequently, when individuals act in ways that affect society, they not do so simply as earlier products of society, as an echo, as it were, of the prior effects of society upon them, but rather as self-changed individuals whose actions can then in principle affect and change society.3 Self-referent behavior and the topic of the individual self-concept, as it is generally termed in the social psychology literature (Rosenberg 1979; Gecas 1982; Demos 1992), dates from the late 1970s, and reflects the nature of social psychology as an interstitial social science jointly created by sociologists and psychologists. The literature, perhaps not surprisingly, seems to have had greater practical impact in psychiatry and mental health than theoretical impact in the social sciences. Nonetheless, the self-concept has been said to be “rapidly becoming the dominant concern within social psychology … as part of the general intellectual shift from behavioral to cognitive and phenomenological orientations” in both sociology and psychology (Gecas 1996). Those with a more sociological orientation within social psychology usually concentrate on the social antecedents of individual behavior, whereas those with a more psychological orientation within social psychology rather tend to concentrate on the social consequences of individual motivation and behavior. Social psychological research on the self-concept as a whole, however, combines these two orientations by generally investigating the multiple ways in which the individual is reflexively constituted and reconstituted out of a socially influenced process of self-evaluation, self-imagining, and self-awareness engaged in by the individual. The self-concept might thus be thought of as a concept of individual identity understood as “situationally variable,” yet also relatively stable, as suggested by the metaphor of a “moving baseline” subject to fluctuation and change (Demos 1985). The concept also reflects social psychology’s comparatively more dialectical mode of investigation in comparison with the more unidirectionally oriented dynamics of sociology and psychology taken alone, which examine either the impact of social structure on the individual or the impact of the individual on social structure. If the self-concept in the most general sense is the understanding that individuals acquire of themselves as a result of the influence of social structures upon their processes of self-appraisal (Rosenberg 1979), then it can also be investigated (a) in terms of self-appraisal processes in which individuals think of themselves as a subject, an “I,” and (b) in terms of self-appraisal processes in which individuals see themselves as an object for others, a “me.” Social psychologists with the former, more psychological orientation explain the self-concept in terms of individual desire to establish a sense of self, using the term “self-esteem” to emphasize “the evaluative and emotional dimensions of the self-concept.” Social psychologists with the latter, more sociological orientation use the concept of individual identity to “focus on the meanings comprising the self as an object” within a social context (cf. Gecas 1982: 4).
116
The embedded individual
The former emphasis on self-esteem makes the individual activity or agency that produces a self-concept central, whereas the latter emphasis on identity gives greater weight to society’s influence on individual self-evaluation. The distance between these two types of conception creates a basis for describing the creation and recreation of individual self-concepts in society as a potentially conflictual one, thus allowing for incorporation of additional theoretical frameworks such as power and control in society. Both orientations, however, assume that how others judge an individual strongly influences the formation of the individual’s self-concept (whether in the self-esteem sense or in the identity sense), and use this to argue that the self-concept involves a sort of “reflected” self-appraisal. At the same time, it is also generally agreed that the concept of “reflected” self-appraisal still lacks good empirical foundations, since, while it is plausible to say that people’s self-concepts reflect the way they believe others see them, there nonetheless seems to be little evidence supporting close correspondence between people’s self-concepts and the way people are actually seen by others. The problem is that it is not obvious just how people’s “self-evaluations are influenced by the feedback received by others” (Shrauger and Schoeneman 1979). Possible reasons for this have been suggested for both the self-esteem and identity understandings of the concept. The way in which individuals establish selfesteem may be selective with respect to good and bad opinions of others as well as with respect to whose opinions are observed. Alternatively, the way in which individuals establish an identity concept may reflect norms of social interaction in a culture, whereby customary practices condition the form of appraisal from others and create differences between how others perceive an individual and express their views of that individual (Goffman 1959). Thus, in addition to the basic psychological and sociological types of approaches to the self-concept itself, there are also distinct psychology-based and sociology-based strategies for explaining the ways in which others’ opinions feed back upon and affect individuals’ self-appraisal.4 For the purposes here, one especially important aspect of more recent social psychology literature on the self-concept is its shift towards greater emphasis on the dynamic effects of interaction between the individual and society or agency and structure. Thus, while the earlier social psychological literature on the self-concept tended to emphasize constancy and stability in individuals’ self-concepts, in order to develop a coherent conception of a “maintained” self, there now appears to be increasing agreement that emphasis should also be placed upon a “plasticity of characteristics previously assumed to be stable throughout the life course,” implying a more “‘open’ or ‘unfinished’ character of the human organism in relation to its environment” (Dannefer 1984: 107; cf. Demos 1992). For example, one such view widely influential in the social sciences and health fields is of the self-concept as an organized succession of different types of age, such as “developmental” age as a set of meanings associated with personal passage from youth through old age, “social” age as the social meanings of age at various transitions and turning points in social life,
The embedded individual
117
and “historical” age as meanings associated with the experience of different cohorts and generations in history (Elder 1991). However, there are many other ways in which the plasticity of the individual could be understood when taking the idea of the individual as a sort of “moving baseline.” Thus, in principle the social psychology self-concept should be amenable to employment in economics frameworks that emphasize dynamic, mutually constituting relationships between individual and society, and that thereby treat the nature of individuals as largely endogenous to the economic process. Let me emphasize, however, that applying the self-concept to structure–agent models to improve the socially embedded individual conception is not a matter of taking over any specific content on the subject from current social psychological research. Rather, what this literature offers social theory and economics is an openended account of individuals as reflexive (and therefore active) beings. The idea, that is, is to incorporate a structure of reflexivity into our understanding of socially embedded individuals as is implied by the general understanding of self-referent behavior. Selfreferent behavior, or the idea that individuals can take themselves as objects (just as they take others as objects), obviously assumes a variety of possible forms. The question for economics, then, is what forms are important in economic life, and in what ways are they important?
6.4 The socially embedded individual conception in economics Though there are mainstream economics conceptions of individuals in social-institutional settings, they nonetheless all assume that individuals are first to be defined atomistically or in an internalist manner. In contrast, most heterodox accounts of the individual begin by defining individuals as socially embedded or in an externalist manner. For example, as Polanyi put it, economic agents and their actions are “embedded and enmeshed in a variety of institutions” (Polanyi 1968: 217). Or as Tool states, “The central contention here is that the particular attitudes, motivations, beliefs, and behaviors which human beings manifest are largely determined by the character of their interaction, conformist and non-conformist, with the intellectual, cultural, social and physical environmental matrix to which they are exposed and from which they emerge” (Tool 1979: 54). In this section, I briefly discuss a number of recent representative accounts of the socially embedded individual from institutionalism, social economics, critical realism, feminism, intersubjectivist economics, and reflected rationality theory. The common denominator of all but the last is that each can be interpreted as providing the general outline of a conception of the socially embedded individual based on a general structure–agent framework incorporating some understanding of individual reflexivity. The last account lacks the structure–agent framework but offers valuable insights on individual reflexivity as a form of social behavior.
118 6.4.1
The embedded individual Institutionalism
One effect of the emergence of new institutional economics in the 1980s was to throw the contrasting strategies of old and new institutional economics into sharp relief. Rutherford’s examination of the similarities and differences between the old and new institutionalism frames these contrasting strategies in terms of the debate between methodological holism and methodological individualism (Rutherford 1994). In their more extreme versions, methodological holism is the functionalist idea that “the behavior of individuals is to be explained in terms of social or historical laws, forces, functions, or purposes” that are not the intended or unintended outcome of individuals’ decisions and actions, while methodological individualism is the reductionist idea that “any fully adequate theory of social phenomena must run solely in individualistic terms” (Rutherford 1994: 30, 36). However, more moderate versions, Rutherford argues, exist on both sides of this debate, and holism and individualism “can and should be combined into a middle way,” which he associates with Agassi’s “institutional individualism.” Indeed, “there are large bodies of work within both traditions,” he points out, “that take institutions and institutional change as the outcome of the actions of individuals, while, at the same time, maintaining that the existing institutions constrain and affect the situations and goals of individuals” (Rutherford 1994: 50; emphasis added). Thus, though there are indeed strong holist positions within old institutionalism, it is better to characterize old institutionalism as having a particular view of the individual than as a theory of autonomous social forces. Hodgson expresses this in terms of the idea of the “institutionalized individual” (Hodgson 2000). Mayhew refers to “enculturated individuals” (Mayhew 2000). Though they agree with Rutherford that strong holist views can be found among many of the best-known contributors to old institutionalism, a closer reading of their work shows that these views were often delivered with the aim of contesting the naive view that the individual is given in economics. Thus, Hodgson asserts: “Instead of a bedrock of given individuals, presumed by the mainstream and new institutional economics, the old institutionalism holds to the idea of interactive and partially malleable agents, mutually entwined in a web of partially durable and self-reinforcing institutions” (Hodgson 2000: 325; cf. Hodgson 1988). Or, as Mayhew puts it, “It is obvious that culture is necessarily a creation of people and that this is so even if we also accept that people are creations of their culture” (Mayhew 1987: 590). How, then, is one to proceed in understanding individuals and society if the polar opposites, methodological individualism and holism, are rejected? Hodgson argues, and Mayhew seconds the idea, that the key in grasping “both the dependence of institutions upon individuals and the molding of individuals by institutions” is to examine both “upward and downward causation” processes (Hodgson 2000: 326; also cf. Hodgson 2003; Mayhew 2000: 331). When individuals “create and change institutions,” this is a matter of upward causation; when “institutions mold and constrain individuals,” this is a matter of downward causation” (Hodgson 2000: 326).
The embedded individual
119
One important implication of accommodating “reconstitutive downward causation” is that, without it, it is difficult to explain learning. Neoclassical economics “treats learning as the cumulative discovery of pre-existing ‘blueprint’ information, as stimulus and response, or as the Bayesian updating of subjective probability estimates in the light of incoming data” (Hodgson 2000: 327). However, learning takes place in institutions and social structures and involves individual development and continuous adaptation. Hodgson does not explicitly treat this as a reflexive process, but it is not difficult to see how it can be so understood, particularly if learning is understood as a recursive social practice that is modified over time. Thus, individuals need not only understand how to participate in social practices but also understand how to do so when those practices are modified over time (Dolfsma 2002b). One way in which this response to change can be explained is in terms of how individuals see their place in a given social practice. Change, then, in the way that social practice operates requires that individuals re-estimate their place in that social practice. However, this requires that individuals have a view of themselves or a self-concept in that practice, and also that they are able to see how that view or self-concept is altered by change in that practice. The ability to form and re-form individual selfconcepts in social circumstances in this way is what was characterized above as selfreferent behavior. This learning-based view of individuals, linked to self-referent behavior, can also form the basis of “an evolutionary view of individuals and human agency.” Hodgson emphasizes the importance of an evolutionary view of individuals and agency incorporating developments from evolutionary psychology (Hodgson 1999; cf. Bunge 1980). In this framework, human psychology evolves through change in both habits and methods of conscious reasoning. However, learning clearly facilitates this general process and also constitutes one of its products. Taking self-referent behavior, then, as not only a form of social behavior but also one with a particular role for learning would seem to allow us better to understand change and development in human psychology. In fact, an earlier tradition within old institutionalism emphasized reflexivity for precisely these reasons. At the beginning of the last century, Cooley’s “looking-glass self” was constructed around the idea that how we see and judge ourselves depends upon how we believe we appear to others (Cooley 1902: 179–85). Mead subsequently described how mind and self are the product of a social process, and asserted that selfconsciousness is fundamentally social (Mead 1934; cf. Albert and Ramstad 1998). Both saw themselves as opponents of the traditional Cartesian disjunction between mind and world. Thus, there is good precedent for developing evolutionary institutionalist views of individuals and society that both operate in a general structure– agent framework and employ reflexive views of individuals.
120 6.4.2
The embedded individual Social economics
Social economics, as its name indicates, begins from a social perspective, and substitutes the idea of the individual as Homo socio-economicus for the idea of the individual as Homo economicus (O’Boyle 1994). The latter notion, Lutz emphasizes, is best thought of as referring to “asocial” individuals, whereas the former notion refers to “persons as social individuals [who] are embedded in a web of constitutive social relations” (Lutz 1999: 6). These “constitutive social relations” can be understood to operate on a number of different levels, with some social relationships encompassing and encapsulating others. In particular, noneconomic community, family, and wider social relationships, in which social values are fundamental to explaining the interaction between individuals, encompass and encapsulate economic relationships, while the economy in the broadest sense, including production, consumption, and distribution, encompasses and encapsulates the market economy, understood as the domain of exchange. Social economics, then, is “anti-imperialistic” in comparison with neoclassical economics in that it reverses the latter’s hierarchical social order by explaining economic relationships in terms of social relationships. The idea that individuals are “embedded in a web of constitutive social relations” is also susceptible to a structure–agent interpretation. On the one hand, that social relationships are prior to and more fundamental than market relationships implies that society cannot be seen as simply the aggregate of all individuals. As Lutz argues, treating society as merely a collection of individuals ignores the separate contribution made by social interaction and commits the fallacy of composition (Lutz 1999: 7). One particularly important consequence of this error is that it limits attention to those social values applicable to the market alone—particularly the gains-from-trade efficiency value—and prevents examination of that wider set of values—justice, equity, human rights, the common good, and so on—relevant to our understanding of society as a whole. On the other hand, though social economists’ broader focus on social relationships that encompass and encapsulate the economy and the market shifts attention to those social values associated with society as a whole, many nonetheless still regard individuals as having an intrinsic value and being centrally important. Indeed, for many social economists the explanation of the individual is a starting point and primary objective (Danner 2002; O’Boyle 1998). The difference from neoclassical and mainstream economics is that, since the individual is seen from the outset as Homo socio-economicus, the individual is always seen as a subject of the wider array of values relevant to society as a whole. In this respect, it is interesting that one influential view of the individual in social economics—the dual-self view—employs an explicitly reflexive view of the individual. The individual is understood to possess both first-order preferences, such as are associated with ordinary wants and desires, and which are often socially constituted, and also second-order preferences, or metapreferences, which have firstorder preferences as their object (Lutz 1999; George 2001). If individuals are then
The embedded individual
121
identified with their metapreferences, they can then be understood to be continuously engaged in a process of self-evaluation whereby they inspect and judge their first-order preferences. Lutz is explicit about adding this idea that individuals actively evaluate their first-order preferences, characterizes this as a process of “self-determination,” and considers it to be “at the heart of dual-self conception of the active self” (Lutz 1999: 158). However, that individuals are influenced by social structures, and in particular by the market, means that they can also be influenced in the way that they engage in self-evaluation. George thus argues that markets promote firstorder preferences, discourage the exercise of second-order metapreferences, and leads to the development of “undesired desires” (George 2001). Thus, while individuals act upon social structure when they evaluate their socially constituted first-order preferences, they are also acted upon by society when the market limits and constrains that evaluative capacity. 6.4.3
Critical realism
Lawson explicitly develops a critical realist conception of the agency–structure model in the third part of Economics and Reality (1997). His agency–structure model is specifically critical realist in that, as a realist model for a social science, the model makes “social structure dependent upon human agency … open to transformation through changing human practices which in turn can be affected by criticising the conceptions and understandings on which people act” (Lawson 1997: 158). That is, social structure changes because human practices change as a result of people critically evaluating the ideas that these practices reflect. At the same time, this does not imply that human agency is an unmediated force acting on social structure free of all social influence. First, “social structures (rules, relations, positions)” are real in the sense of being irreducible to the actions of individuals (Lawson 1997: 161), and second, social structure is a precondition for individual action in that individuals draw upon it when acting. Thus, in place of traditional rationality theory, Lawson substitutes a theory of situated rationality whereby “[a]t any given point in time any individual is situated in a range of positions” associated with which are “rules to draw upon, obligations to fulfil, structures of power to utilise and be influenced by” (Lawson 1997: 187). These positions, and the rules, obligations, and structures of power upon which they depend, constitute the social preconditions for individual action that help explain how human agency depends upon social structure.5 Thus, human agency depends upon social structure just as social structure depends upon agency. It accordingly avoids the (twin but opposite) errors of voluntarism and reification/determinism, where voluntarism explains social structures as essentially the creation of individuals (as in some methodological individualist accounts), and reification/determinism sees individuals as essentially the product of social structures (as in some methodological collectivist accounts). The way in which the model achieves this, Lawson argues, is by explaining the complex character of
122
The embedded individual
human agency as intentional activity. Two points are emphasized. First, “most structural reproduction and/or transformation arises as the unintended product [of human intentional activity], whether or not desired or even recognized” (Lawson 1997: 169). Though human agency creates social structures, intentional activity cannot fully explain the evolution of social structures (the error of voluntarism), because its results are in significant degree unintended. Second: If the reproduction/transformation of social structure is rarely an intended project, it is equally the case that the individual agents are not always aware, certainly not discursively or self-consciously so, of the structures (such as language rules) upon which they are drawing. (Lawson 1997: 169)
This reinforces the idea that, though social structures influence human agency, they do not determine intentional human activity (the error of reification/determinism). Yet if this account depends upon treating human agency as intentional activity, what are we to say about the pervasiveness of highly routinized forms of activity in human society? Lawson allows that much of human behavior does not take on the aspect of being intentional. If we distinguish between discursive and tacit consciousness, where the former involves conscious reflection on a particular subject, then it seems obvious that individuals often rely upon a body of tacit knowledge and skills in order to do such things as follow rules, observe norms, and act in accordance with conventions without reflecting directly upon doing so. However, such activity is still purposeful, potentially conscious, and thus also fairly treated as a form of intentional activity. Were the individual to be interrupted in the performance of an episode of rule following, typically a reason for following the rule could be given. This verdict is important not just because routine is such an important part of human behavior, but because understanding routinized behavior as a form of intentional activity is important for developing a more complete understanding of the individual and human subjectivity in the agency–structure model. Generally, Lawson’s discussion of the processes of reproduction and transformation in the agency–structure model tends to associate these processes with the evolution of social structures. However, to be complete, the processes of reproduction and transformation of agency and social structure should also be associated with the evolution of human subjectivity, the agency side of the model. Here, his discussion of routines as a form of intentional activity opens the door to one way to go about this. Why, he asks, is it that so much human activity is routinized. Part of the answer lies in the simple fact that we seem unable to apply discursive decision-making to all occasions in life where it might be possible to do. However, if this is necessary to explain routines, it cannot be sufficient, since other responses to this inability such as inactivity are also available. A sufficient explanation that Lawson offers is that this human inability causes individuals anxiety and stress, and that individuals accordingly rely on
The embedded individual
123
routinized behavior to reduce anxiety and stress. Then the fuller account of the agency–structure model that includes an explanation of the evolution of human subjectivity would involve describing how routines evolve together with individuals’ evolving strategies for combating anxiety and stress. Yet there is no great difference between this and the social psychology idea of the individual self-concept. Indeed, Lawson adds that individuals may be thought to have a need for “inner security” which in turn may reflect a need for “a significant degree of continuity, stability and sameness in daily affairs” (Lawson 1997: 180). One way of understanding continuity and stability in life is in terms of how it promotes continuity and stability in an individual’s self-concept. Recalling the “moving baseline” expression, individuals might be said to rely on routinized activity because doing so promotes continuity and stability in their self-concepts. Lawson in fact suggests just such an argument when (following Laing and Giddens) he treats a need for inner security as a need for an “ontological security,” where the “psychological origins of ontological security are to be found in basic anxiety-controlling mechanisms … hierarchically ordered as components of personality” (Giddens 1984: 50; quoted in Lawson 1997: 183). Ontological security, that is, applies to a subject, or to the individual. Thus, the processes of reproduction and transformation that are explicitly associated with the evolution of social structures in Lawson’s agency–structure model appear equally applicable to the evolution of human subjectivity that can be described in social psychology self-concept terms. 6.4.4
Feminism
A dominant theme in much feminist work is the social construction of group identities, and how this often generates an identity/difference “politics of recognition” (Taylor 1992; Benhahib 1996; Fraser 1997). A problem with the “politics of recognition” is that it is often “officially recognized, sanctified, legitimized, and accepted by the state and its institutions” to create manageable “corporate identities” (Benhahib 1999: 298). These “corporate identities” conceal internal heterogeneity within groups, and also minimize individuals’ multiple cross-group associations. Thus, while individuals’ identification with groups is an important element in determining individual identity, it can also be subject to manipulation and misunderstanding in ways that are disruptive for individuals. One response to this on the part of feminists is to pursue ways in which women can recognize and act on shared interests. Folbre characterizes this general problem as a “structures of constraint” problem in which “individuals are so embedded in a complex structure of individual and collective identities and competing interpretations of these that sometimes they do not even know whose interests they are acting on” (Folbre 1994: 16). Folbre’s framework is neither methodological individualist nor methodological holist. In her earlier work she rejected a number of traditional Marxist ideas regarding household relationships and class consciousness, thus distancing herself
124
The embedded individual
from methodological holism (Folbre 1982). However, in seeing the multiple and often conflicting group associations that women (as well as men) face as “structures of constraint,” she goes on to employ a structure–agent type of model in which both structures and agents have effects on one another: “Choice takes place within certain social structures, themselves the outcome of previous choices and structures” (Folbre 1994: 39). Individuals, that is, cannot be reduced to their group identities, and yet group identities impose significant constraints upon individuals. Folbre’s analysis, moreover, takes a unique slant on how this mutual influence and constitution occurs. On the one hand, by treating social structure as a combination of competing and conflicting social groups, she makes the pressures acting upon individuals complex and multidimensional. On the other hand, by explaining individuals’ relative autonomy as an ability to move between social groups and negotiate discontinuities in social structure, she creates a conception of agency that has particular policy implications. This thinking about the complexity of social structure and how individuals negotiate it also creates a foundation for thinking about individuals as reflexive beings. Clearly, the idea that individuals identify with others points towards a kind of selfreferent behavior involving one type of self-concept. When individuals acquire a group identity, their self-concept includes a sense of likeness with others in that group. Individuals know who they are by knowing who they are like. However, the dynamic that Folbre investigates in connection with how individuals move between and negotiate membership in multiple social groups involves an additional type of self-referent behavior. Essentially, individuals acquire a personal identity selfconcept by seeing themselves as maintaining a continuous sense of self across and apart from their multiple group associations (Davis 2002c). It helps to see how this understanding differs from the utility theory multiple selves approach of neoclassical economics. As discussed in Chapter 4, as long as an individual is simply made of multiple utility functions, the individual lacks a unity. What Folbre adds to this is the idea that individuals negotiate their multiple group associations, thereby adding a concept of agency that is missing from the neoclassical framework. We now also see that this idea of the individual as an agent is tied to an understanding of individuals being reflexively engaged in self-referent behavior. Thus, Folbre combines a structure–agent type of model with an understanding of individuals as reflexive agents. 6.4.5
Intersubjectivist economics
Intersubjectivist economics (or French conventions theory) has its origins in France in the 1980s in the writing of Dupuy, Orléan, and Thévenot (Dupuy 1988, 1992, 2002; Orléan 1989, 1990, 1992; Thévenot 1989), particularly in connection with the analysis of conventions and J. M. Keynes’s treatment of speculative behavior, and has been introduced to Anglophone economists by Fullbrook (Fullbrook 1996, 1997, 2002). Drawing on Keynes’s “newspaper beauty contest” account of specula-
The embedded individual
125
tion in Chapter 12 of The General Theory of Employment, Interest and Money (Keynes 1936: 156), which models investment decisions upon the expected choices of others, individuals are treated as interdependent at the level of consciousness—thus “intersubjective”—and speculation in financial markets and other conventional structures is explained in terms of imitative or mimetic structures of interaction between individuals.6 From this perspective, neoclassical economics errs in regarding “human desires and proclivities as fundamental data, which, like the masses of physical bodies in classical mechanics, are not affected by the relations being modeled” (Fullbrook 2002: 1). The consequence is that interaction between individuals is a purely “intrasubjective” interaction, which, like Newton’s interaction between atoms, still leaves individuals/atoms unchanged. This conception may have been meaningful in an earlier age, but is not compatible with a world of instantaneous mass communication in which “what we think, desire and decide as economic actors depends a great deal on what other actors are seen to think, desire, and decide” (Fullbrook 2002: 2). The intersubjectivist school is consciously neither methodological individualist nor methodological holist. As Fullbrook puts it, not only are agents interdependent, but agents and structures are interdependent as well, “so that conceptualizing one level inevitably involves concepts relating to the other” (Fullbrook 2002: 3). This means that agents and structures influence one another. While individuals alone are agents, “conventions, including social beliefs and institutions, transcend the individual, shaping and constraining his or her behavior” (Levy 2002: 254). Conventions, as the medium of this interaction, must consequently “be apprehended twice: as resulting from individual actions and as constraining subjects” (Revue Économique 1989: 143; quoted in Levy 2002: 254). This framework, moreover, is argued to require a new understanding of rationality. Drawing on the thinking of Habermas, individuals are seen as exercising a strategic rationality, which takes into account the likelihood of cooperation or opposition of others, and a communicational rationality, which operates when individuals make a shared commitment to certain facts and norms (Levy 2002: 262–3). Generally, then, individuals’ behavior is reflexive in that, as intersubjective beings, they re-evaluate their own choices when in interaction with others. Imitation involves this in a particularly strong way, but strategic and communicational rationality also requires that individuals judge themselves in relation to others. Thus, French conventions theory also draws on structure–agent type models, with attention to the reflexive aspects of individual behavior. 6.4.6
Reflected rationality theory
The last approach I discuss concerns the thinking of two economists who have made reflexivity central to their analysis of individuals. One is Hargreaves Heap who focuses on individuals’ “reflective capacities” and their sense of self-worth. The other is Fontaine who examines the different meanings of sympathy and empathy in the history of economics. Both draw on Smith’s “impartial spectator” thinking in The
126
The embedded individual
Theory of Moral Sentiments (1976 [1759]) as a device that individuals may employ to examine themselves in relation to others.7 Hargreaves Heap notes that individuals “reflect on what they do and these reflections give rise to feelings of self-worth, or sometimes the reverse when feelings of shame, guilt and embarrassment are experienced” (Hargreaves Heap 2001: 98). Building on Smith’s idea that individuals’ behavior is influenced by how their actions would appear to an “impartial spectator,” he describes individuals’ sense of selfworth as a “public sense of self-worth that comes from behaviour which conforms with (or breaches) the shared standards or norms of a group” (Hargreaves Heap 2001: 99). That is, “shared standards and norms” constitute benchmarks by which individuals judge themselves, where that evaluation is registered as their sense of public self-worth. Clearly, then, engaging in self-reflection constitutes a form of social behavior. In fact, action motivated by a sense of self-worth can be referred to as “expressively rational” to emphasize that its aim is to express to others the individual’s sense of self-worth (Hargreaves Heap 1989). For example, individuals might act one way or another when wanting to express a sense of self-worth in firms that pay employees according to the effort they are perceived to expend. Actions motivated by a sense of self-worth, then, can be an important motivation in economic behavior. Yet expressive rationality, or, more generally, choice motivated by reflexive considerations, has gone largely unexamined by those concerned with the theory of rational choice. This can lead to problems, since expressively rational behavior cannot be explained as a form of instrumentally rational behavior, and failure to include individuals’ concern for such things as a public sense of self-worth in accounts of interactions between individuals can generate significant “explanatory gaps” (Hargreaves Heap 2001). One example concerns public policy towards the unemployed. Policy informed by instrumental rationality reasoning might call for a tightening of unemployment benefits to increase incentives for job searching. However, if individuals are also expressively rational, lowered benefits may reduce their sense of public self-worth and thereby lead them to reduce job search: “Thus a high level of involuntary unemployment when it persists for long periods can be turned into a permanently high level of equilibrium unemployment as the unemployed develop a culture of unemployment” (Hargreaves Heap 2001: 112). Fontaine’s review of the concepts of sympathy and empathy reveals considerable ambiguity and confusion over the differences between and meanings of these concepts in the history of past and recent economics (Fontaine 1997, 2001). To clarify their meanings, he defines sympathy as a concern for the welfare of others (as associated with interdependent utility functions), and empathy as the process of putting oneself in someone else’s shoes. He also distinguishes two forms of empathy: “‘empathetic identification’, which implies an imaginary change of circumstances and personhood with another, and ‘partial identification’, which implies only a change of circumstances” (Fontaine 1997: 263). Both forms of empathy, it turns out, have played a role in debates over interpersonal utility comparisons in neoclassical welfare economics since
The embedded individual
127
the 1950s. Partial empathetic identification allows those empathizing with others to keep their personal preferences and yet enter the circumstances of others, while empathetic identification extends empathy to others’ subjective features but only at the cost of blurring the personal identity lines between individuals (Fontaine 2001: 388–9). Fontaine does not use the language of reflexivity as Hargreaves Heap does, but it is not difficult to see how empathy can be understood as a form of reflexive behavior. With sympathy, individuals merely register a concern for the welfare of others from their own points of view, and ignore any role those points of view play in their judgments of others. With empathy, however, individuals separate themselves (partially or completely) from their own points of view, and consider the point of view of others with whom they empathize. In terms of interpersonal comparisons, utility or otherwise, the idea is that one might become more objective—indeed, less subjective—in distancing oneself from one’s own point of view. However, achieving this kind of objectivity essentially involves individuals forming self-concepts of themselves, namely self-concepts of themselves as subjective beings who are unable to see beyond their own vantage points and are at most sympathetic towards others. That is, the point is not just to put oneself in someone else’s shoes; the point is also to look back at one’s own shoes when not occupying them. Putting aside a possible loss of identity associated with complete identification, one must always still be oneself, even when out of one’s own shoes. Therefore, also placing oneself in someone else’s shoes involves a form of self-referent behavior, or a technique of getting a look at oneself. The idea is thus like the dual-self notion discussed above, though with the difference that individuals use others’ imagined viewpoints as leverage for judging themselves, rather than a set of internal metapreferences.
6.5
The socially embedded individual conception
The premise of this chapter is that there exists a conception of the individual in the history of economics that is alternative to the atomistic conception, namely a conception of individuals as socially embedded. This position represents a minority point of view, since most interpreters of modern thinking about the individual and society argue that this history rather involves a contest between two fundamentally opposed traditions, methodological individualism and methodological holism, where the former is exclusively about individuals and the latter is exclusively about supraindividual entities. I suggest that the latter characterization probably misrepresents the view of the individual in methodological holism, argue that in any event the logic of development of methodological holism has been increasingly to break out an account of individuals as socially embedded, and finally show how elements of this socially embedded individual conception have taken on increasing prominence in contemporary social theory and heterodox economics. The larger framework in which this socially embedded individual conception operates—one that replaces methodological holism as an organizing framework—is
128
The embedded individual
the structure–agent framework widely adopted in social theory. I characterize this as a framework because its main principle—that mutual causality applies to the relationship between individuals and social structures, or that individuals influence social structures and social structures influence individuals—is so general as to accommodate a variety of more specific theories. Though seemingly a trivial and obvious truth, it is basically rejected by methodological individualists, who argue that social structure is influenced by individuals, but that individuals in various fundamental respects are not influenced by social structures. To be fair, some recent developments in mainstream economics challenge this assumption. However, I think that most from both sides of the fence today would agree that mainstream economics has a considerable distance to go before it can be said to operate within a general structure–agent framework. The other main idea that this chapter offers is an interpretation of the socially embedded individual within the structure–agent framework as a reflexive being. The rationale for this claim is my argument that the structure–agent framework cannot ultimately advance beyond methodological holism, in spite of the disclaimers of its proponents, without developing an understanding of the individual as an active being with powers as an agent that are not postulated arbitrarily in an ad hoc manner. The litmus test of whether an account of individuals as agents is not ad hoc is that the individual’s active powers are explainable as a form of social behavior. Thus, rather than posit free will, agency, or a transcendental power to act, the argument here is that individuals engage in self-referent behavior, a form of social behavior, and that this gives us sufficient reason to think that individuals have the ability to influence social structure. The key to this argument is that, when individuals are influenced by social structure, their self-examination and self-evaluation lead to behavior that cannot be reduced in explanation to the social influences upon them. Thus, action that follows reflexive behavior constitutes an independent source of activity in the structure–agent framework. Of course, skeptics can always say that even selfconceptions are socially determined. However, there are two reasons to reject this as a serious argument. First, it also has a deus ex machina quality as essentially an expression of abstract fatalism from outside the social science discourse in which the issue resides. Second, the social science research on self-referent behavior from social psychology does not support the reductionist argument that action stemming from self-referent behavior can be explained in terms of social influences on individuals. Thus, at the level of social science, reflexive behavior qualifies as a source of agency, where that concept is tied to the recursive character of self-referent behavior. I do not suggest, however, that there is one single account of individuals’ reflexive behavior appropriate to economics. The discussion in Section 6.4 of different contributions to structure–agent thinking and reflexive behavior is meant to indicate a direction of research instead of a substantive result. The embedded individual conception, in fact, is in an early state of development, and it is unclear how it will be developed in the future, if at all. Indeed, it is important to appreciate that the idea of
The embedded individual
129
the embedded individual is a conception rather than a systematic theory of the individual. As such, it possesses unworked-out assumptions that may not bear close scrutiny on more careful examination. Like the idea of the individual as atomistic, the individual as socially embedded is largely metaphorical. Moreover, the process of systematically working out the ideas in the atomistic individual conception has gone on much longer and with far greater effort than has been the case with the newer embedded individual conception. Unfortunately for the atomistic individual conception, this process of development and examination has largely undermined it as a theory of the individual, and the initial intuition behind the metaphor has been shown to have been mistaken. The following two chapters test the embedded individual conception using the same identity requirements applied to the atomistic individual conception. Thus, for the embedded individual conception to graduate to the status of a theory of the individual, it needs to be shown that, with this conception, individuals can be distinct from one another, that is, individuated, and also that individuals, if they can be distinguished as relatively independent beings, can also be reidentified across change. The first test involves the greater challenge, since the basic idea behind individuals being socially embedded is that they are not distinguished from one another but all jointly embedded in social relationships. Thus, Chapter 7 addresses this issue first. Chapter 8 addresses whether socially embedded individuals who might be distinguishable might remain so across change. It could be the case that socially embedded individuals can be seen to be relatively independent in an abstract, theoretical manner, but that this kind of relative autonomy ceases to be meaningful in a realistic view of economic life as a dynamic, historical process. Finally, in Chapter 9, I take stock of the socially embedded individual conception, and make general arguments about what it and the atomistic individual conception tell us about economics and ethics.
7 Individuation
Collective intentionality
We think, therefore we are. (Shaftesbury, 1963 [1900]: vol. 2, 275)1
In this chapter I investigate whether socially embedded individuals can be distinguished and individuated from one another in heterodox economics as I have characterized it in the previous chapter. As with my examination of the problem of individuation in connection with the neoclassical atomistic individual conception, this requires first stating what the relevant conception of the socially embedded individual is, and then asking whether there are reasonable means of distinguishing individuals from one another within the framework of this conception. Thus, in this chapter I begin by going beyond the general characterization of individuals as reflexive beings in the structure–agent framework discussed in the last chapter to advance one specific account of socially embedded individuals. There are of course other accounts of socially embedded individuals than the one advanced here. An advantage of this one is that it explicitly begins with an account of individual behavior in a particular kind of social setting, and thus clearly is about individuals rather than social entities. This makes it possible to formulate the individuation test for at least one account of socially embedded individuals, which is the main issue with which I am concerned in this chapter It may seem odd, however, to propose that a discussion of socially embedded individuals begin with an account of individual behavior, even if in some social setting, since the idea that individuals are socially embedded seems to require at the very least that the starting point of the investigation be the relationships between individuals. However, the specific account of socially embedded individuals I employ—shared or collective intentionality analysis—is not an account of individuals outside social relationships, but one that instead sees social relationships as embedded in individuals. The basic idea is that, while only individuals form intentions, alongside those intentions expressed from a first-person singular point of view, individuals also express shared or collective intentions from a first-person plural point of view. When we speak of shared or collective intentionality, we focus on a kind of intentional behavior
Individuation: collective intentionality
131
different in nature from the kind of intentional behavior that underlies the atomistic conception of the individual. In effect, intentions expressed in first-person plural terms make social relationships internal to the individual, because the individual’s use of the first-person plural puts the individual automatically in a relationship with all other individuals to whom it applies. At the same time, that individuals can express themselves in first-person plural terms does not mean they do not also express themselves in first-person singular terms. Thus, my argument in this chapter is not that all individual behavior is based on shared or collective intentionality thinking; my argument is that socially embedded individual behavior is based on this kind of thinking, and a complete account of individual economic behavior should consequently explain both relatively autonomous and socially embedded experience. Determining the extent to which individuals rely on collective intentionality reasoning in economic life is thus a matter of establishing the extent to which, and in what particular sorts of social settings, individuals express themselves in first-person plural terms. In the discussion here, I restrict my attention to relatively cohesive, institutionally well-structured social settings—social groups—on the assumption that shared intentions are arguably more reliably formed in these circumstances than in larger, more diffuse, loosely organized social settings. Compare, for example, the need an individual feels in a place of employment to correctly express intentions that are likely to be shared by other employees and an employer (“we face a production deadline”) versus the lesser need an individual feels in larger, more loosely organized social settings to correctly express intentions that may only possibly be shared by others (“we’re concerned about inflation”). Basically, since consequences of misstating shared intentions have stronger force in the former type of setting, individual behavior in such settings is more likely to be socially embedded. I thus concentrate on individuals in relatively cohesive, institutionally well-structured social settings to provide the strongest demonstration of the logic of first-person plural thinking in economic life. A key consequence of treating individuals as socially embedded in groups is that their behavior can no longer be explained strictly in instrumentally rational terms. Individuals socially embedded in groups sometimes also act in an instrumentally rational manner, forming intentions in a first-person singular way just as they do in more autonomous social circumstances, but, on those occasions when they express first-person plural intentions, they clearly exercise a different sort of rationality. Essentially, unless we suppose that individuals are always duplicitous, expressing a shared intention commits the individual to an intention held by others. In effect, expressing a shared intention introduces an element of obligation into the individual’s decision making—not in the form of an external constraint on instrumentally rational behavior, but as an alternative kind of motivation. The key point is that a shared intention is also the individual’s intention. Thus, the individual is bound, but voluntarily so. Without implying any moral connotation, I thus distinguish this alternative kind of behavior flowing from shared intentions as a deontologically rational or a rationally principled type of behavior.
132
Individuation: collective intentionality
Section 7.1 introduces the concept of shared or collective intentionality, relying primarily on the thinking of Toumela, one of the leading contributors to the subject. In Section 7.2, I return to the themes of the last chapter to ask whether the collective intentionality framework, as set forth by Tuomela, can be integrated into the structure– agent model, particularly as that model was amended in the previous chapter’s discussion to accommodate a treatment of individuals as reflexively engaged in self-referent behavior. Section 7.3 turns to how the socially embedded individual conception explained in collective intentionality terms entails a revised view of individual economic behavior. Section 7.4 investigates whether socially embedded individuals as developed in this and the last chapter can be distinguished and individuated from one another. Finally, Section 7.5 compares the atomistic and socially embedded individual conceptions of the individual in the ways that they each address the problem of individuation.
7.1 7.1.1
Collective intentionality analysis Intention
Why is it that collective intentionality thinking is not more familiar and well established than it is? The distinction between shared intentions and ordinary individual intentions seems relatively straightforward when we think of everyday language usage. People regularly and commonly use “we” language in an intentional manner when they express intentions that have to do with their being part of some social group (“what we want,” “what our department decided,” “what we in the community have chosen to do,” etc.), while they use “I” language when they see themselves acting independently of others (“what I want,” “what I believe,” etc.). Yet until recently the philosophical literature on intentionality, with few exceptions (e.g., Sellars 1968: 223–6), has ignored the differences between these two forms to concentrate on the abstract concept of intentionality. Analytic philosophers in the Anglophone tradition for the better part of the twentieth century have explained the concept of intentionality in terms of the abstract idea that mental states have an “aboutness” to them (e.g., Anscombe 1979 [1963]; Davidson 1980). Equally general is the idea that having an intention involves settling on a course of action, either by deciding to act in some way or by excluding considerations that favor alternative courses of action. Also pitched at a high level of generality is the idea that intention involves forming a commitment to action, and that intentional action only occurs when an agent aims to bring something about by acting in a purposeful way. Nothing in all of this helps us understand the different conditions under which individuals might express first-person singular or first-person plural intentions. Indeed, nothing in the standard literature suggests there is any substantive difference between first-person singular and first-person plural sorts of intention. At the same time, in the social sciences and especially in economics, proponents of individualist explanations have long argued that it makes no sense to ascribe inten-
Individuation: collective intentionality
133
tions to groups of individuals on the grounds that only individuals can have intentions. This familiar and not unreasonable critique of the more extreme sort of holist argument does not apply to shared or collective intentionality as it is understood here, because shared intentions are individual intentions about groups, not the intentions of groups. Nonetheless, one effect of this critique has been to substitute an individual intentions–group intentions dichotomy for a first-person singular– first-person plural dichotomy, thereby closing off investigation of the specific conditions under which individuals express shared intentions. This critique also conflates the question of intentionality and the question of agency. The implication is that, since only individuals form intentions, only individuals can be agents. However, casual inspection of the world strongly suggests that groups (nations, political parties, firms, armies, social groups, movements, etc.) have been and will continue to be agents over and above the individuals that make them up. One way in which this could be reconciled with saying that only individuals form intentions is to investigate how individuals’ shared intentions facilitate groups acting as agents. 7.1.2
Social groups
Social groups that are relatively cohesive have been extensively investigated in sociology, anthropology, and social psychology. A social group may be characterized as (a) a plurality of individuals tied to one another by (b) some principle of membership that implies (c) a system of individual rights and obligations. A social group is not the same thing as a social category, which is a researcher’s grouping of individuals according to some set of shared characteristics that the researcher selects. Groups, rather, are collections of individuals whose shared characteristics constitute the basis for their interaction with one another. Of the characteristics of social groups surveyed in the recent literature on the subject, I emphasize the following as particularly relevant to the analysis of groups in economic life: that individuals engage in repeated interaction, that they define themselves as members of a group, that they are defined by others as belonging to the group, that they share and observe group rules and norms, and that they participate in a set of interlocking roles that are central to how the group functions (Cartwright and Zander 1968: 48). One advantage of construing groups in these terms is that it allows us to say that groups need not operate on a face-to-face basis nor necessarily be small in size. Much sociological, anthropological, philosophical, and social psychological research focuses on small groups that exhibit regular face-to-face contact. Indeed, in the limit a relationship between two people can be seen as a kind of group if those individuals see themselves as being in some type of repeated interaction with one another and observe rules and norms that determine roles for them in the relationship. Gilbert, also a leading contributor to collective intentionality analysis, uses as one of her main examples the idea of two people “taking a walk together” (Gilbert 1989). Were “taking a walk together” a regular interaction between two individuals, on the
134
Individuation: collective intentionality
understanding here they would constitute a group. More long-lasting relationships of all kinds between two individuals, then, would also qualify as instances of groups. In economic life, however, groups are generally seen as being larger and somewhat more impersonal, and they not infrequently involve limited or even no face-to-face contact (firms, unions, cartels, governments, etc.). This is the sort of case I focus upon in order to concentrate on individuals’ social embeddedness. The two-person limiting case group would still be worth investigating in economics to explain recurring market interactions between individuals, where trust relationships are better modeled along the lines of group behavior than in atomistic individual terms. However, I do not address this case here. 7.1.3
Tuomela
The recent philosophical literature on collective or shared intentionality distinguishes we-intentions, corresponding to the use of “we” expressions, from I-intentions, corresponding to use of “I” expressions. We-intentions are then explained as a structure of mutually reinforcing, reciprocal attitudes shared by individuals in a social group. The main contributors have been Bratman (1993, 1999), Gilbert (1989, 1996), Searle (1990, 1995), and Tuomela (1991, 1995). Others, such as Etzioni (1988), have distinguished “I” and “we” thinking but do not employ the idea that shared intentionality can be described as a structure of mutually reinforcing, reciprocal attitudes. I rely on Tuomela’s work, which is the most extensive and well developed. Tuomela’s analysis is also conservative in that, though he sometimes makes informal reference to intentions being shared, he emphasizes that this is not meant to imply that we-intentions exist in society apart from individual we-intentions. Rather, a we-intention is an individual’s attribution of an intention to the members of a group to which the individual belongs, based on that individual both having that we-intention and also believing that it is held by other individuals in the same group. That is, I can only use “we” language that pertains to you and me if I think that you would similarly apply it to you and me. Thus, expressing we-intentions is a matter of whether there exists a set reciprocal attitudes, not whether there is an actual sharing of attitudes. Indeed, in the limit an individual could have a we-intention that no other individuals have, if that individual were mistaken about others’ we-intentions. Thus, a we-intention is not a supraindividual group intention separate from the attributions that individuals make to groups, and when people use expressions such as “the intentions of the group,” this is just a shorthand device for referring to a collection of individual we-intentions on the part of individuals in the group. Two characteristics of we-intentions are central to Tuomela’s analysis. First, an individual expressing a we-intention believes that this intention is held by other group members; second, the individual expressing this intention believes it is mutually held by members of the group. Consider the case in which an individual’s weintention is rooted in an attitude (“fear”) that the individual believes other group
Individuation: collective intentionality
135
members also attribute to the group. For an individual A who is a member of a group G, “A we-fears that X if A fears that X and believes that it is feared in G that X and that it is mutually believed in G that X is feared in G” (Tuomela 1995: 38). “X is mutually believed” if not only do I believe others believe X but also they believe that I believe X.2 On this basis, A might suppose that “group G has some intention” reflecting “G’s fear of X” (say, whether the group will avoid some danger). Of course, A can only surmise that others in G have the same fear and also that the fear of X is mutually believed by members of the group. Ideally, that X is “mutually believed” would involve saying that the fear that X is believed by absolutely everyone, but Tuomela allows “mutual” to have strong and weak interpretations, because groups themselves have strong and weak criteria for supposing their members share a belief, attitude, or intention. The main point is that we-attitudes are a group attitude not in the sense that a group over and above its members has an attitude towards something, but in the sense that individuals “generally” in a group have some such attitude that they express in “we” terms. Thus, saying that they “generally” have a we-attitude depends not just on the mutual belief condition but on both conditions which when combined give us reason to suppose that individual members of a group are justified in saying what they (that is, “we”) intend.3 Tuomela uses this framework to distinguish between rules and norms. Rules are the product of an explicit or implicit agreement brought about by some authority, and used to determine a distribution of tasks and activities to individuals. Rules may be formal and written, such as laws, statutes, regulations, charters, bylaws, etc., or they may be informal agreements between individuals, sometimes orally established and sometimes silently agreed to. In contrast, in the case of norms a network of mutual beliefs substitutes for actual agreements between individuals in determining distributions of tasks and activities across individuals. As with we-intentions generally, mutual beliefs are beliefs reciprocally established between individuals, such that each believes that others have the same belief, and each also believes that others think the same about the others, and so on in a structure of reinforcing, mutually held beliefs. Rules and norms are both understood to have motivational force, meaning that they constitute reasons for action on the part of the individuals who accept them. Indeed, rules and norms are typically framed as “ought” principles and impose requirements on individuals as members of groups in the form of specific prescriptions for individual action. Formally, individual A feels obliged to do X because A is a member of a group with a we-intention that can be represented in terms of a rule or norm to the effect that “we believe members of the group should do X.” However, rules and norms are different by virtue of the different means by which they enforce a distribution of tasks and activities among individuals (Tuomela 1995: 22–4). The prescriptive force of rules derives from there being sanctions that apply, whether formal/legal or informal, to those individuals who do not observe them. In contrast to rules, sanctioning with norms takes the form of approval or disapproval on the part of others. Because norms are internalized by individuals, in that they themselves
136
Individuation: collective intentionality
accept them as reasons for acting, individuals apply others’ potential disapproval to themselves, as when feeling shame or embarrassment. In Tuomela’s framework, then, it can be said that rules are the basis for institutions, and norms are the basis for social values. While it is true that many institutions also involve norms, as relatively settled social arrangements, institutions generally place greater reliance on rules. In contrast, since social values are rarely rooted in agreements, even informally, they usually place little weight on rules. Rather, social values reflect systems of mutual belief about individuals’ interaction with one another. Thus, when individuals create and/or change institutions, they adopt new rules and produce new we-attitudes defining group action within an institutional framework that can be characterized in terms of agreements and corresponding sanctions. When individuals develop and/or influence social values, they adopt new norms and produce new we-attitudes defining group action within a social value framework on the basis of their mutual beliefs and systems of approval and disapproval. In both frameworks, rules/institutions and norms/social values, we-intentions are the foundation for understanding group action. Individuals thus influence institutions and social values as members of groups, and group action is the intermediate link between individual action and supraindividual institutions and social values missing from mainstream accounts of individuals’ influence on institutions and social values. Finally, it is important to emphasize that rules and norms create obligations for individuals in terms of how different tasks, rights, and positions apply to different individuals in groups. Tuomela characterizes an individual’s position within a particular group in terms of that individual’s tasks and rights within that group. An individual’s tasks and rights are then further distinguished according to whether they flow from rules or norms operating within the group, that is, whether they are rule-based tasks and rights or norm-based tasks and rights. In contrast, across groups, individuals’ social positions are understood in terms of the whole array of actions that individuals are required and permitted to carry out across various economic and social settings. These social positions assign individuals a variety of different tasks whose performance is in each instance protected by rights, where these task–right combinations may themselves exist within established modes of implementation that are also understood in task–right terms. The overall framework thus explains individual rights and duties within and across groups in terms of task–right pairs that ultimately have we-attitudes in groups as their foundation. In Section 7.4, however, I concentrate on individuals’ positions within groups in order to focus on task–right combinations that have relatively clear implications for economic behavior.
7.2 Collective intentionality and the structure–agent framework Can the collective intentionality framework, as set forth by Tuomela, be integrated into the structure–agent model, particularly as modified to allow conceptualization
Individuation: collective intentionality
137
of individuals as engaged in self-referent behavior? As seen in the last chapter, the main idea behind the model itself is that individuals influence and are influenced by social structure, while self-referent behavior exists when individuals take themselves as an object in one way or another. First, then, I address what collective intentionality tells us about the structure–agent model, and second I discuss what collective intentionality might imply about self-referent behavior. When individuals form shared or collective intentions in groups, they commit themselves to what those shared intentions require of them. On one level this is a matter of observing rules and norms. Whatever type of rationality shared intentions involve (a matter I address in the next section), rules and norms limit and influence individual behavior and are thus evidence of the effects of social structure on individuals. Furthermore, since groups (especially the more cohesive and structured ones) are almost always organized in ways that distinguish different positions for their members, having shared intentions in groups is also a matter of participating in division of labor and responsibility which differentially allocates tasks and rights to individuals on the basis of their respective positions within groups. Not only does social structure in the form of rules and norms limit and influence individual behavior, but it also limits and influences individuals through the medium of the tasks and rights that are allocated to them in groups. Thus, in spite of shared intentions being fully individual intentions and therefore voluntary in nature, individuals having shared intentions occasions them being influenced by the social structure of those groups in which those intentions are expressed. Individuals forming shared intentions, however, also influence the social structure of groups, since shared intentions change as a result of the interaction between individuals, and this can change the social structures of groups. Tuomela’s representation of a shared intention as a set of reciprocal attitudes means that shared intentions are not something invariably settled and fixed that stand over and apart from the individuals who have them, but are rather sets of continually converging and reconverging individual attitudes that may exhibit both stability and instability. Thus, shared intentions and the social structures based upon them can be relatively settled, but they can also break down, fail to form, exist for relatively brief periods, have conflicting aspects, obtain for only subsets of individuals in a group, operate across groups, and so on. Indeed, in contrast to ordinary intentions, an especially interesting aspect of shared intentions is that they may conceal acts of deception on the part of individuals who express them but nonetheless privately reject them and either have different shared intentions or simply act on instrumentally rational grounds. Consequently, on a collective intentionality understanding of individuals and social structures, individuals also influence social structures. Does collective intentionality analysis, then, also provide a way of saying that individuals engage in self-referent behavior? Alternatively, do individuals engaging in self-referent behavior exercise a kind of rationality appropriate to the structure– agent framework that we can explain in collective intentionality terms? We saw in
138
Individuation: collective intentionality
the last chapter that rationality in the structure–agent framework can be characterized as a “situated rationality” (Lawson 1997: 187). Individuals who express shared intentions, however, can be said to exhibit a kind of situated rationality in that their expressing those shared intentions depends upon other individuals expressing them as well. Tuomela’s two conditions for expressing we-intentions, recall, are that individuals assume that their we-intentions are held by other group members, and assume that these we-intentions are mutually believed. To see how this involves a situated rationality, suppose these conditions do not hold, as when an individual expresses a mistaken we-intention (Tuomela’s limiting case). Since a shared intention is a set of reciprocal attitudes across a group, other members of the group are likely to object that this individual’s we-intention cannot be ascribed to the group: “You say we will do this, but we have no intention of doing it.” Feedback such as this, however, generally compels individuals in error to re-evaluate what they say and believe. But if this re-evaluation takes the form of their asking themselves why their views are at odds with those of others, then the adjustment process that feedback initiates depends on individuals engaging in a form of reflexive behavior. This adjustment process need not be thought to occur only in the extreme circumstance in which one person is completely at odds with everyone else. Because a group of individuals is continually converging and reconverging on a whole system of shared intentions, we should suppose that all group members are continually engaged in evaluating and re-evaluating whether they are correctly expressing group intentions. Shared intentions in groups, then, constitute a type of social system that integrates its members specifically through their exercise of self-referent behavior. We can extend this picture of feedback effects and self-referent behavior by also recalling that Tuomela explains the positions that individuals occupy in social groups and the task–right combinations associated with these positions in terms of shared intentions. Occupying a position in a group involves participating in a division of labor and responsibility within that group. Thus, individuals generally have an understanding of the overall organization of those groups of which they are members, and this enables them to express shared intentions about how those groups function in their different aspects. This means that the degree to which individuals reflexively evaluate themselves as members of groups is really quite extensive. Every task an individual has, every right one feels entitled to invoke, has as a background a system of shared intentions tied to group members’ understanding of how each individual fits into the overall division of labor and responsibility of the group. If we again emphasize that groups involve whole systems of shared intentions upon which individuals are converging and reconverging, we need to suppose that all group members are more or less continually engaged in evaluating and re-evaluating whether they are correctly expressing group intentions on how the group is organized, and how tasks and rights are allocated in it. Not only, then, is the collective intentionality interpretation of socially embedded individuals consistent with the general structure–agent framework, but it also
Individuation: collective intentionality
139
provides a plausible understanding of individuals as reflexive beings within that framework. Individuals influence and are influenced by social structure, and the way in which this occurs involves individuals in an adjustment process with respect to shared intentions that leads individuals to reflexively evaluate how their we-intentions compare with those of others. Clearly, this general account of socially embedded individuals—structure–agent framework, self-referent behavior, collective intentionality analysis—is different in a number of respects from the conception of individuals as atomistic beings. In the following section, then, I turn to how the socially embedded individual conception requires that we revise our view of individual economic behavior.
7.3
A revised view of individual economic behavior
The behavior of atomistic individuals is understood in instrumentally rational terms because individual objective functions are defined solely in terms of individuals’ own preferences. With no basis for action other than their own preferences, and putting aside that they might act out of habit or behave irrationally, atomistic individuals can do nothing other than seek to satisfy their own preferences. In contrast, when we treat individuals as socially embedded, we no longer say that individuals act only on their own preferences, because they also act in accordance with those rules and norms that function as “ought” principles—what I have termed a deontologically rational or a rationally principled type of behavior. However, proponents of the atomistic individual conception understand rule following and norm observance in instrumentally rational terms. Might “ought” principles operating in social groups then still be explained in instrumentally rational terms? Is instrumental rationality a sufficient explanation of individual behavior? There seem to be three objections to saying that something other than instrumentally rational behavior is involved here. 7.3.1
First objection
One way in which this might be argued is to say that observing rules and norms that have the force of “ought” principles still involves being instrumentally rational, though with the difference that the individual becomes subject to constraints additional to those usually assumed in standard constrained optimization analysis, namely, constraints associated with group membership. Though these additional “social group” constraints further narrow individuals’ choice sets, individuals would still maximize preferences, suggesting that socially embedded individuals are not significantly different from atomistic individuals. This argument, however, ignores what is involved in saying that individuals observe rules and norms on account of their sharing intentions with others regarding those rules and norms. Shared intentions are those intentions that individuals ascribe to the groups of which they are members. However, as intentions, they must stem from individual objective functions rather
140
Individuation: collective intentionality
than constitute constraints on those objective functions. That is, shared intentions are like our ordinary intentions in expressing what individuals choose to do rather than what they are limited to doing. It is true that individuals in groups are more constrained in their behavior compared with when they act outside groups. However, this type of constraint has an intended aspect and is consequently different from the usual sort of constraint which is entirely external to the individual. 7.3.2
Second objection
A second argument for explaining shared intentions in instrumentally rational terms accepts that shared intentions stem from individual objective functions, but argues that they express individual we-preferences, just as ordinary individual intentions express an individual’s own preferences (or I-preferences). We-preferences have been analyzed by Sugden and characterized as team preferences (Sugden 2000; also cf. Bacharach 1999). Sugden explicitly rejects collective intentionality analysis as developed by Tuomela, Gilbert, and others, on the grounds that it assumes individuals are bound by obligations or “ought” principles, which he regards as inconsistent with an account of instrumentally rational behavior (Sugden 2000: 189–90). To preserve the latter, he reasons, the former has to go. This implies that rules and norms are things that members of teams prefer to observe rather than believe they ought to observe. Moreover, if we-intentions are really the product of we-preferences, then it seems it is no longer necessary to say that individuals in teams (or groups) need to be treated as socially embedded, since the obligations or “ought” principles they observe are what they prefer. Sugden essentially draws this conclusion when he argues that the “existence” question regarding whether teams and other groups exist (and therefore can act as agents) is independent of the theory of instrumental rationality enlarged to include we-preferences. Were groups thought to be agents over and above their members, there clearly would be a stronger case for saying that their obligations and “ought” principles were not always preferred by their members. Sugden’s argument, accordingly, depends on supposing that we-preferences do not really impose obligations or “ought” principles on individuals. Why is it, then, that Tuomela and other proponents of collective intentionality analysis claim that this is a necessary dimension of we-intentions? The answer lies in their analysis of shared intentions as sets of reciprocal attitudes across individuals in groups. Though shared intentions are indeed individual intentions, unlike team preferences, which represent only what an individual independently prefers for the team, an individual’s shared intention is one element in a set of reciprocal attitudes. Thus, when individuals ascribe intentions to groups of which they are members, this represents not what they prefer to ascribe to the group, but rather what they believe to be the group’s intention on the basis of what they believe that they and other group members believe to be the group’s intention. By this token, shared intentions imply “ought”
Individuation: collective intentionality
141
principles because individuals share an intention over which they have very limited influence. Indeed, this combination of sharing an intention and having it stand over oneself helps explain the particular quality of “ought” principles as binding precepts that individuals nonetheless embrace. Preferences, by comparison, have but one master, namely, the individual. It is true that team preferences have a shared aspect to them. However, without a set of feedback connections between individuals, such as Tuomela describes for shared intentions, the shared nature of team preferences is simply the result of an accidental alignment of individuals’ we-preferences about teams of which they are members. Team members may happen to share preferences about their teams. Yet if these preferences regarding the team begin to diverge, there is nothing in the interaction between individuals that brings about an adjustment in individuals’ preferences regarding the team. Indeed, there are many examples of teams in the real world that operate on the basis of Sugden’s team preferences, and as a result break down, simply because individuals are driven by what they prefer rather than by what they believe obligates them. The problem, basically, is that, with we-preferences, just as with ordinary individual preferences, de gustibus non est disputandum. That is, individuals retain their atomistic status, and the “teams” of which they are members do not exist as teams in the customary sense of the term. 7.3.3
Third objection
These conclusions, however, suggest a third argument regarding how instrumental rationality might be sustained vis-à-vis collective intentionality analysis. Suppose that we treat rules and norms as conventions understood as coordination equilibria (Lewis 1969). Then, using evolutionary game theory, individuals can be seen as instrumentally rational players who seek the best possible response to one another’s individual strategies (a Nash equilibrium), and rules and norms can be explained as endogenously determined sets of reciprocal expectations. This would allow for a feedback/adjustment process, as operates in collective intentionality theory, but it would not explain this process in terms of “ought” principles. Rather, following Hume’s view of conventions, individuals find it in their interest to conform to rules and norms to which they expect others will conform. There are different ways of explaining why individuals would find this in their interest. Hume relied on sentiments of approval and disapproval, and indeed used this as the basis for his theory of justice. Since a system of justice implies “ought” obligations, this game theoretic/ instrumental rationality framework can also be argued to explain the “ought” content of rules and norms, whether in moral or pragmatic terms. However, in contrast to collective intentionality analysis, “ought” principles in this instance derive from what individuals find to be in their interest. In collective intentionality analysis, “ought” principles derive from shared intentions, and shared intentions are explained in terms of individuals’ use of first-person
142
Individuation: collective intentionality
plural “we” language. Barring cases of deceit, first-person plural “we” language cannot be explained in terms of first-person singular “I” language unless one denies elemental differences in human language, and engages in a reductionist sort of reasoning that has no support among linguists. In collective intentionality analysis, the reason that shared intentions imply “ought” principles is that they require a commitment on the part of the individual that is absent in the case of ordinary intentions that can be expressed in first-person singular terms. Thus, it seems clear that “ought” principles that derive from collective intentionality analysis are not reducible to “ought” principles that might emerge from a Humean framework. However, this does not imply that the latter involves an unacceptable account of “ought” principles, or that this account should be eliminated to produce one of “ought” principles cast exclusively in shared intention and commitment terms. Rather, it seems that both reductionist arguments should be rejected, because neither is successful, and because both are part of the view that thinking about individuals and society can be explained in terms of two inalterably opposed intellectual traditions: methodological individualism and methodological holism. Indeed, both accounts of “ought” principles arguably have real-world foundations. Just as there are teams that operate (often poorly) in terms of individual we-preferences, so there are “ought” principles based on instrumentally rational behavior. Just as there are social groups that operate (usually more successfully) in terms of we-intentions, so there are “ought” principles based on individual commitment. My position is that individual behavior is complex in being rooted in both types of intention. The challenge economists consequently face is to determine the mix of types of behavior associated with different kinds of intention and to properly ascribe each kind of behavior to the correct real-world circumstances. Much mainstream economics, because of the dominance of the atomistic individual conception, imperialistically imposes instrumental rationality arguments on social settings where it does not apply. In using the wrong explanation in the wrong circumstances, mainstream economists impose “thin” institutional explanations that overlook how the functioning of some social groups and institutional structures depends upon “ought” principles more strongly than can be explained in instrumentally rational terms. The holist economics tradition, in contrast, has at times been equally imperialistic, though in the reverse direction, in using social whole-type explanations in circumstances for which they do not apply. These “thick” institutional explanations overlook the extent to which individuals are relatively free of shared intentional experience, as well as individuals’ need for navigating across social structures. I do not attempt here to set forth a specific account of individual economic behavior as complex. There are a variety of different ways in which the two can be related.4 However, consider an example. Suppose an employee in a business is assigned a set of rule- and norm-based tasks associated with doing a particular job. If one rule is to invoice customers by the end of the month, and the norm for how this is to be done is to include in the invoice a complete description of all purchases made by those customers, the individual assigned these tasks is likely still free to perform
Individuation: collective intentionality
143
them in a variety of ways (inquire as to customer satisfaction, pursue follow-up orders, institute new record-keeping practices, etc.). How well individuals do their assigned jobs, then, can be a matter of the extent to which they also act on their own preferences regarding the way a job is best done. They consequently act in an instrumentally rational way when already behaving in a rationally principled manner.
7.4
The individuation of embedded individuals
I argued in Section 7.2 that socially embedded individual agents influence social structure. However, to say that they do not simply have this influence in the aggregate requires showing that socially embedded individuals can also be distinguished from one another and individuated as relatively autonomous economic agents. Certainly there are good reasons for thinking that they cannot be distinguished as such. Whereas the atomistic individual conception is presumptively about individuals in explaining them in terms of individual characteristics, namely, subjective inwardness, the embedded individual conception explains individuals in terms of social characteristics, namely, how they are related to one another, or how they are embedded in a common social space. Thus, the challenge this conception faces is to explain how individuals can be fundamentally related or like one another and yet also be distinct beings. At the same time, any such explanation must treat individual behavior as a kind of social behavior, since the structure–agent framework, in which I assume embedded individuals must be understood, presupposes that individuals are influenced by the social structures that they also influence. 7.4.1
Different positions?
One argument for distinguishing embedded individuals proceeds from the assumption that social groups and institutions possess an internal heterogeneity that creates distinct positions for the individuals who occupy them. Consequently, though all individuals would be related to one another in light of the way that positions are related in a group or institutional structure, and also be like one another in sharing this same embedding, they might yet be thought distinct from one another by virtue of occupying different positions. However, there are obvious objections to this argument. First, groups and institutions often have different people occupying the same kinds of position. From the point of view of the organization, they are indistinguishable, even if they are different individuals. Second, even in the unlikely event that every individual occupies a different position, it is still possible for essentially identical individuals to occupy different positions. Thus, institutional heterogeneity per se does not help us understand how socially embedded individuals might be distinguished from one another. But this is as one would expect, since the general strategy attempts to employ social characteristics of individuals as an individuating device when social characteristics rather explain commonalities between individuals.
144 7.4.2
Individuation: collective intentionality Different abilities?
This suggests that embedded individuals need to be somehow distinguished from one another in the same basic way in which atomistic ones are, that is, in terms of some set of individual characteristics. I propose, then, that the focus on individuation shift to how particular individuals come to occupy different positions within groups. Rather than simply look at individuals in positions, the focus should be on how individuals and positions are matched, since such an account in principle would combine social characteristics with individual ones. Are there, then, characteristics of embedded individuals that are not simply social characteristics? One possibility is differences in experience and abilities. Individuals come to occupy different positions in groups according to differences in their qualifications. There is certainly truth to this, but at the same time such an answer is accompanied with significant difficulties. In the first place, there is no agreed-upon account of what the difference is between natural abilities and acquired abilities. Since the latter are socially influenced, it could always be argued that “individual” characteristics understood as differences in experience and abilities are really social characteristics, and thus that these differences do not distinguish individuals. Second, however, even were these differences somehow sufficient to distinguish individuals, it is not clear that this would help in the larger project here of showing how individuals are agents who influence social structure. Though differences in individual ability are manifest in different individuals having different effects on social structure, this just reduces the account of agency to the unexplained existence of “ability.” This would not be much of an improvement on the sorts of explanation that assume free will, a transcendental power, or some other metaphysical analysis. 7.4.3
Self-referent behavior and intention
There may, nonetheless, be a reasonable way of thinking about individual characteristics on the basis of the discussion of self-referent behavior in the previous chapter. According to the social psychology literature, self-referent behavior is a kind of individual behavior that is also a form of social behavior. Recall, it was argued in the last chapter that the existence of self-referent behavior could be a source of agency, since it breaks the chain of social determination on the individual. Though society influences individuals’ self-concepts, there is no evidence that it determines them. However, the argument that individuals being able to form self-concepts might locate a source of agency is an abstract one, and does not include any account of a mechanism by which this might occur. That is, it does not demonstrate how individuals who engage in self-referent behavior might actually then act in social networks as social individuals and succeed in influencing social structures. Thus, since this chapter employs collective intentionality analysis to explain individual behavior in social settings, here I attempt to interpret this particular analysis in terms
Individuation: collective intentionality
145
of self-referent behavior, in order to ask whether embedded individuals can be distinguished from one another as relatively independent beings and thereby influence social structures. One immediate advantage of proceeding in this way is that it makes the question of how individuals come to occupy positions a matter of their own intentions. The “abilities” strategy leaves individual intention out of the picture, and thus opens the door to the notion that abilities are socially determined. However, by definition, intentions, shared or otherwise, are individual characteristics. Thus, when an individual comes to occupy a given position, in the process accepting whatever tasks and responsibilities this involves, in the traditional understanding of intentionality this is a matter of the individual deciding to do something, forming a commitment to act in some way, or aiming to bring something about in a purposeful manner. This traditional understanding is enlarged when we focus on shared intentions. As first-person plural intentions that concern what is ascribed to the group, shared intentions are also social in nature in that they implicate others in the individual’s use of “we” language, just as they require the individual to attend to the ways that others would use that language. 7.4.4
Self-imposing tasks and rights as individuating
How, then, might individuals be distinguished from one another in collective intentionality terms? The special character of shared intention, in contrast to ordinary intentions, is that it combines voluntariness and compulsion. When individuals use “we” language, they bind themselves to whatever that language implies, but they do so intentionally. By including themselves in the “we,” they limit their own discretion, but do so at their own discretion. Of course, there are cases where this characterization does not apply. Individuals can express we-intentions deceitfully, and secretly not include themselves in whatever would be required of themselves along with those others to whom these false shared intentions apply. At the other extreme, individuals can sign on to we-intentions that they do not accept, and then only feel the force of compulsion. However, the great middle ground of first-person plural speech has a dual-sided nature. When individuals use it, they bind themselves by precepts that they also embrace. That they are able to do this demonstrates that they have a capacity to impose things upon themselves. The language “impose upon oneself” or “self-impose,” I suggest, captures this dual character and refers to a variety of types of circumstance in which people freely make themselves do things. Here, I am only interested in one sort of circumstance in which this occurs, namely, expressing shared intentions. Thus, when individuals express shared intentions, they “self-impose” upon themselves whatever those shared intentions imply, in this instance, that they carry out various tasks and exercise associated rights in a division of labor and responsibility in a social group.
146
Individuation: collective intentionality
What is important about this capacity in the current context is that only individuals can self-impose upon themselves. The group cannot self-impose on the individual, though it certainly can impose upon individuals against their will. Put in we-intention terms, for the great middle ground of first-person plural speech, only individuals can voluntarily impose upon themselves what such intentions imply. Thus, the capacity to self-impose upon oneself is necessarily a characteristic of individuals. However, in the particular context of shared intentions, it is a characteristic of individuals that is also a social characteristic. In this chapter I have explained shared intentions thinking in connection with social groups. In this connection, the capacity that individuals have to impose things upon themselves is coextensive with their accepting what is involved in participating in a social group. Accordingly, exercising this capacity tells us how individuals come to occupy positions in groups. Since how individuals come to occupy positions in groups is the basis proposed above for distinguishing individuals from one another, we now find that individuals’ capacity to selfimpose upon themselves provides a way of saying how socially embedded individuals can be distinguished from one another. Is this an adequate account of how socially embedded individuals are distinguished from one another? Does the capacity to self-impose or voluntarily impose things upon oneself in a division of labor and responsibility individuate socially embedded persons? The answer seems to be that it does by virtue of the special meaning of “self-impose.” Only individuals can voluntarily impose things upon themselves. When others impose things on individuals, from the individual’s point of view these things are done involuntarily. Thus, if individuals express we-intentions in ways that position them in social groups, they do so on a one-by-one basis or one individual at a time, and are consequently distinguished from one another in their exercise of this capacity, even though it is also part and parcel of their social embedding. Contrary to most people’s intuitions about the meaning of social embeddedness, then, it can be argued that individuals understood to be socially embedded in a collective intentionality sense are also distinguishable from one another. Moreover, since the means by which this occurs involves the exercise of one kind of self-referent behavior, combining collective intentionality and self-referent behavior provides a mechanism by which individuals can be said to influence social structure in a structure–agent model. Social psychological research, as we saw in the previous chapter, employs a selfconcept based on the idea that individual identity is reflexively constituted and reconstituted out of a socially influenced process of self-evaluation, self-imagining, and self-awareness engaged in by the individual. In the most general sense, the selfconcept is the understanding that individuals acquire of themselves as a result of any process of self-appraisal (Rosenberg 1979). This is not quite the same idea as imposing something upon oneself, since one can engage in these processes without taking on any responsibility in activities with others. However, it seems reasonable to say that engaging in processes of self-appraisal is a prerequisite for voluntarily imposing upon oneself group responsibilities, since interacting with others in a way
Individuation: collective intentionality
147
that integrates one into a set of shared activities can only occur voluntarily if one has a sense of oneself. Social psychology, then, provides the general framework for understanding individual reflexive behavior, but this general understanding needs to be complemented with an account of social processes in which this behavior might operate, if we are to speak of a capacity individuals possess to self-impose things upon themselves in interaction with others. Collective intentionality analysis provides this account of social processes in its treatment of first-person plural speech in social settings. Two of the commentators encountered earlier in this book, in fact, make the idea of individuals voluntarily imposing things on themselves in the form of making social commitments central to their accounts of what distinguishes individuals. Gilbert does so in connection with her concept of the plural subject. As she puts it, there exists a reflexive, individualizing capacity in individuals in their having a “readiness” (Gilbert 1989: 185–6) or a “willingness” (Gilbert 1996: 54–6) to be members of groups. In criticizing Lewis’ 1969 account of conventions as coordination equilibria and regularities in behavior in which individuals tend to follow precedents blindly, she argues that such an account leaves out any role for persons as distinct beings who not only make an individual commitment to participation in groups but may also employ personal standpoints, or a “personal principle model,” to determine which group rules they individually will adopt and observe (Gilbert 1996: 3, 51ff.). Taylor, whose Sources of the Self account of the origins of the atomistic conception of the individual was reviewed in Chapter 1, argues that individuals create identities for themselves by taking positions in regard to a society’s constitutive goods: “The full definition of someone’s identity thus usually involves not only his stand on moral and spiritual matters but also some references to a defining community” (Taylor 1989: 36; emphasis added). Both Gilbert and Taylor, that is, make individual commitment the means by which individuals distinguish themselves from one another. They explain individual commitment, moreover, in terms of having a conception of oneself that assists one in one’s interactions with others.
7.5 Individuation in the atomistic and socially embedded individual conceptions In closing this chapter, I draw two comparisons between the atomistic and socially embedded individual conceptions with respect to how they each address the problem of individuation. First, it is helpful to see how the two conceptions are similar and how they are different in regard to their respective understandings of human subjectivity. How they are similar is they both treat individuals as reflexive beings. Locke’s conception of the individual also employs reflexivity and a form of self-referent behavior in supposing that individuals have unique access to their own subjective states of mind or, in neoclassical terms, their own tastes and desires. Individuals can always ask themselves whether they prefer one thing to another or are indifferent between the two, and this unique access is the basis on which individuals
148
Individuation: collective intentionality
would be distinguished from one another were it adequate. However, this is a very different kind of reflexivity from that involved in the embedded individual conception in which individuals ask themselves whether to impose one responsibility or another upon themselves. In the Lockean/neoclassical view, individuals know how they are affected by their tastes and desires. Embedded individuals know how they will act with regard to participation in social groups and institutions. We might thus say that, in the atomistic conception, individuals are distinguished by virtue of a subjective inwardness, whereas, in the embedded individual conception, individuals are distinguished by virtue of a subjective outwardness. This difference deserves further emphasis. In the case of the embedded individual conception, nothing needs to be said about how individual minds work—other than that they produce intentions—to explain this form of reflexivity. Whereas the social psychology self-concept discussed in the last chapter does indeed draw on psychology, the account of individuation in this chapter only draws on the basic idea of reflexivity to explain embedded individuals in nonpsychological terms. Individuals are deduced to have a reflexive capacity to impose things upon themselves simply from the existence of shared intentions. Rather than psychology, collective intentionality analysis employs philosophy of language and a modest amount of sociological theory regarding groups. This has the advantage of locating the discussion of individuals in a public social space as opposed to a private psychology space. As Wittgenstein made clear in his critique of the notion of a private language, the risk of elaborating arguments in terms of fully private psychologies is that the idea of inner language that this requires is arguably incoherent (Wittgenstein 1958 [1953]). Economists in the neoclassical tradition may have sensed something like this when they criticized cardinal utility and the inner psychology it involved as a basis for choice.5 But whether or not they did, they have not appreciated that the idea of the individual as an atomistic being understood in terms of subjective inwardness is fundamentally problematic. While Locke was right about individuality requiring some sort of reflexive capacity, he was wrong in understanding it as subjective inwardness. The second comparison between the two conceptions in terms of how they address individuation concerns the question of necessity. Because the atomistic conception uses private access to one’s own tastes and desires to distinguish individuals, and because individuals—if they indeed have private access—always have it, individuals are necessarily distinct from one another in these terms.6 In contrast, because the embedded individuals conception relies on the idea that individuals have a capacity to impose social responsibilities upon themselves, and because they may, for various reasons, not exercise this capacity, individuals are only contingently distinct from one another. This means that embedded individuals’ relative autonomy can be an object of social policy. Institutions can evolve and be designed either to impede or to promote individuals’ exercise of this capacity. Individuals may be embedded in social groups and institutions involuntarily or at their own discretion. Arguably, the rise of social production and collective consumption since the nineteenth century
Individuation: collective intentionality
149
has made economic life more socially dense than earlier in history. If so, this suggests that the importance of whether individuals are able to exercise a capacity to selfimpose social participation upon themselves has increased, and should become an increasing focus of social-economic policy. However, a shift in focus of this kind would also need to query what conditions were required for individuals to be able to exercise this capacity on a sustained basis. That is, if embedded individuals can be distinguished from one another on this basis, can they also be reidentified as such through change? If the capacity to form social commitments and participate in groups and institutions is intermittent and episodic, it is difficult to say that embedded individuals are relatively autonomous beings. Therefore, I turn in the next chapter to the remaining question of whether individuals thus understood can also be reidentifiable through change.
8 Reidentification Capabilities
Ruby said there were many songs that you could not say anybody in particular made by himself. A song went around from fiddler to fiddler and each one added something and took something away so that in time the song became a different thing from what it had been, barely recognizable in either tune or lyric. But you could not say the song had been improved, for as was true of all human effort, there was never advancement. Everything added meant something lost, and about as often as not the thing lost was preferable to the thing gained, so that over time we’d be lucky if we just broke even. Any thought otherwise was empty pride. (Frazier 1997: 380) Individuals cannot be located by a single set of coordinates, because they operate in many different collective dimensions, within many different chosen and given groups. Nor can they be located by a list of all the given groups to which they belong, by a simple “adding up” of separate positions. The interaction between different dimensions of collective identity affects the choices individuals make about which collective interests to pursue. (Folbre 1994: 52–3) Acting freely and being able to choose may be directly conducive to well-being, not just because more freedom may make better alternatives available. (Sen 1993: 39)
Chapter 3 investigated the reidentification problem in connection with Becker’s time allocation model of the individual. Becker’s model is important, both because of the influence that human capital theory has in and outside economics and because human capital theory provides one of the few accounts in neoclassical or mainstream economics that explicitly raises the issue of how individuals are changed through time as a result of the actions they undertake. This latter feature of the model makes it possible to raise the reidentification problem for the atomistic conception, and to do so in a such way as to bring out important but largely unexamined assumptions regarding the nature of ability and the influence of education on individuals that underlie that conception. Chapter 3 argued that atomistic individuals are not successfully reidentified across change in the model because the assumptions about
Reidentification: capabilities
151
ability and education that the model makes turn out to be implausible. In this chapter, the embedded individual conception is put to the same reidentification test in connection with the issue of how embedded individuals are changed through time as a result of the actions they take. This investigation asks whether there are unexamined assumptions in this conception that also fail to stand up to reasonable scrutiny, and that preclude the reidentification of socially embedded individuals. In spite of the similarity in the two investigations, there are important differences between the two conceptions of the individual in terms of how they treat change in the individual across time. The embedded individual conception, I will argue, does not focus upon a process of human capital investment and embodied capital accumulation, but builds on the reflexive individuating capacity discussed in the last chapter to shift attention to how individuals develop human capabilities across change. It might seem that individuals’ investment in human capital and their development of human capabilities concern essentially the same subject, since both frameworks seek to explain how individuals develop their skills and abilities. However, this surface similarity does not withstand closer examination. Chapter 3 demonstrated that the concept of ability is not well formulated in the time allocation model, whereas the concept of ability has been given more extensive attention in the capabilities framework. Moreover, the main orientations of the two approaches are quite different (cf. Sen 1999: 292ff.). The human capital framework explains investment as response to change in prices in light of individual tastes and desires, while the human capability framework—in Sen’s interpretation—explains capability development as an enlargement of individual “functionings,” particularly as facilitated by expanded entitlements. The human capital framework, moreover, involves an essentially affective reaction to market signals, while the human capability framework primarily involves a policydriven strategy of increasing individuals’ space of functioning activity. The kinds of change that the two conceptions of the individual address are also quite different. In the case of the atomistic individual time allocation model, change occurs in the individual’s internal make-up as some preferences are enhanced or heightened by investments that embody capital in the individual. In the case of the embedded individual conception from the last chapter, change occurs both within and across the individual’s social group affiliations and institutional locations. Thus, whereas the issue facing the atomistic individual conception is whether individuals distinguished in subjective terms continue to be distinguished in subjective terms as they engage in human capital accumulation, the issue facing the embedded individual conception is whether individuals distinguished in terms of a capacity to impose tasks and responsibilities upon themselves continue to be distinguished in this way as they develop capabilities associated with their changing involvement in groups. Recalling the internalist–externalist dichotomy between approaches to the individual, in the atomistic conception, change occurs in individuals’ “internal” space, while, in the embedded individual conception, change occurs in individuals’ “external” space.
152
Reidentification: capabilities
Section 8.1 briefly outlines Sen’s capability framework, distinguishing his different concepts of individual advantage, and drawing attention to the emphasis he places on the role of freedom and real opportunities to explain his preferred understanding of individual capabilities. Section 8.2 turns to the problem of whether individuals can be reidentified across change in their membership in groups. An important dimension of this discussion is how we see the increasing complexity of the social world. Does it overwhelm and “oversocialize” individuals, or do they learn to manage its requirements, thus maintaining a relative autonomy? To make the best possible case for the latter result, I formulate the reidentification test for the embedded individual in terms of Sen’s understanding of capabilities in the real opportunities sense. However, upon examining a key implicit assumption involved in doing so, I conclude that the embedded individual conception does not pass the reidentification test. According to a reasonable view of the way the world works, socially embedded individuals are not reidentifiable as relatively independent agents across change in their social group associations. Section 8.3 then asks why the embedded individual conception fails the reidentification test as opposed to why the atomistic conception fails the test, and then discusses the status of the former as an ideal. Section 8.4 briefly compares Sen’s thinking about the individual and the embedded individual conception. Section 8.5 concludes with a discussion of a key reason why these two traditions of thinking about individuals are so different in character.
8.1
Sen’s capability framework
Sen’s work is usually associated with more traditional views of the individual and not with the embedded individual conception. Indeed, with its focus on individual advantage, it might not seem a promising framework to apply to the questions at hand. Nonetheless, a case can be made for saying that Sen’s work is sufficiently at odds with neoclassical and mainstream economics to provide adequate resources for the current discussion. Much of Sen’s work has been devoted to criticizing the traditional understanding of the individual as a utility-maximizing agent, especially in terms of how this understanding frames standard welfare economics and normative reasoning in economics generally (cf. Pressman and Summerfield 2000). His longstanding interest in poverty and basic needs, particularly in connection with economic development, is contrary to the standard view that wants and needs cannot be systematically distinguished, and that the latter have no place in economics. He has consistently defended interpersonal comparisons, which have been off limits for most economists since Robbins. Perhaps most importantly, the concept that ties his entire capability framework together—freedom (cf. especially Sen 1999)—has no real equivalent in neoclassical and mainstream economics. In fact, his focus on freedom provides a normative framework alternative to efficiency analysis. An early insight of Sen’s was that individual choice and well-being need not always be conjoined, as had been generally assumed in neoclassical economics in its charac-
Reidentification: capabilities
153
terization of individuals as utility-maximizing agents (Sen 1970, 1977). One of Sen’s critiques of welfare economics is that it is “welfarist,” meaning that overall states of affairs are judged solely according to individual utilities. However, individuals pursue a variety of goals in life besides that which immediately serves their own interests and well-being, reflecting such things as their commitments, their sympathies for others, and their desire to freely direct their own lives even though doing so may not always produce the best results for themselves. Recognizing the diversity of individual goals led Sen to make two distinctions regarding the nature of individual advantage and motivation that ultimately came to underlie four relatively independent forms of valuation and ways in which he believes his capability framework can be applied to the world (Sen 1993: 35ff.). First, since a person’s goals may concern matters distinct from that person’s own well-being, it is important to distinguish between a person’s pursuit of individual well-being and a person’s pursuit of overall agency goals, where this latter concept includes all the other reasons that motivate individuals to act in the ways they do. Here, the term “agency” is not being used as in neoclassical principal–agent analysis, but rather in its “older—and ‘grander’—sense as someone who acts and brings about change, and whose achievements can be judged in terms of her own values and objectives” (Sen 1999: 19). Second, since being free to pursue a variety of goals is valuable to people, it is also important to distinguish between the freedom to achieve one’s goals, whatever their nature, and the freedom to simply pursue one’s goals. This latter distinction, we will see, is at the core of Sen’s understanding of his capability approach, and represents in his own view one of his principal departures from neoclassical economics. These two distinctions yield four sometimes overlapping but relatively distinct concepts of individual advantage for Sen (1985): 1 2 3 4
Well-being achievement. Agency achievement. Well-being freedom. Agency freedom.
The first represents the traditional concern of neoclassical economics in individuals’ interest in satisfying their own preferences. The second, as noted, concerns individuals’ ability to achieve goals that need not involve their own well-being. The third concerns individuals having the freedom to pursue their own well-being. The fourth concerns individuals simply having the freedom to pursue all their goals, whether or not they are successful in achieving them. Against the charge that this multigoal framework is insufficiently determinate for social policy analysis, Sen emphasizes the flexibility it provides in being able to address the great variety of different types of valuation problem that social policy confronts. For example, when considering governmental policies regarding individuals’ well-being, such as in connection with
154
Reidentification: capabilities
old-age pension, well-being achievement may be a more relevant standard than wellbeing freedom. In contrast, well-being freedom may be more relevant than wellbeing achievement when policies regarding individuals’ lifetime occupational choices are concerned. In the first case, government might wish to guarantee certain types of outcomes for individuals, while in the second case government might wish to assure that individuals have certain kinds of opportunities to develop their skills and talents. In Sen’s view, a multigoal framework is needed to make these sorts of distinctions, and this argues against having one standard to apply to all cases.1 Applying the capability framework to these four cases involves recognizing how different functionings and capabilities pertain to each case. Functionings are defined in an Aristotelian manner as the “beings and doings” of people—the different things that people can be and do. Capabilities are defined in two ways: either as realized functionings—what a person is actually able to be and do, or as the set of alternatives an individual has—the person’s real opportunities for being and doing different things (Foster and Sen 1997 [1973]: 199–200). Thus, were we, for example, to ask how the capability framework applies to the question of poverty, we would focus on well-being achievement, set forth a set of basic needs regarding such things as nutrition, shelter, literacy, health care, etc., characterized as “crucially important functionings” that individuals need in order to escape poverty, and then put this all in terms of a set of “basic capabilities” to be realized rather than simply constitute real opportunities (Sen 1980). Indeed, it is this particular application of the capability framework and of the meaning of the concept of capability that has probably gained widest attention in its use to construct the United Nations Human Development Index (HDI) which is calculated and reported annually for countries across the developing world in the United Nations Development Program Human Development Report.2 However, other applications of the capability framework do not refer to “basic capabilities,” nor need they make use of the HDI treatment of capabilities as realized functionings. In fact, equating the value of a person’s capabilities to a realized set of functionings is termed “elementary evaluation” by Sen, because it represents the simplest way of evaluating capabilities (Sen 1993: 38). Essentially, that a person achieves a chosen set of functionings demonstrates that that person has the capabilities needed to achieve those functionings. However, if we also place value on having opportunities that are not taken up, a person’s capabilities then need to be seen as the range of alternatives they have, even if none of these alternatives would have been preferred. “Acting freely and being able to choose may be directly conducive to wellbeing, not just because more freedom may make better alternatives available” (Sen 1993: 39). Freedom, that is, is something that matters in itself, and is valuable irrespective of its contribution to good consequences. Thus, “[t]he ‘capability approach’ can be used either with a focus on what options a person has—given by the capability set—or by the actual functioning combination she chooses—given by the chosen functioning vector” (Foster and Sen 1997 [1973]: 200). The options focus, Sen emphasizes:
Reidentification: capabilities
155
is contrary to the one typically assumed in standard consumer theory, in which the contribution of a set of feasible choices is judged exclusively by the value of the best element available … . Even the removal of all the elements of a feasible set (e.g. of a “budget set”) other than the chosen best element is seen, in that theory, as no real loss, since the freedom to choose does not, in this view, matter in itself. (Sen 1993: 39)
Accordingly, when the capability framework is used to emphasize freedom—or agency freedom as Sen characterizes it in the context of four ways in which individual advantage can be understood—we are able to take note of “counterfactual” situations that individuals may value investigating even if they do not advance well-being. One of Sen’s favorite examples to illustrate this concerns fasting. Fasting is not just starving, but starving because one has rejected the option of eating. One consequently would not want to treat fasting as a problem of hunger, at least not in the way one would treat hunger as a problem brought about by famine. Making this distinction depends upon understanding capabilities in the options or real opportunities sense which allows for counterfactual possibilities. The individual who fasts could avoid hunger. At the same time, individuals facing starvation because of famine lack the opportunity to eat. We cannot decide how to address these different situations if we operate with the more restricted interpretation of capability as realized functionings. That more limited sense of capability has value in particular contexts, such as when the issue is how to address poverty as a welldefined problem of specific unmet needs. However, in more complex, open-ended contexts, in which even evaluation issues are often not sharply defined, it seems a mistake to eliminate at the outset important information about individuals regarding their real options. Indeed, in the human world, it seems that counterfactual possibilities are a fundamental dimension of life. This may be especially true of individuals’ membership in groups, since the complexity of their interaction in groups presumably creates numerous possibilities that are never pursued, but that individuals might want freely to investigate. It is this idea—that negotiating a complex social world may depend upon a capability for free investigation—that is taken up in the following section in addressing whether individuals distinguished from one another in terms of their capacity to impose responsibilities upon themselves can be reidentified across change in group membership.
8.2 The capability framework applied to the reidentification of embedded individuals Here, I first set up the reidentification problem for the embedded individual conception in a manner parallel to the way the problem was posed for the atomistic conception. Second, I ask why should we be pessimistic about the possibility of reidentifying embedded individuals. Third, I propose applying Sen’s real opportunities sense of capabilities to embedded individuals’ social group experience to make what I regard as the best possible case for individuals’ relative autonomy.
156 8.2.1
Reidentification: capabilities The reidentification problem
To see the nature of the problem, first recall why reidentification failed in the case of the atomistic individual time allocation model. The time allocation model extends the basic pure preferences view of the individual by allowing that some preferences are heightened or enhanced as a result of investments in human capital. Individuals are still defined in terms of preferences, but are no longer reidentifiable in terms of their preferences, because their postinvestment, heightened preferences reflect factors external to those preferences in connection with the social determination of ability and the influence of education on opinion. The individual is transformed in nature by human capital investment, and accordingly ceases to be distinguishable merely as a collection of preferences. That is, the preferences-based definition of the individual does not stand up to a process of change which that conception itself employs. The embedded individual conception, in contrast, defines individuals as distinct in terms of their each having a capacity to self-impose group participation upon themselves. However, since collective intentionality analysis explains how individuals use “we” language in groups, and since the use of “we” language is continually changing, individuals’ membership in groups must always be changing. Over time, individuals enter new groups and leave others, and, just as importantly, when they remain in groups their roles within those groups often change. At the same time, institutions themselves are also continually in a state of change, so even a relative stability in individuals’ group affiliations is likely to involve change in the character of their participation in these groups. In Tuomela analysis, since different groups operate according to different combinations of rules and norms, and since these different combinations of rules and norms imply different task–right combinations for individuals, individuals’ continual change in group associations implies a continual change in the rules and norms they observe and their associated sets of task–right combinations. Thus, is the individual transformed in nature by this process of change? 8.2.2
“Structures of constraint”
How might this continual change in group associations affect individuals? Many, particularly in sociology, have argued that individuals’ experience in groups or social life tends to reduce individual autonomy, whether understood as it is here or in other ways. Indeed, this conviction lies behind the social science critique of the Enlightenment view of the individual discussed in Chapter 1 which sees individuals as primarily concerned with “social identification.” An especially clear way of putting this is to say, adapting Granovetter’s terminology (1985), that socially embedded individuals become “oversocialized” over time so that they cease to function as independent, distinguishable individuals, and instead become enmeshed (in the constraining sense of the term) in their ever-changing social relationships. That is, rather than be able to exercise a reflexive capacity to self-impose social group require-
Reidentification: capabilities
157
ments upon themselves, others impose social requirements upon individuals. In effect, the experience of continually moving from one group membership to another over time is thought to accustom individuals to passive acceptance of whatever social requirements their roles in groups entail. Thus, though collective intentionality analysis, philosophically speaking, might be able to distinguish individuals from one another in terms of the idea of their possessing an individuating reflexive capacity to self-impose responsibilities upon themselves, in this widely held view of society’s oversocialization effects, in fact most individuals over time lose or never develop this capacity, and are consequently not reidentifiable as distinct individuals across change. Add to this picture the possibility that the social world is also becoming increasingly complex. In many social settings, individuals now seem responsible for everlarger sets of task–right combinations which are governed by multiple sets of rules and norms that may conflict, overlap, and be vague enough often to be in need of interpretation. Thus, when individuals experience continual change in their group and social involvements, one complex set of task–right responsibilities is replaced by another, which is itself replaced by another, which is in turn replaced by another, and so on. Part of what is understood by saying that individuals become “oversocialized” over time would then be that they simply become overwhelmed by an increasingly dense social world. Folbre in fact regards this as essentially the kind of world we live in, and characterizes it as one in which “structures of constraint” dominate individual experience. As she describes it, “individuals are so embedded in a complex structure of individual and collective identities and competing interpretations of these that sometimes they do not even know whose interests they are acting on” (Folbre 1994: 16). The difficulty is that individuals are simultaneously affiliated with groups and in motion between groups, so that their “collective identities” are both in competition and in a state of constant flux: Collective action helps individuals constitute their own identity. At the same time, individuals have to mediate many conflicting forms of loyalty, and their behavior can seldom be explained simply by their “objective” position or membership in a group. (Folbre 1994: 28)
The question for Folbre is how do individuals maintain and manage their collective identities without also becoming lost in them? This is essentially the issue of whether socially embedded individuals are reidentifiable across change. 8.2.3 Using Sen’s real opportunities sense of capabilities to explain reidentification A possible answer to Folbre’s question, I suggest, lies in applying Sen’s real opportunities understanding of capabilities to the capacity individuals have to voluntarily impose group participation upon themselves (Davis 2002c). The idea of a capacity, as
158
Reidentification: capabilities
I used the term in the last chapter, is the idea of a power that may or may not be exercised. In contrast, the idea of a capability, as Sen understands it, is the idea of a power that has been developed and cultivated so as to be susceptible to regular exercise. One thus develops certain capabilities when one has the capacities for doing so, and also when there exists social arrangements that support this development, such as a system of entitlements regarding the use of individual and social resources. Consider, then, the capacity to voluntarily impose group participation upon oneself. In an “oversocialized” type of world, individuals are unable regularly to exercise this capacity, and accordingly do not develop the capability for doing so. To be more specific, failing to develop this capability means that individuals do not become adept at mastering new and changing combinations of tasks and rights across new and changing group affiliations, fail to recognize how one group participation resembles or differs from another, and in general move inflexibly across their different social involvements—a circumstance compatible with being “oversocialized” and passively accepting assignments dictated by others. Using Sen’s real opportunities sense of capabilities to understand how individuals might develop a capability for participating in groups and also for moving across their many group affiliations is then a matter of recognizing what may be involved in individuals being able to consistently master new task–right combinations. In a complex social world in which old routines and procedures are often either irrelevant or need to be adapted, mastering new task–right combinations depends upon individuals being able to explore how their actions can be best organized. One way of understanding this is to say that they need to be able to freely consider counterfactual possibilities associated with those new combinations in order to master them. Basically, in complex situations, individuals need to be able to consider how things might be done in order to figure out how things ought to be done. The best way of doing something, that is, is often only arrived at after considering other options that by comparison are subsequently seen to be less advantageous ways of doing things. How, then, are socially embedded individuals reidentified? The answer is that, should individuals succeed in developing the capability for freely participating in and moving across their various social group affliliations, this capability would constitute the basis on which they would be reidentifiable as relatively independent individuals across this process of change. On the understanding of capacity and capability above, developing this capability is equivalent to learning how to consistently exercise the individuating capacity that people inherently possess to freely impose social group participation upon themselves. However, if individuals learn how to consistently determine the terms on which they participate in groups, they thereby assure themselves of always participating in groups on a relatively autonomous basis. Thus, acquiring this special kind of capability would allow them to be reidentified across change in their group involvements as relatively independent beings. Note, however, that this conclusion depends upon our being justified in using the second rather than the first of Sen’s two definitions of capability development.
Reidentification: capabilities
159
Would the same conclusion obtain were we to employ the first of Sen’s capability definitions? Sen’s real opportunities sense of capability involves individuals freely exploring their options, whereas his realized functionings sense of capability involves individuals simply developing their skills and abilities without any implication regarding their freedom in doing so. Human capital theory employs this latter sense of capability, because it employs “elementary evaluation” in equating the value of a person’s capabilities to a realized set of functionings. Nothing in this analysis requires that individuals freely explore their options en route to this outcome. Similarly, if we think that individuals negotiate their changing social affiliations by developing capabilities understood in the realized functionings sense, nothing in this analysis involves their freely exploring how they would take on or reject social and group responsibilities. However then, the key to saying they remain relatively independent beings— that they freely determine the terms of their involvement in groups—is no longer part of the analysis. That is, this particular understanding of capability does not support the idea that individuals can be reidentified across change in this way, because it does not include the idea of individuals being able to exercise the individuating capacity freely to self-impose group participation upon themselves. Therefore, were this realized functionings interpretation of the concept of capability the appropriate one, embedded individuals would not be reidentifiable across change in their social involvements. Does the socially embedded individual framework, then, pass or fail the reidentification test? Formulating this test for the atomistic individual conception helped bring out assumptions implicit in human capital theory regarding the nature of ability and education. Formulating the test for the embedded individual conception helps bring out the implicit assumption in this conception that capability needs to be understood in the strong, real opportunities sense for individuals to be reidentifiable. Is this a plausible assumption? Do individuals, as a matter of fact, tend to develop capabilities in this strong sense across their changing experience in social settings? Or is their capability development better described along human capital lines in terms of their cultivation of realized functionings? A fair response, I think, is that many, perhaps most, individuals in the world today are not able to acquire this stronger kind of capability. Rather, their capability development is more narrowly guided in such a manner as to leave them little discretion in how their capabilities develop. In Folbre’s language, then, most people today see group associations as “structures of constraint,” and find themselves unable freely to move between and negotiate their multiple social responsibilities. Most people do not find themselves increasingly skilled in managing their conflicting social circumstances, but rather feel themselves increasingly overwhelmed by them. They do not self-impose the terms of group participation upon themselves, but instead generally find it imposed upon them by others. They are for the most part, as Granovetter puts it, “oversocialized.” Thus, our conclusion must be that the embedded individual conception also fails the reidentification test. The standard for success is, as it were, too high for the world we live in.
160
Reidentification: capabilities
8.3 The embedded individual conception as an ideal conception Both the atomistic and embedded individual conceptions accordingly fail the reidentification test needed for saying that individuals are relatively independent agents. Yet there is an important difference between the ways in which they each fail. With the atomistic conception, the test fails because individuals cannot in principle be reidentified in subjective terms across change. Basically, the entire project of defining individuals subjectively—already questionable in the pure preferences view—breaks down when preferences are seen as heightened through socially influenced human capital investments. With the embedded individual conception, however, the reidentification test fails not because that conception breaks down, but because, as a matter of fact, it does not reflect the real-life experience of most people in the world today. That is, reidentification fails because of the way the world happens to be organized, not because of any inherent difficulty in the idea that individuals might maintain a relative autonomy by developing a capability for freely participating in their multiple social involvements. Indeed, that individuals can be thought reidentifiable in this way follows from the fact that, while most people in the world experience social settings as “structures of constraint,” there nonetheless are also individuals in the world today who can be described as being generally successful in maintaining a relative autonomy in social group circumstances, precisely because they have developed capabilities for determining the terms on which they participate in social groups. These latter individuals, in comparison with the great majority of people, are regarded as privileged with respect to freedom and relative autonomy. Their existence thus functions as a standard for others, and this implies that the embedded individual conception is partly descriptive and partly normative as a conception. That is, the embedded individual conception as understood here primarily explains how individuals could be distinct and reidentifiable, and as such it functions as an ideal conception. Interestingly, many rational choice theorists see the same sort of combination of descriptive and normative content in the “rational economic man” idea. In the face of much experimental evidence that real-world individuals often do not behave rationally, the theory of rational choice is characterized as a theory that explains how individuals ideally ought to behave in order to be rational (Sugden 1991). This implies that the atomistic individual conception also functions as an ideal (albeit, according to the arguments here, as one that is flawed in principle). If rational choice is the ideal, then the conception of individuals who engage in it involves an ideal conception of the individual. How are we to look upon ideal conceptions in economics? The traditional view of the matter is positivist in drawing a firm line between normative, value-laden discourse and descriptive, value-free discourse. Good science operates on the latter terrain, ideals have no place in science, and values are only brought in from outside science in the formulation of policy. By this token, science is neutral with respect to
Reidentification: capabilities
161
values, and good explanation in science never biases policymakers towards one set of values or another. This view has supporters among the ranks of both orthodox and heterodox economists. In the case of the latter, this has arguably been manifest in a preference for holist explanations. If, in fact, individuals are generally embedded in institutions in ways that minimize their character as relatively autonomous agents, then a purely descriptive approach should take supraindividual entities as its object. Ironically, orthodox economists, who have been the most steadfast in defending the positivistic divide between facts and values, while also claiming to be engaged in value-free descriptive science, have generally taken an ideal state of affairs—rational choice—as their object.3 Moreover, contrary to the idea that good science should never bias policymakers towards one set of values or another, the preferences-based utility conception of the individual itself has also been narrowly interpreted on questionable psychological grounds to rule out normative judgments involving interpersonal utility comparisons and consequentialist arguments that go beyond efficiency reasoning. However, the lesson in this, it seems, is not that we ought to work harder to banish ideal conceptions from our explanations in order to be more determinedly value free. Policymaking, which draws upon ideals, tends to influence the world by restructuring it in line with its ideals. Science thus examines the world as influenced by ideals, and explanations of the world thus exhibit ideals when they describe the world (cf. Dolfsma 2002a). Of course, policymaking also has many unintended consequences, and the extent to which it is effective in influencing the world is controversial. However, it would be naive to suppose that unintended consequences eliminate all systematic influences that policy has, or to think that there is no imprint of ideals on the world. Take a strong case. The dominance of free market, gainsfrom-trade/efficiency policy recommendations in many contemporary economies has increased the importance of markets, and consequently created incentives for individuals to adjust their behavior to the logic of markets. Were economic policy in these economies driven by goals such as social provisioning, human flourishing, and freedom, social behavior would presumably be different. Thus, policymaking tends to actualize ideals in the world, and, especially in periods of policy dominance, actual behaviors and policy-conceived ideal behaviors increasingly converge. It can also be argued that ideals are actualized in the world through nonpolicy avenues. The argument of mainstream economics that rational behavior is essentially self-interested behavior is well known and not without influence in contemporary society. However, this argument is generally advanced independently of policy recommendation as part of the explanation of how markets work. At the same time, most choice theorists today also accept that rational choice behavior is an ideal, and that choice theory is normative in stating how individuals ought to behave in order to be rational. One could argue that rational choice theory should be given up in favor of a more behavioral approach. But one would need to address the mainstream defense (widely held since Friedman’s methodological pronouncements) that the logic of markets over time compels individuals to behave rationally and self-interestedly.
162
Reidentification: capabilities
From this perspective, it is appropriate to describe rational behavior in ideal terms, because that ideal behavior is in a process of being realized in market economies. Descriptive explanation, then, cannot avoid making use of ideals, irrespective of the impact of policy. Thus, consider again the embedded individual conception as an ideal. On one level it has the status of an ideal because it fails to describe the social circumstances of most people. On a second level, however, it has the status of an ideal because it operates as a standard that has influenced and may continue to influence social-economic policy aimed at increasing the number of people in the world who are relatively autonomous and yet socially embedded. Yet, were policies with this aim increasingly adopted, the embedded individual conception would then begin to actually describe more individuals in the world, so that individuals might more generally be reidentified as having developed capabilities for freely participating in their social involvements. The point here, however, is not suddenly to substitute utopian reasoning for the identity analysis that this book has employed. The point is rather that this identity analysis is complicated by the interaction of policy and explanation, and this requires that we see the embedded individual conception as functioning on two levels: as a pure ideal and also as a potentially realizable state of affairs. Another way of making this point is to recall what the structure–agent framework implies about the agency of embedded individuals. Agents influence social structures in a variety of ways, one of which is through the formulation of policy in response to an evaluation of structures that then influences those structures. In this kind of framework, the relationship between social structures and individual agents is best understood dynamically, since, as they mutually influence one another, neither ever achieves a determinate, final state. However, this means that ideals can gain or decline in significance over time. If certain ideals are influential, social structure increasingly reflects or actualizes them; if they are not influential, they exist more as pure ideals. It is interesting, then, that proponents of the structure–agent framework see it as an alternative to holist explanations, which, without the structure–agent framework, would have constituted for many a preferred form of explanation. In effect, their view seems to be that the grounds for including individual agency effects on social structure are strong enough to abandon monocausal holism. Put in terms of the embedded individual conception, their judgment is that individuals appear to have acquired sufficient autonomy from the social structures in which they act to be able to influence those structures. Put in terms of the status of the embedded individual conception as an ideal, for structure–agent proponents this conception seems less a pure ideal and more one increasingly susceptible to actualization. Consider, then, that much of interwar American institutionalism was closely linked to the progressive liberal reform movement, and often justified its critical examination of business and market institutions of the time in these terms (Rutherford 2000: 295). Institutionalists pioneered legislation in public utility regulation, unemployment insurance, worker’s compensation, public health insurance,
Reidentification: capabilities
163
and Social Security, all of which were arguably aimed at reducing the “structures of constraint” that operated on individuals embedded in business and market institutions. Thus, while on the one hand the idea that individuals might exercise a relative autonomy in economic life was little more than an ideal, on the other hand the theoretical and practical efforts of institutionalists contributed to the actualization of that ideal. Or consider feminist economists in the 1970s and 1980s who critically reexamined Chicago School “new home economics” thinking about household production, marriage, divorce, fertility, and the gender division of labor (Seiz 2003). Exposing the circularity of arguing that “women specialize in housework because they earn less in the labor market, and they earn less in the labor market because they specialize in housework” (Ferber and Birnbaum 1977: 20), they gave new impetus to discrimination law as a practical strategy for breaking the “structures of constraint” upon women of neoclassical economics. They thus drew upon an ideal conception of women able to exercise a relative autonomy in moving between home and work, and helped change social policy to actualize that ideal.4 One longstanding principle in heterodox economics, dating at least to Marx, is to use theory to change the practical world. Moreover, the perspective often taken on this transformation is broadly “liberationist” in seeking to replace structures of “unfreedom” with structures of freedom. The value of freedom, of course, has a long history in the development and emergence of democratic society, but my argument here is rather that the value of freedom has a particularly important role in the embedded individual tradition. In contrast to the orthodox tradition, heterodox thinkers generally begin from an understanding of the way society functions as a whole, and this often has the effect of bringing to the fore system-wide relationships between groups or classes of individuals. In the heterodox tradition, then, not only is individual freedom juxtaposed to “structures of constraint” as in the liberal tradition, but just as importantly the well-being of those who are privileged in escaping those “structures” is juxtaposed to those who are unsuccessful in this regard. This has the effect of shifting the focus on freedom to system-wide change, and helps give the heterodox tradition its “liberationist” character. It also gives the tradition its methodological orientation on the status of ideals, producing a more complex view of fact and value than is available in positivism. Reasoning dynamically, heterodox thinking judges that, if most individuals are not self-determining, that some are creates an agenda for replacing a pure ideal with one that can be made real.
8.4 Sen’s thinking about the individual versus the embedded individual conception Sen’s thinking about the individual is capabilities-based rather than preferencesbased. However, the embedded individual conception advanced here is different from Sen’s understanding in a number of respects. First, Sen’s thinking originates within the liberal tradition of ideas about individuals and society that takes individuals
164
Reidentification: capabilities
as pregiven independent agents. The embedded individual conception, however, as its name implies, sees social context as central to understanding individuals. Thus, whereas a challenge to the liberal tradition is to understand society’s effects on individuals, a challenge to embedded individual thinking is to understand individuals’ relative autonomy and influence on society. Second, Sen does not actually have a theory of the individual. His resistance to the idea that thinking about human capabilities should be structured around a single view of human flourishing demonstrates that his goal is not to produce a theory of the individual, but simply to contribute to thinking about individual capabilities. Moreover, what has been employed in this and the last chapter to produce a theory of the embedded individual has no place in Sen’s thinking. That is, Sen makes no use of collective intentionality analysis, does not emphasize individuals having a reflexive capacity to self-impose social group participation upon themselves, and does not focus on developing a particular capability for doing so across change in one’s social group involvements. There is, nonetheless, important shared ground between Sen’s thinking and the embedded individual conception here. First, his emphasis on freedom is close to the idea that individuals have a relative autonomy when they determine the character of their participation in groups. The difference is that he sees an expansion of freedoms in the form of enhanced capabilities as a human flourishing based on freedom having intrinsic value. On the argument here, however, developing a capability for freely negotiating one’s social involvements across a lifetime is instrumental in maintaining the status of being relatively independent as an individual. It is the latter that has intrinsic value. Simply claiming that freedom has intrinsic value does not guarantee the integrity of the individual per se. Second, Sen’s thinking and the embedded individual conception share opposition to the preferences-based vision of the individual. Both hold that Homo economicus could engage in rational choice and yet not act freely. Both hold that “free” choice in the preferences-based conception amounts to affective response to market stimuli, and that the standard view lacks any understanding of individuals as active beings.5 Third, there is also shared ground between Sen and the embedded individual conception in terms of the attention placed on relations to others. Though Sen does not see individuals as socially embedded, his treatment of how individuals exercise their capabilities is dependent upon seeing them as interacting with others. Just as thinking of individuals in terms of their preference turns the investigation away from social interaction, thinking of individuals in terms of their capabilities turns the investigation towards social interaction.
8.5 Concluding remarks on the atomistic and embedded individual conceptions Here, I close by briefly addressing a central reason why these two traditions of thinking about individuals are so different in character. The identity analysis employed in this book might be thought to affect the two theories differently in that,
Reidentification: capabilities
165
while both tests can be formulated for both theories, it might seem that the reidentification test ill-fits the atomistic conception. Indeed, without human capital theory, which does permit a treatment of the individual through time, the issue of reidentification of individuals might seem to be beside the point in neoclassical economics. Since most of neoclassical theory is based on counterfactual comparative statics analysis, being able to reidentify individuals is not a concern, since the only question is what choices would individuals make were prices and income different. However, this is not a very satisfactory response, since implicit in the idea that individuals would make different choices in different circumstances is the idea that the same individuals are involved across different circumstances. Otherwise, comparative statics would not be comparative! Furthermore, implicit in the idea of making any kind of policy recommendation is the idea that the same individuals are around to benefit from policy improvements. Thus, unless the neoclassical model is interpreted in an entirely formal and agnostic manner regarding the properties of objective functions, the problem of reidentification still applies to it. The more interesting issue, it thus seems, is why is it difficult to apply reidentification identity analysis to the atomistic individual conception. My argument thus far has emphasized the limiting nature of the atomist conception as subjectivist. In a natural world, however, reidentification is not just about the survival of minds; it is also about the survival of minds-in-bodies, as indeed much of the philosophical literature on personal identity attests. Accordingly, here I draw attention to a further feature that differentiates the atomistic and embedded individual conceptions, namely, their entirely different views of time. The latter view, pitched as it is at the level of society, is an implicitly historical analysis. Virtually every discipline that makes society its object of investigation admits a time frame of reference. In contrast, when individuals are understood subjectively, they are removed both from their bodies and from the world of bodies, namely society. This chasm between seeing the individual as mind alone and the fact that minds exist in a world of bodies helps explain why the time allocation human capital model of the individual is so awkward on central issues. It is interesting, nonetheless, to speculate over how a view of the individual as bodiless mind could emerge in human history, which is very much a long record of physical suffering. For this I go back to Chapter 1 and Taylor’s account of the emergence of modernity’s theory of the individual. Under the influence of Enlightenment science, Descartes expelled God from nature to give space for mechanical explanation. However, he did not naturalize the soul—that job was left to Locke—and he preserved a role for God in constituting human reason in the form of divine guarantee of clear and distinct ideas. Thus, the human mind remained the human soul. That is, the mind’s fundamental character was as a bodiless and transcendent entity. Locke brought the fruits of empiricism to bear on this conception, but he retained Descartes’ divine image of mind. Such a mind is out of time, not in it, since it transcends nature, which is in time. How did this story turn out? Neoclassical economics
166
Reidentification: capabilities
gave way to mainstream economics when subjectivism was abandoned. However, the new orthodoxy still lacked the resources for understanding individuals in time. Formalism, accordingly, filled the gap, and Descartes’ timeless mathematics was given a new domain of operation. Heterodox theory, however, does not originate in the vision of Descartes and Locke. It begins in the nineteenth century with increasing awareness of the impact of industrialization on society, and thus takes society as its object. Society understood from this perspective is an historical object. Its passage through time is irreversible and cumulative. Individuals may also not count in this story, since, barring an anchor in timelessness as eternal souls, in nature individuals are constantly being eliminated and replaced. Small compensation would be if their short stay could be made less painful on the model of those few who are relatively privileged. A further extension of this privilege might then be achieved by recognizing realistic prospects for changing the structure of society, and then gradually adjusting policy and institutions to one another. However, this difficult project takes time.
9 Before the fall
Value in economics
It is neither that “the good” comes first, and then “rights and duties”, nor that rights and duties congeal first followed by the good, but that they are linked concepts that demand simultaneous consideration. (Sen 2001: 61)
Chapter 5 ended Part I with the argument that the subjective conception of the individual in neoclassical economics was followed by the cognitive science abstract individual conception in early mainstream economics, which was in turn followed by a loose collection of ad hoc ideas about individuals in subsequent mainstream game theory, bounded rationality, experimental economics, behavioral economics, evolutionary economics, new institutional economics, and other recent strategies. My general argument was that the way in which the problematic character of the subjective conception of the individual was addressed in neoclassical economics—that is, by eliminating subjectivism—produced in early mainstream economics the abstract individual conception, and that the latter, itself highly problematic as a cognitive science account of individuals, gave rise to the multiple directions of current mainstream economics, none of which offers a theoretical or systematic view of the individual. This chapter ends Part II, and provides an evaluation of the direction of development of heterodox economic thinking regarding the socially embedded individual conception. The focus of this chapter, however, is different from Chapter 5 by virtue of the very different origins of this conception. First, whereas orthodox views of the individual, both neoclassical and early mainstream, are internalist and originate in thinking about mind, heterodox economic views of the individual are externalist, and originate in thinking about society. Neoclassicism understands mind as a collection of desires or preferences, and early mainstream theory treats the mind as a symbolprocessing computer. Heterodox thinking originates in nineteenth century holist views of the individual and society, and undergoes development in the twentieth century in the form of structure–agent thinking. Second, orthodox economic thinking tends to be positivist (in spite of its implicit reliance on ideals), and argues that values should be located outside explanation. Its emphasis on desires and preferences and on
168
Before the fall: value in economics
computation is naturalist and formalist. Heterodox economics sees values at work in social behavior, and includes values in explanation. Its emphasis on society and social interaction produces historical explanation that includes social values. This chapter addresses the second of these two differences between orthodox and heterodox economics, and uses it to suggest the direction of development of heterodox thinking generally and with respect to the individual. The last chapter discussed the role that values play in explaining the status of the embedded individual conception as an ideal conception. This chapter seeks to show how the embedded individual conception also involves positions different from the atomistic conception in regard to the fact–value distinction and, even more importantly, in regard to which substantive values are addressed in economics. Section 9.1 uses the internalist–externalist distinction and collective intentionality analysis to argue that the positivist view of facts and values is more likely to be associated with proponents of the atomistic conception, and that the opposing idea that facts and values are related in explanation is more likely to be associated with proponents of the embedded individual conception. Sections 9.2 and 9.3 then turn to the alternative normative horizons that are associated with each of the two conceptions of the individual. Section 9.2 draws on a recent exchange between Scanlon and Sen to illustrate how neoclassical welfarism and the consumer sovereignty principle are tied to the atomistic individual conception. Section 9.3 discusses the expanded normative space that is made possible when one employs an embedded individual conception, and gives particular attention to Sen’s thinking about agency freedom to emphasize “individuals matter” as a central value for heterodox economics. Section 9.4 looks at the role of power in society, and how it interacts with contemporary identity/difference “politics of recognition.” The purpose of this section is to ask why it might be normatively important that individuals maintain individual identities across change. Section 9.5 suggests that heterodox thinking about individuals and social values dates back to an earlier classical economics perspective that antedates the late nineteenth century rise of neoclassical economics.
9.1
Facts and values
Critiques of the positivist dichotomization of facts and values in economics often proceed by questioning the value-free status of facts and descriptive statements, and argue that the selection of concepts and problems, judgment of evidence, and determination of allowable methods of investigation all reflect value judgments that economists make (e.g., Blaug 1998; Wilber 1999). As reasonable a proposition as this is, it is either paid lip service, ignored, or flatly rejected by most orthodox economists. In contrast, most heterodox economists accept it. This difference in thinking, I argue, can be associated with which of the two different views of the individual one employs. I first explain this in terms of the difference between internalist and externalist definitions of individuals, and then look at what collective intentionality analysis might tell us about why orthodox and heterodox economists differ on the issue.
Before the fall: value in economics
169
9.1.1 Internalist versus externalist definitions and the fact–value distinction An internalist definition of the individual, which explains the nature of individuals apart from or in isolation from one another, can be understood in either of two ways. On the one hand, an internalist definition may countenance the existence of other individuals, but suppose that as a matter of fact the things most important about individuals have nothing to do with their relations to other individuals. Sartre’s existentialist understanding of the individual, for example, countenances the existence of others in characterizing hell as the existence of the other, but is still an internalist view in making the anxiety of individual freedom the most important thing about being an individual (Sartre 1956). On the other hand, an internalist definition can also define the individual in a more essentialist manner, maintaining that the existence of others is irrelevant to the very concept of the individual. In effect, the individual is defined in terms of properties specific to the individual alone. The atomistic conception of the individual that defines individuals psychologically in terms of subjective preferences is this latter sort of internalist definition. Understanding the individual as a subjective being is not just the most important thing about the atomistic individual. Whether or not other individuals exist is altogether irrelevant to explaining the nature of the individual as a subjective type of being. In internalist definitions of the individual of this latter sort there is a built-in immunizing argument against the standard critique of the fact–value dichotomy. That critique depends upon it being accepted that there are always rival selections of concepts and problems, rival judgments of evidence, and rival methods of investigation. However, a strong internalist account of the individual, by nature, rules out there being multiple ways involved in seeing things. Since individuals are defined as unique, independent subjectivities, there is no basis on which one might compare their different perspectives. This undermines the premise upon which the standard critique relies to create a role for value judgments, and casts doubt on the argument that all explanation needs to refer to values. The atomistic conception of the individual, of course, is not the only important concept in orthodox economics. However, it is arguably one of the most important concepts in orthodox economics, and may thus be said to set the tone for its overall character, thus predisposing orthodox economists to adopt the traditional Humean view of facts and values. In contrast, externalist definitions of the individual involve a quite different logic. Explaining individuals in terms of their relationships to others usually involves advancing particular theories of those relationships, where in social science these particular theories have historically been very different from one another. Consider, for example, how different a Freudian externalist theory of the individual formulated in terms of parental sexual relationships is from a Marxist externalist theory of the individual formulated in terms of class relationships. Thus, those who think about individuals within the externalist framework tend to appreciate that their definitions
170
Before the fall: value in economics
and theories rival and are rivaled by those of others. This in turn usually compels a realism about the role of values in social science explanations that leads many to reject the Humean separation of facts and values. Thus, whereas an internalist definition of the individual tends to predispose orthodox economists to a positivist view of facts and values, an externalist approach to the individual rather tends to predispose heterodox economists to accepting the idea that values have a place in economic explanations. 9.1.2
Collective intentionality and the fact–value distinction
However, an even stronger argument for a postpositivist understanding of facts and values is available to heterodox economists in connection with the specific externalist account of socially embedded individuals advanced above. That account explains individuals as socially embedded through a collective intentionality analysis of their we-intentions. When individuals form we-intentions about the groups of which they are members, they take into account others’ thinking together with their own, since individual we-intentions are the product of sets of interdependent beliefs about what members of groups think. Consequently, any use of language involving collective intentions requires that an individual consider both what can and what should be said according to the state of shared and interdependent views of members of the group. However, this means that, even when an individual uses an ostensibly value-neutral, descriptive language to explain group intentions, that language inevitably reflects other individuals’ values regarding what language is thought appropriate in the circumstances (as well, indeed, as the values of the individual, who is no less a member of the group contributing to shared we-intentions). In a we-intention group context, then, facts and values everywhere interpenetrate and relate to one another, because individuals are interdependent. This is not to say that facts and values are indistinguishable, since we can still define and characterize them in distinct terms. However, descriptive fact-stating language, which positivists have claimed functions independently of values, at least in we-intentional, group contexts, is clearly influenced by values.1 We can extend this reasoning to the explanations that economists themselves produce. Economists, like other groups, have views about what they believe economists believe as a group. Moreover, when they advance their theories and arguments about the economy, they recognize that they are influenced by other economists, and suppose that they are contributing to the thinking of other economists. Accordingly, we-intentions also underlie economists’ explanations and descriptions of economic life, and consequently these explanations and descriptions are value laden in the same way that the explanations and descriptions of other kinds of groups are value laden. Others have drawn similar conclusions about values in scientific communities in sociology of scientific knowledge terms with respect to the nature of economics as a social practice operating in a social context (e.g., Mäki 1993; Hands 2001: 186–212).
Before the fall: value in economics
171
Individual economists, just as other scientists, are influenced and guided by the values embedded in scientific communities in their research. Here I merely add that this same conclusion applies when social practices in which individual economists engage are understood from the perspective of an externalist definition of individuals as socially embedded in a collective intentionality sense.
9.2 Welfarism, utilitarianism, and the atomistic individual conception In the following two subsections, I discuss the extent to which, just as our understanding of the relationship between facts and values depends upon our particular understanding of the individual, so our understanding of which particular normative values are important in economics also depends upon our particular understanding of the individual. To do so, I review a recent exchange between Scanlon and Sen to illustrate how neoclassical welfarism with its roots in classical utilitarianism relies on an internalist account of the individual. I then turn to the broader normative framework—characterized by Sen as “deontic-value inclusive consequential reasoning”—that becomes possible when one employs an externalist, embedded individual conception. 9.2.1
Scanlon versus Sen
The normative posture of neoclassical economics originates in classical utilitarianism, which in its current form is understood in terms of Pareto optimality judgments. Sen sees classical utilitarianism as having three theses: welfarism, sumranking, and consequentialism (Sen 1987: 39–40). Welfarism is the idea that a state of affairs must be judged by the individual utilities in that state, where there are three different ways of defining individual utility, namely, pleasure, desire fulfillment, and choice (Sen 1982: 7–13). All three are subjectivist measures of individual well-being. Sum-ranking is the idea that individual utilities are additive and count equally. In cardinal utility accounts, the issue was how to make interpersonal utility comparisons to permit net summation over individual utilities. With ordinalism and Pareto optimality judgments, the emphasis on the individual is sustained in the notion that a state of affairs is preferable if at least one individual is made better off (in that individual’s estimation).2 Consequentialism is the idea that states of affairs are judged in terms of their consequences rather than in terms of such matters as rights, duties, and obligations. In effect, what makes an act right is that it is productive of good, rather than that it comports with any set of deontological principles, whatever their basis might be. Scanlon, however, argues that welfarism—the idea that a state of affairs must be judged by the individual utilities in that state—is the key to the whole of utilitarianism: “Although these [three] theses are logically independent, it seems to me that their appeal is not independent, and that in fact it is welfarism (or, historically, hedonism) that provides the support for the other two” (Scanlon 2001: 40). Sum-ranking
172
Before the fall: value in economics
makes sense in terms of welfarism, because valuing utilities naturally suggests that more utility is better than less. This applies no less to Pareto optimality judgments. Consequentialism also makes sense in terms of welfarism, because if the subjective states that characterize utility are the “sole ultimate values … what morality requires is that we produce those states of affairs with the greatest possible value” (Scanlon 2001: 40). However, utilitarianism’s general dependence on welfarism, Scanlon argues, can perhaps most clearly be seen in its bias towards regarding utility states— and not the actions that produce them—as the sole objects of value. As he sees it, there is no reason not to consider the actions that produce utility states as objects of value, other than that doing so would undermine welfarism. Sen is also bothered by welfarism’s exclusive emphasis on utility states to the exclusion of attention to the actions that produce them. Indeed, in his view, only to consider utility states, or the utility outcomes that actions produce, as valuable should be characterized as “an arbitrary exclusion” (Sen 1982: 29). “Had welfarism not been additionally imposed,” he argues, “consequentialism could have coexisted with taking note of such things as the values and disvalues of actions through the valuation of the states, which include these actions” (Sen 1985: 182). If we suppose that actions have value because they possess pleasure or satisfaction in themselves, then actions will sometimes add to the value of a utility state and at other times lower the value of a utility state. The situation becomes more complicated if actions have value for reasons other than that they give pleasure or satisfaction. However, in either case, Sen argues, the value of actions ought to be taken into account in consequentialist reasoning, since actions can have value apart from the consequences they produce. Welfarism, however, in its internalist preoccupation with individual utility states alone, rules this out. Thus, Sen recommends a consequentialist framework that is not welfarist—and thus not strictly internalist individualist—though perhaps still utilitarian in a broader sense.3 Scanlon, nonetheless, takes a critical view of even a broadened consequentialism, and suggests that considering the values of actions themselves ultimately points us towards a larger normative framework: It is one thing to include, as a determinant of the value of a state of affairs, the fact that it includes an action that an agent would much prefer not to have performed; but it is something else again to take into account the fact that this action was wrong, or violated someone’s right. (Scanlon 2001: 41)
Basically, consequentialism comes into question once one begins to consider the intrinsic moral value of actions, since then consequences alone no longer determine which actions we ought to pursue. At bottom, Scanlon believes, the problem is the logic of a “Foundational Consequentialism,” as he calls it, that “might be taken to start from the belief that the ideal of ‘the good’ has clearer and more obvious justifi-
Before the fall: value in economics
173
catory force than ‘the right’” (Scanlon 2001: 39). Welfarism, with its individualist foundation, is the stopgap that prevents consideration of this type of consequentialism, since it forecloses consideration of the value of actions. 9.2.2
A “deontic-value inclusive consequential reasoning”
Thus, it seems not so much that welfarism has been maintained by neoclassical economists in order to shore up a commitment to consequentialism as that a commitment to value residing solely in individual states of mind—and an internalist view of the individual—has had the effect of shoring up consequentialism as a particular form of utilitarianism. The scope of Scanlon’s argument about utilitarianism being welfarist at root can thus be extended to say that utilitarianism is also subjective individualist at root. Perhaps the most tangible manifestation of this lies in the related popular doctrine of consumer sovereignty, which directly links subjectivity (in the idea that individuals alone know their own desires and preferences) to an exclusive emphasis on producing the greatest possible sum-ranking of individual utilities. Scanlon might thus say that, if we are to question welfarism, we must question consumer sovereignty as well. By extension, if we are to question subjectivist views of the individual, we must question consumer sovereignty as well. Does this imply that consequentialism also must go? Scanlon thinks this follows, but Sen, in rejecting subjectivism, still retains a consequentialist framework, though one enlarged to accommodate not only the value of actions but also the very sorts of values (freedoms, rights, duties) that Scanlon favors. Indeed, he rejects Scanlon’s claim that employing consequentialist reasoning implies that “the ideal of ‘the good’ has clearer and more obvious justificatory force than ‘the right’” (Scanlon 2001: 39). In the sort of enlarged consequentialist framework that Sen prefers: It is neither that “the good” comes first, and then “rights and duties,” nor that rights and duties congeal first followed by the good, but that they are linked concepts that demand simultaneous consideration. While considerations of freedoms, rights and duties are not the only ones that matter (for example, well-being does too), they are nevertheless part of the contentions that we have reason to take into account in deciding on what would be best or acceptable to do. The issue surely is simultaneity. (Sen 2001: 61)
Sen terms such a framework “deontic-value inclusive consequential reasoning” (Sen 2001: 64). It certainly includes in Sen’s mind reasoning about functionings and capabilities. It lacks, he does not hesitate to say, the neat algorithmic decisionmaking procedure that is often attributed to either pure utilitarian or pure deontological reasoning in ethics. Yet, in Sen’s view, moral reasoning is complex and not subject to simple rules, as desirable as having such rules might seem to be. This also tells us that moral reasoning relies on social frameworks that facilitate exchange
174
Before the fall: value in economics
of moral views and a process of exchange between individuals whose outcome are individual and social moral decisions. In the section that follows, then, I use these ideas to argue that the heterodox conception of the individual as socially embedded— understood in a collective intentionality sense—requires just such an expanded normative account along the lines suggested in the passage above. As with the atomistic conception and mainstream economics, then, normative thinking appropriate to heterodox economics can also be seen to depend on the associated theory of the individual.
9.3 Normative reasoning and the socially embedded individual conception We have seen that the normative value emphasized in much of Sen’s work on functionings and capabilities is his concept of agency freedom which is linked to his real opportunities sense of the concept of capability (e.g., Sen 1993). The argument of the last chapter was that individuals exercise this particular form of freedom when they accept combinations of rights and responsibilities associated with membership in social groups. However, if this type of freedom is to be thought a goal for individuals who are socially embedded in groups, then we must ask ourselves whether social groups can indeed promote this sort of freedom. This raises the larger issue of how normative values arise and operate not just in social groups but in organizations and institutions generally, or, as it has recently been expressed, whether we may treat “values as partly endogenous to the economic system, and economic systems and their performance as partly functions of people’s values” (Ben-Ner and Putterman 1998: xvii). To see how values operate in social contexts, then, I begin by contrasting the orthodox view of how they operate in organizations made up of atomistic individuals. Essentially following Hume, the orthodox account attempts to explain how conventions that lack a normative character in themselves can nonetheless come to acquire the status of moral norms. 9.3.1
“Moral” sentiments in organizations and institutions
Hume took a system of justice to be a set of conventions that arise when individuals come to expect one another to behave in regular and predictable ways (Hume 1888 [1739]). Individuals abide by a system of justice because they find it in their interest to conform to its rules when they expect others to conform to them as well. The idea that such rules are “conventional” comes from supposing that there are different possible rules of justice, and those that actually come about reflect a history of contingent interactions between people. Nothing a priori moral underlies actual systems of justice, making them for Hume not partly but entirely endogenous to the economic system. But why, then, should such rules be thought normative in nature? Why should they be thought to be anything more than simply persistent regularities
Before the fall: value in economics
175
in social behavior akin to other regularities that no one believes have normative content? Hume’s view, based on eighteenth century Scottish School psychology of sympathy, was that conformity with such regularities evokes sentiments of approval, and failing to conform with them evokes sentiments of disapproval. When these sentiments become widely shared and become attached to an idea of the general good, he believed they may then be characterized as a moral approval and disapproval. A system of justice, then, is ultimately nothing more than a relatively settled set of conventional expectations between individuals chiefly concerned with their own interest that is reinforced by sentiments of approval and disapproval. This conception has been modified and redeveloped in recent years by Lewis and Sugden. For Lewis, conventions are coordination equilibria (Lewis 1969). Coordination equilibria can be explained in game-theoretic terms, with players acting on individual strategies to achieve a common expectation regarding which individual strategies offer the best reply to one another (that is, they are Nash equilibria). Hume’s psychology of sympathy is replaced by a characterization of individuals in terms of strategies, but any norms that emerge are still conventional and entirely endogenous to the economic system. Sugden similarly explains conventions in terms of individuals’ expectations of one another conforming to regularities in behavior, but adds a concern individuals are said to have over incurring others’ resentment as an emotion underlying conformity to conventions (Sugden 1986, 1989). When this emotion operates widely to reinforce individuals’ adherence to conventions, Sugden then believes that normative expectations obtain among them (Sugden 1998). However, against this it seems it might be said that the emotion of resentment no more deserves the label “normative” than Hume’s talk of approval and disapproval deserves the label “moral.” Sugden argues in reply that this criticism misses the point behind providing a Humean naturalistic analysis of values: “In such an analysis, the definition of a moral sentiment has to be naturalistic; one cannot then object that some of the sentiments allowed by the definition are not ‘really’ moral” (Sudgen 1998: 84). In my view, such a response is questionable. The claim that “moral sentiments” are just that, namely, somehow “moral,” needs a stronger defense than the suggestion that it should be possible to explain moral values naturalistically, and that therefore there must exist such things as “moral” sentiments. Indeed, making this sort of argument seems to involve making exactly what G. E. Moore famously labeled the “naturalistic fallacy” (Moore 1903). Note also that the program of producing a naturalistic account of normative values is closely associated with the aim of producing a positivist interpretation of moral life. Sugden is explicit about this connection, asserting that economists “trained in a positivist tradition” must seek to explain normative values without “assuming the existence of moral facts” (Sugden 1998: 76). A moral fact is a fact about something being right or wrong, good or bad, etc. For example, one might say it is a fact—specifically a moral fact—that it is wrong to needlessly harm another person. To deny that moral facts exist is to say there is nothing in society or nature that can be described in normative terms as a matter of
176
Before the fall: value in economics
fact. Normative values, rather, are subjective judgments, and must accordingly be explained in terms of some sort of “moral” sensibility that people exercise and impose on the world. Two obvious problems arise with this sort of approach. First, the idea that human society can or should be described naturalistically is highly questionable and has not surprisingly been long contested in the history of social science. Sugden and others in this tradition generally do not explain why a natural science approach to social science might be plausible, and I am skeptical that any good arguments exist on this score. Second, this approach creates a very strong problem for making the transition from “is” to “ought” (thus, Moore’s naturalistic fallacy). Sugden addresses this problem by simply insisting on a redefinition of “moral” in naturalistic terms. Whether this is a reasonable redefinition of “moral,” however, depends on whether the account of moral behavior that emerges on these terms captures what we ordinarily think is bound up with the normative. What is it, then, that is most characteristic of the normative domain? We can begin to answer this question by emphasizing the implied content of the moral “ought.” When “ought” appears in an expression in a normative way, it indicates the presence of a moral obligation. If individuals do something because they believe they morally ought to, they do so out of a sense of moral obligation operating upon them. There are many ways of understanding what this sense of “ought” involves, but, following Kant (Kant 1948 [1785]), I take the minimum essential idea to be a matter of doing something because it is required irrespective of one’s inclinations or desires. But then the idea of acting out of a sense of obligation is not what is involved in acting on a “moral” sentiment. If one is motivated to respect a norm or convention, because one fears others’ resentment or disapproval for failing to do so, it is not a sense of obligation but an inclination that operates as one’s motivation. Rather, fear of others’ resentment or disapproval has replaced acting out of a sense of obligation. The closest Sugden comes to referring to a sense of obligation, then, is when he argues that one of the virtues of his analysis is that it “allows us to consider cases in which normative expectations and self-interest pull in opposite directions … cases in which individuals follow conventions even though this is contrary to self-interest” (Sugden 1998: 83). However, this is not evidence of acting out of a sense of obligation, since one might well be inclined to observe a convention that was contrary to self-interest. Thus, if we take the idea of acting out of a sense of obligation to underlie what is involved in moral “ought” thinking, a naturalistic, moral sentiments type of approach does not capture what we ordinarily think is bound up with the normative. This gives us good reason to conclude that the Lewis–Sugden type of development of Hume regarding how values operate in organizations and institutions as conventions is not an account of distinctively normative expectations. This in turn tells us that a naturalistic approach probably cannot explain how normative values arise and operate in organizations and institutions. It is worth emphasizing, then, that what the Humean tradition offers on this score derives from its starting point in the notion that individ-
Before the fall: value in economics
177
uals are naturalistically described as isolated from one another, and generally acting in their own interest. Let us, then, begin with a different starting point by describing individuals as embedded in social groups, and then ask how normative values arise and operate in organizations and institutions. The following questions arise: 1 2 3
Does this strategy enable us to talk about what is most characteristic of the normative domain, namely, a sense of obligation that individuals have when they use “ought” language? Does it provide us with an account of the variety and types of relationships between normative values in social life along the lines of Sen’s “deontic-value inclusive consequentialist” framework? Does it suggest a role for Sen’s emphasis on agency freedom in this framework?
9.3.2
Question 1: moral obligation in organizations and institutions
The first question can be answered by explaining what is involved in saying that socially embedded individuals, understood in a collective intentionality sense, have we-intentions as well as I-intentions. The Humean framework, by taking individuals as fundamentally isolated from one another, operates exclusively with I-intentions. As previously argued, the use of “we” language generally creates obligations for individuals—and not just in the moral sense. The collective intentionality framework, particularly as developed by Tuomela, shares with the Lewis–Sugden account of convention the idea that individual expectations are established within a system of reciprocal expectations between individuals. However, Tuomela’s account is different in that individual expectations have as their object reciprocal sets of weintentions rather the I-intentions implicitly involved in the Humean framework. We-intentions, as previously argued, create obligations for individuals, because the successful use of “we” by an individual needs to conform to how others use that same “we.” Outside a requirement to use language correctly, this obligation does not exist for the use of “I.” Of course, some obligations that individuals recognize are pragmatic, and consequently do not have moral content. However, with the interpretation here, collective intentionality analysis is not naturalistic and certainly not motivated by positivistic aims. Thus, it is as reasonable to suppose that moral facts exist as to suppose that they do not. From this it follows that some of the obligations that individuals observe are indeed moral in nature. Though the dividing line between pragmatic and moral obligations may often be difficult to draw, and though it may change over time, it seems there are many clear cases of each, and thus fair to say that individuals who form we-intentions and use “we” language often operate under a sense of moral obligation. Therefore, a collective intentionality framework, by operating with a conception of socially embedded individuals rather than atomistic ones, makes it possible to include a sense of moral obligation alongside individual inclination as a form of
178
Before the fall: value in economics
individual motivation. Turning to the second question above, then, what does the collective intentionality framework and the conception of individuals as socially embedded tell us about the range and variety of normative values in social life and the relationships between them? 9.3.3
Question 2: an expanded normative domain
The emphasis on moral obligation thus far has rested on looking at moral obligation as something that particular individuals recognize. However, a fuller characterization of the concept needs to regard these obligations not just from the point of view of the individuals who have them but also from the point of view of the individuals to whom they may apply. This is tantamount to saying that the concept of moral obligation is an externalist-individual type of normative concept, since it emphasizes relationships between individuals. It contrasts with saying that the concept of well-being is an internalist-individual type of normative concept, since (at least in orthodox economics) well-being is essentially a property of the individual. Indeed, the Pareto efficiency standard emphasizes this basically by judging states of affairs according to whether at least one person is better off (though the older cardinal utility framework which made interpersonal utility comparisons employed an externalist-individual type of normative concept). Externalist-individual normative concepts, it can be argued, just because they emphasize relationships between individuals, generally require that we give attention to an entire range of normative concepts that go beyond whatever particular normative concepts (say, regarding what is good) constitute a particular individual’s moral view. Thus, to give any kind of detailed explanation of the moral obligations that one has to others, one typically also needs to have an understanding of one’s own and others’ rights. However, systems of rights are themselves embedded in broad social commitments to such ideals as freedom, equality, fairness, human dignity, community, and justice. Thus, employing an externalist-individual type of normative concept typically commits one to examining an entire range of accompanying normative concerns. This means that the connections between these normative concerns cannot generally be mapped out with any high degree of precision. However, following Sen I believe this to be an inescapable aspect of moral life, and one, moreover, that has been mostly overlooked by economists accustomed to the sort of highly determinate normative judgment made possible by internalist-individual normative standards such as the Pareto concept. The idea of an expanded normative domain, then, is linked to an understanding of individuals’ social embeddedness in terms of their involvements in social groups. Social groups generally have goals that help define them. Thus, their members’ weintentions often concern a consequentialist type of moral reasoning, as when something is regarded as right on account of its helping to bring about some outcome desired by the group. However, this sort of consequentialist moral reasoning, when it is expressed in we-intention terms, also has independent concepts of moral obliga-
Before the fall: value in economics
179
tion associated with it. Thus, a particular individual using “we” language in regard to what potential good consequences a group wants to bring about operates both with an idea that what is right is a matter of bringing about the relevant outcome and with an idea that what is right is a matter of observing obligations upon oneself in using we-intentions. This latter sense of right draws in turn upon other ideals such as justice, equality, and dignity. This is therefore one way of talking about a “deonticvalue inclusive consequential reasoning” (Sen 2001: 64). As Sen emphasizes, in such a framework, “[i]t is neither that ‘the good’ comes first, and then ‘rights and duties’, nor that rights and duties congeal first followed by the good, but that they are linked concepts that demand simultaneous consideration” (Sen 2001: 61). 9.3.4
Question 3: Sen’s concept of agency freedom
Sen does not derive the view above from a particular theory of the individual, but rather reaches it on the basis of the interrelatedness of our normative concerns. In closing this section, however, we might ask what role his concept of agency freedom—which figures so prominently in his account of individual capabilities—has in this framework and also in relation to our normative thinking about individuals regarded as being socially embedded in groups. The emphasis on the concept of agency freedom in Sen’s real opportunities understanding of individual capabilities rests on individuals being able to investigate counterfactual opportunities: “Acting freely and being able to choose may be directly conducive to well-being, not just because more freedom may make better alternatives available” (Sen 1993: 39). That is, freedom is something that is desirable by virtue of the value of simply being able to explore the possibilities that life offers—whatever the consequences of doing so. For Sen, accordingly, we develop our individual capabilities as real opportunities by being able to act freely in this strong sense. If individuals, then, are enmeshed in countless cross-cutting social group relationships, one way we may understand individual freedom is in terms of their being able to move back and forth comfortably across these relationships. This would imply that individuals are not fragmented and lost amid these relationships, but, as argued in the last chapter, sustain an individual identity across them. Sen’s idea that individuals develop their capabilities by exercising a freedom to explore possibilities in life provides one way of thinking about individual identity in such circumstances. Individuals may be said to acquire an identity across and independent of their various group involvements when they are able to freely investigate and embrace the various obligations associated with their membership in social groups. Their being able to do this—a matter of their developing a capability in this regard—would identify them as independent individuals. Agency freedom, which underlies Sen’s thinking about capability, can accordingly be thought central to understanding the personal identity of socially embedded individuals. What this argument adds, then, to the discussion above regarding an expanded normative domain is a particular focus within that domain that gives special weight to
180
Before the fall: value in economics
one particular social value, namely, that “individuals matter.” My suggestion here is that heterodox economists, who have sought to replace the atomistic conception of the individual with something richer, have, at least implicitly, made the value “individuals matter” central to their understanding of the expanded normative space they see as operating in economic life. Sen’s particular attention to agency freedom facilitates this, particularly when we combine this attention with an understanding of what it means for individuals to be socially embedded in groups. In the section that follows, then, I comment on why the concept of individual identity should be an element in the view that “individuals matter”.
9.4
Individual identity as a normative concern
The idea that individuals have an individual identity has been framed in terms of whether they are reidentifiable across change in their group involvements. However, should we even have a normative concern for whether individuals maintain an identity across change? Why, specifically, is individual identity important? My claim that it is important is based on the fact that many power relationships in society work to individuals’ disadvantage by confining them to particular roles and narrow identities associated with those roles. When individuals with power over others use that relationship for their own self-aggrandizement, they often seek to exercise their control by limiting the scope of others’ activities. Limiting the scope of other individuals’ activities is tantamount to imposing particular roles with well-defined boundaries upon them. In the last chapter I explained this in terms of Granovetter’s idea (Granovetter 1985) that individuals who are unable to sustain individual identities across change in their various group involvements become “oversocialized” in that others are able to determine their rights and responsibilities. Here, I add to this characterization an emphasis on the other side of this relationship, namely, that there always exist individuals in the world who—not only out of self-interest but sometimes also to promote their own moral visions—are actively engaged in imposing the status of being “oversocialized” on others. Broadly speaking, their objective is to create identities for those individuals whose activities they wish to control and direct. Recognizing their existence and making them part of one’s thinking about normative theory constitutes a further consequence of defining individuals in externalist terms. Consider patriarchal relationships between women and men. Historically, women have been assigned one type of identity, namely, an identity exclusively associated with the domain of the household and a responsibility for caring activities, though they have long sought identities beyond this domain, most recently particularly in connection with a whole range of market employments unrelated to their traditional household identities. Men who promote this exclusive assignment of household identities to women exercise power over women not only in reserving a set of preferred activities for themselves but also in providing themselves identities across their different social group involvements. Indeed, men enjoy the status of fathers,
Before the fall: value in economics
181
husbands, and heads of households in large part because they are not confined solely to household identities. The rationale behind antidiscrimination law in the United States which bars employers from inquiring about women’s marriage and family status is often thought specifically to be a matter of protecting women’s rights as individuals in the labor market. I suggest, however, that what is at issue is not just fair treatment in the labor market but a larger right of individuals to develop and pursue their life goals as they choose, where this involves the opportunity to develop an identity across their multiple social group involvements. Generally speaking, then, since US antidiscrimination law bars employers from using gender, race, age, disability, and religion in the determination of employment decisions, it represents a broad public policy aimed at promoting human development premised on the idea that individuals seek to develop their individual identities across their various social group involvements. However, similar arguments can be made in connection with virtually all of the ways in which individuals are characterized in group terms—race, class, sexual preference, age, religion, ethnic or national origin, etc. Essentially, when individuals acquire some social or group identification, this creates opportunities for others to reduce access or opportunities for those individuals outside their social identification. The irony is that, while social identification, and an identity/difference “politics of recognition” (Taylor 1992; Benhabib 1996; Fraser 1997), is seen by many individuals as a means of escaping social domination, group identities are also often “officially recognized, sanctified, legitimized, and accepted by the state and its institutions” to function as “corporate identities” (Benhabib 1999: 298). These institutionalized identities entail a systematic “misrecognition” that both reifies dominant culture conceptions and suppresses internal group heterogeneity (Fraser 2000). One way of seeing this latter problem is as a matter of an imbalance between individuals’ social identities and their individual identities. Individuals gain from social identification up to the point at which this recognition undermines their other characteristics not reflected in that group identity. Beyond this point social identities become burdensome, and the individual’s identity apart from the group is sacrificed. From this perspective, what is normatively significant about the concept of individual or personal identity is not just whether individuals maintain identities across change but also whether individuals are able to freely and flexibly self-impose the responsibilities of group membership upon themselves across groups and in moving from one group identity to the next. This helps reinforce the importance of Sen’s real opportunities sense of capability that emphasizes the freedom to investigate counterfactual possibilities. One reason why an emphasis on being able to investigate counterfactual possibilities is important is that social groups vary sufficiently from one to the next that individuals continually find their experience in and across groups to be complex. In continually encountering new situations in new groups (and in the changing nature of existing groups), individuals are forever in need of considering how things might be done in order to figure out how they ought to be done. That is,
182
Before the fall: value in economics
individuals need to be able to reason in terms of counterfactual possibilities because of the inherent (and arguably increasing) complexity of life today which includes multiple and changing group involvements. Thus, another way of saying that individual identity is normatively significant is to say that it is normatively significant whether individuals successfully develop a capability in this real opportunities sense for moving through society over their lifetimes. The focus on individual identity, then, acts as an organizing framework for a number of normative concerns. Sen emphasizes freedom in his account of human capability development. However, he also generally favors a normative framework sufficiently comprehensive as to permit a “deontic-value inclusive consequential reasoning” in which good consequences count just as do rights and duties. Basically, he argues for this larger normative framework on intuitive grounds, in that we do not seem to be able to address the variety of complex moral issues that the world involves with narrower, one-principle type normative frameworks. The suggestion here, however, is that we can go beyond this intuition regarding the need for broad scope in our normative framework by filling in the social value that “individuals matter” with an understanding of how they can be made to matter in individual identity development terms. Human and individual development is very much a matter of individuals being able freely to negotiate their different social group involvements; being able freely to negotiate one’s group involvements is a matter of having an individual identity across them. Thus, underlying a comprehensive normative framework that premises human development is a concern with individual identity.
9.5
A classical world
I argued in Chapter 2 that Smith partially resolved Locke’s dilemma regarding how individuals’ inner subjective worlds linked up with the outer, Newtonian objective world. Smith used the Scottish Enlightenment concept of unintended consequences and the idea that the market worked as if by an invisible hand to show how individuals’ private subjectivities might explain the market as a Newtonian-type process understood in terms of equilibria and causal connections. However, Smith also believed that the market operated within a moral system that he had described in The Theory of Moral Sentiments written before The Wealth of Nations (Evensky 1993). Thus, while he explained the market in terms of self-interest, he also saw self-interest as companion to moral interest, and accordingly saw the mechanical working of the market occurring within a social framework made possible in part by moral principles. Smith’s “other” book, of course, has been largely forgotten, leaving the market and self-interest to be explained entirely naturalistically. This helped produce a vision of economics as a positive science long before Robbins narrowed the subject’s policy space. Robbins merely applied an established principle that normative value did not enter into explanation in economics.
Before the fall: value in economics
183
Smith’s classical world, then, was one in which good explanation depended upon accounting for moral values at work in economic life. However, unlike those who later amended this world as neoclassical, Smith was comfortable with what this implied about the relationship between facts and values in economic explanation: that explaining a world in which moral values are at work inevitably makes explanation value laden. Basically, interpretation of a value-laden world relies on one’s understanding of values. Thus, one can only avoid value-laden explanations by supposing that the world one is examining is value free. Smith did not believe this to be true of the world he lived in, but his confidence that his values reflected his society’s values made him confident that his explanations were good ones. The positivist response is to point to the dangers of circularity in this, and then call for explanations that avoid this sort of value nexus. The problem with this response is that it presupposes that the economic world can be explained without referring to individual and social values. If this assumption is false, then the strictures against circularity are irrelevant. Smith thought the assumption false. The neoclassical world removes normative value to the policymaking periphery beyond economics by reconstituting the subject matter of economics. The key to doing this was treatment of the individual in fully nonnormative terms—something that could only be accomplished by conceptualizing the individual as an atomistic being isolated from other individuals. This strategy, however, ultimately breaks down as a theory of the individual. The effect is to leave us without the resources needed for understanding individuals in today’s complex world which ranges from circumstances in which individuals are relatively isolated to circumstances in which their interaction with others is changing, demanding, and multisided. If the social world is complex, our theory of individual behavior also needs to be complex. However, if individual behavior is additionally complex in including normative content, our view of economic explanation also needs to be complex. Neither atomism nor positivism succeed on these counts. Thus, despite their long history of association with liberalism and defense of the individual, they offer an inadequate defense of the individual in the world today. However, elements of a defense of the individual may be available in a return to the classical world of the socially embedded individual.
10 Revisiting the issues
My experiences are of people and houses and pinchings and aspirins, all things which I understand, in large part, in terms of their nature. (Cartwright 1989: 60) And then she thought that you went on living one day after another, and in time you were somebody else, your previous self only like a close relative, a sister or brother, with whom you shared a past. But a different person, a separate life. (Frazier 1997: 422)
The aim of these concluding remarks is to comment briefly upon the framework in which the book’s arguments are developed. I begin by revisiting the main methodological issues introduced in Chapter 1 in an effort to clarify the motivation for the book’s focus on ontology and identity analysis. My goal in this regard is also to make a case for the approach adopted here in the hope that it will add credibility to the book’s arguments and conclusions. I also provide an overview of the historical methodology associated with the accounts given in Chapters 2 and 6 of the two main traditions of thinking about the individual, both in order to put these accounts in better perspective, and perhaps to be clearer about what those accounts were meant to achieve. Finally, I close with a summary discussion of what I believe to be the central conclusions of this book. Section 10.1 discusses the nature of ontological investigation which the book employs, makes a case for ontological thinking in economics as an important branch of economic methodology, and then explains identity analysis as a form of ontological investigation. Section 10.2 discusses an issue suggested by the book’s exclusive focus on the identity of individuals in economics, namely, how we might look upon the relationship between individual identity in economics and personal identity. In Section 10.3, the historical methodology implicit in the accounts in Chapters 2 and 6 is discussed in terms of the metaphorical nature of the atomistic and embedded individual conceptions. Section 10.4 briefly summarizes five central conclusions of the book.
Revisiting the issues
10.1
185
Ontology in economics
What distinguishes the arguments in this book from many others on the topic of the individual in economics and in economic methodology is their explicitly ontological focus. This may give the arguments here an appearance of unfamiliarity, since ontological issues in economics are less often discussed than epistemological ones. Moreover, when ontology is the focus, the discussion is more often carried out at what might be termed a metaontological level in regard to how such positions as realism, idealism, instrumentalism, operationalism, constructivism, etc., underlie theories, assumptions, and models in economics. In this book, however, following Cartwright’s intuition regarding the importance of capacities as opposed to laws in science (Cartwright 1989), the aim has rather been to carry out an ontological analysis in terms of those “things” in the world that economics is typically about, and accordingly the discussion has been pitched at the object level in asking how a particular type of “thing” taken to be central to much of economics—the individual— might be investigated in an ontological manner. Though this type of investigation has implications for metaontological views, these have been put aside in an effort to focus on the points of contact that particular theories in economics might or might not have with the world through their different conceptions of the “things” in economic life. Put differently, I see ontological investigation in economics as a way of explaining how the object terms in theories refer to things in the world, and thus take the subject of reference to be central to the evaluation of economic theories. Most treatments of reference, however, operate from a philosophy of language perspective in emphasizing how language works. Thus, the investigation in this book is also distinct in examining reference ontologically, in that it examines the circumstances under which those objects that economics describes can indeed be said to refer to or pick out real-world existents. This involves our first saying what kinds of thing we are interested in, since, as Shwayder tells us, one cannot “say anything about an object without allowing for the possibility of the question, ‘What kind of thing is it?’” (Shwayder 1965: 41), or, as Quine records the standard wisdom on the subject, there can be “no entity without identity” (Quine 1969: 23). Once, then, we have a view of the kinds of thing we are interested in—here, the individual in economic life—we are then in a position to outline the requirements for that kind of thing to exist—here, that individuals are distinct and reidentifiable. This is what I have treated as identity analysis and as a framework in which successful or unsuccessful reference can be explored. In sum, if we want to know whether an object term in a theory picks out a real thing in the world, we characterize that object as being of a certain kind, ask what existence tests that sort of thing as described must satisfy for it to be said to exist, and finally check whether the conception in question satisfies the appropriate tests. The advantage in proceeding in this way, in my view, is that it gives ontological analysis a possibly significant role in economic methodology. Metaontological
186
Revisiting the issues
analysis, which makes claims at the level of theories, assumptions, and models, does not easily discriminate between different approaches in economics in terms of whether they succeed in being about the world. The problem is that theories, assumptions, and models are almost always thought to have some relation to the world. However, that they always have “some” relation threatens to overwhelm systematic consideration of the issue altogether, so that, if we suppose that “good” economics is determined in part in terms of how well a theory relates to the world, then most economics might end up qualifying as “good” economics. It is true that economic methodologists are ambivalent about whether methodology should be normative or descriptive, and so perhaps the question of what constitutes “good” economics will strike some as irrelevant (cf. Weintraub 1990). However, it might also be the case that methodologists’ ambivalence about the character of economic methodology is partly the result of disenchantment over failure to agree about standards, and that this in part is a product of a long history of failure since Kuhn to get beyond epistemological arguments to ontological discussion of the ways in which economics may or may not succeed in being about the world (Davis 2003). Some will no doubt say, that identity analysis is but one more voice—neither more clear nor more commanding than others—added to the already large number of competing voices in economic methodology. Moreover, identity analysis may not seem a very strong candidate for producing clarity in thinking about economics and the world, because ontology in general and identity analysis in particular have received relatively limited attention from economic methodologists. In response, again I emphasize the central intuition behind the arguments in the preceding chapters, namely, my belief that theories, assumptions, and models in economics and the social science are in considerable degree formulated in terms of and constructed around thinking about their objects. The reason for this, I suggest, is that, as a social science that examines cause-and-effect relationships, economics is primarily concerned with behavior. Behavior, moreover, is associated with entities, or more properly agents, which are consequently the main objects of investigation. This focus, it seems, is what creates a place for identity analysis in economics. Let me, however, add another reason for this approach rooted in the particular arguments of this book. If economics is indeed chiefly constructed around thinking about its objects, and if “good” economic theory simply does a good job in explaining the nature of the things in the world, then economics has to be able to differentiate between those occasions when it is about one kind of thing from those occasions when it is about the other kind of thing. In terms of the problem of individuation, “good” economics needs to be able to persuasively say when it is about single-person and when it is about multiperson agents. However, here lies an important dilemma for political economy and economics since Smith. In over two hundred years of theorizing about economic life, political economists and economists have never been able to agree on when agency resides in single individuals and when it resides in groups of individuals. Rather, one or the other type of agent is typically
Revisiting the issues
187
assumed from the outset, and the economist then proceeds as if this agent is the world’s exclusive existent. Thus, it is no accident that political economy and economics exhibit a long history of unresolved opposition between methodological individualism and methodological holism. To make matters worse, these two traditions have had very little serious communication with one another, no doubt leading disinterested observers of economics to wonder how coherent a discipline can be that is so schizophrenic at the most elementary level of how it relates to the world. Basically, then, economics is not very effective in saying what it is about because it does not address its own most basic ontological concerns. Moreover, since it cannot really say in a reasonably rigorous way when it is about single-person agents and when it is about agents made up of many persons, its arguments are often too blunt to provide reliable cause-and-effect analysis. Perhaps even more disturbing is that the individual, which economics has long claimed to be successful in explaining, turns out to be something of a black hole in the subject—a place into which vast amounts of intellectual energy has historically disappeared without much in the way of return. Thus, it seems there is a good case for saying that there ought to be more attention paid to ontological analysis in economics. Or in simpler terms, what is required of economics is that its objects be more carefully defined, where this ultimately involves saying how one picks out something in the world from among many other kinds of things that exist when using one’s definition of it. My method of definition, then, involves using identity conditions and existence tests. My general framework combines ontology, reference, and identity.
10.2 Individual identity in economics and personal identity I introduced the topic of individual identity in economics in Chapter 1 by comparing it with the topic of personal identity investigated by philosophers. The two topics are similar in that both are concerned with what must be assumed in order to be able to ascribe an unchanging identity to an individual or person. The two topics are also different in that economics faces an additional problem in needing to be able to say how or when individuals are distinct from one another as independent economic agents. Nonetheless, since Chapter 7 provides a means of saying how individuals in economic life might be thought distinct from one another, the individuation problem ultimately does not seem to separate the two topics in an important way, and they might accordingly be thought to converge on the question of what makes an individual relatively unchanging. From this perspective, we can then ask how individual identity, understood as a matter of reidentification across change in economic life, relates to personal identity, understood as a matter of reidentification across change in one’s life generally. One view, which it seems should be rejected from the outset, is that the two domains are essentially indistinguishable, so that explaining reidentification in
188
Revisiting the issues
economic life also explains it in life generally. While it is certainly true that one’s character as an economic agent bears on one’s individual nature as a whole, very few appear to believe that the latter can be reduced to the former. Indeed, most, I think, would prefer that their status as economic agents be seen as a means to achieving their goals as individuals. That is, should individuals be reidentifiable in economic life across the variety of their lifetime group affiliations and associations, they would hope that this would enable them to maintain “the greatest personal well-being possible” in their lives as a whole. I place this latter expression in quotation marks to avoid saying that what individuals wish for is a sense of personal identity across their lives outside what identity they may establish in their economic lives. Some may indeed wish to maintain personal identity in this manner. However, others may be equally content with the notion that when “you went on living one day after another … in time you were somebody else, your previous self only like a close relative, a sister or brother, with whom you shared a past. But a different person, a separate life” (Frazier 1997: 422). Parfit, in fact, revolutionized the philosophical literature on personal identity by raising the possibility that something other than identity might be an individual’s fundamental objective (Parfit 1986). Thus, I believe that the question of whether personal identity is important must be left open when we think about the relationship between individual economic identity and the balance of individual life. However, individual economic identity still seems an important objective for the reasons I discussed in the last chapter. There I argued that power operates in the world in ways that create advantage for some at the expense of others, and that one important way in which this occurs is the practice of the powerful confining the less powerful to narrow identities inimical to the latter’s long-run development. When this occurs, an individual’s economic life increasingly determines the character of that individual’s life in general, and one’s status as an economic agent ceases to be a means to achieving one’s goals as an individual. Thus, individual economic identity, either as understood in Chapter 8 in terms of the development of a capability for embracing the rights and responsibilities of group membership, or in some other way, remains normatively important. Still, how individuals might look upon success in this regard in relation to the question of personal identity seems better left open.
10.3
Two historical traditions
In Chapters 2 and 6, I gave accounts of how ideas regarding the nature of the individual evolved historically in orthodox and heterodox economics respectively. In my view, each tradition begins with a conception of the individual that is essentially metaphorical in nature. The conception of the atomistic individual trades in ideas drawn from Descartes’ philosophical-theological views framed in the context of Newton’s mechanical science thinking. The conception of the embedded individual depends on an image of individuals being located somehow “within” networks of
Revisiting the issues
189
social relationships. However, as metaphors and the contexts in which they operate generally evolve after their initial appearance, so these two metaphors and the contexts in which they each operate have also clearly evolved. This particular evolution, however, can be contrasted with the relatively benign sort of evolution that characterizes the history of many metaphors in economics and elsewhere. For example, the original notion that money “circulates” like water has been eliminated largely without notice in the gradual integration of the concept of money into the theory of a market economy, thereby producing a “dead” metaphor with its meaning determined by its theoretical context (Lagueux 1999). In the case of the atomistic and embedded individual conceptions, however, the central importance of these notions within their surrounding conceptual contexts seems to have sustained their metaphorical lives for a longer time. This metaphorical persistence in turn has meant that the conceptual tension that metaphors produce when they combine meanings from one context with those of another has played a role in the evolution of the atomistic and embedded individual frameworks which can be charted as a succession of efforts to effectively tame or discipline these two metaphors. Thus, my historical arguments in Chapters 2 and 6 portray the evolution of both theories as a struggle on the part of their respective proponents to increasingly adjust metaphorical ideas to the contexts into which they were inserted. In the case of the atomistic individual conception, the idea of the individual as subjective inwardness could not easily be reconciled with an understanding of the economy as a cause-andeffect system. In the case of the embedded individual conception, the idea of the individual as always embedded within social structures could not easily be reconciled with the idea of being an individual. Thus, the evolution of both conceptions involved recurring critique of the central metaphor. In the atomistic individual conception, this involved first Smith’s invisible hand application of cause-and-effect logic to individuals’ external interaction with one another in the market, then the marginalists’ extension of that same logic into the mind in the development of the theory of choice, and finally the elimination of subjectivity altogether in functionalist characterizations of the individual in early mainstream economics. In the embedded individual conception, this initially involved hesitation over the purview and extent of holist explanations of the world, then attempts to invest individuals with some form of agency, and finally (in my view) the beginnings of efforts to understand individuals’ reflexive psychological processes. However, these two evolutions in ideas had quite opposite destinations in regard to their treatment of individuals. Because the atomistic conception equated the individual with subjective inwardness, disciplining that account had the effect of removing the very basis on which individuals were conceptualized as unique and independent. In contrast, because the embedded individual conception began with a holistic account of individuals effectively lost “within” social structures, disciplining that account had the effect of separating out a conceptualization of relatively independent individuals. The irony in this, on the one hand, is that the liberal tradition,
190
Revisiting the issues
with its deep historical roots in political and ethical thinking about individual rights and freedom, ultimately failed in its task of developing a supporting scientific account of individuals because of its particular metaphorical origins. In effect, science ultimately overruled ethics. On the other hand, the holist tradition, with roots in political and ethical thinking about needs and rights of entire groups of people, seems to have produced a conception of individuals as separate and independent within an increasingly dense social world. Science also motivates this latter development, since the fundamental issue is ultimately how structure–agent relationships are to be explained, but in this instance science may be compatible with an ethics that promotes the integrity of individuals. I have told these two stories largely as histories of how ideas about individuals have developed. However, in my view it is right to see the history of economics in wider terms that explain the adoption and transmission of ideas as a social-historical process. That larger framework is missing from Chapters 2 and 6 because my goal there was first to chart the opposed conceptual evolutions that characterize the two main traditions of thinking about individuals in economics. The larger framework encompassing this pair of histories, however, is not far from the surface. The long history of conflict between methodological individualism and methodological holism is very much part of an equally long historical contest since the Enlightenment between competing visions of individuals and society. Economics does not stand apart from politics in this history, nor does it stand apart from ethics. Thus, the fuller story of the evolution of these two traditions of thinking is a story of the modern world in a struggle over the organization of society and the place of individuals within it.
10.4
Significance
The five subjects I regard as central to this book are: 1 2 3 4 5
The reappraisal of the treatment of the individual by orthodox economics. The argument that heterodox economics, contrary to what many suppose, has important elements of a theory of the individual. The emphasis on reflexivity in individual behavior. Collective intentionality analysis as a means of explaining economic behavior as complex. The vision of the embedded individual in practical and normative terms.
10.4.1
The emperor’s clothes
Orthodox economics has long represented itself as individualist in both scientific and normative terms. Indeed, its position in this regard has always been key to its broad ideological support in society. I argue, however, that orthodox economics is now individualist in name only, and that, like the emperor of the famous fable, its past
Revisiting the issues
191
glories no longer really conceal its current shape and form. I previously suggested (Davis 2002b) that, just as the emperor was the last one to learn the truth, so it will be with mainstream economics. Fable and reality, however, may part ways when it comes to how the eventual discovery of the truth is greeted. The emperor was embarrassed. My suspicion is that orthodox economics will not be. However, the more interesting question may be, how will others react? 10.4.2
Heterodox economics
Heterodox economics has undergone continuous decline over the last decade, largely because, I believe, its underlying vision of society has frequently been associated with failed Soviet socialism. The silver lining in this is that, as time passes, this association is less and less plausible (though it was never accurate). Nonetheless, the general perception both in and out of economics is still that heterodox economics has little to offer in a scientific or normative way. What I think the arguments in the second half of this book suggest, then, is that heterodox economics may have a more promising future than most imagine. If the elements of an alternative conception of the individual described here coalesce around an increasingly resonant set of concerns regarding individual life in today’s socially complex world, the better intuitions that heterodox economists have about institutions and social structures could place them in a position to speak with greater authority about society’s concern over the increasing vulnerability of individuals. My point, though, is certainly not to recommend wholesale substitution of heterodox economics for orthodox economics. I prefer a pluralistic environment in which multiple research strategies compete openly without any one strategy monopolizing access to research, education, and influence, as is currently the case. However, there may be some reason to think that this kind of academic world will emerge in the future. Mainstream economics, without its anchor in a theory of the individual, has become an eclectic mix of different kinds of research strategies. Pluralism, then, may actually find an advocate precisely where it has so long been opposed. Heterodox economics would arguably prosper in this sort of environment. 10.4.3
Reflexivity in individual behavior
I argued in Chapter 6 that the way to make the structure–agent framework viable is to reinforce its implicit account of individual agency with a treatment of reflexive or self-referent behavior. This could involve importing more social psychological thinking into social theory and economics, in order to supplement what thinking exists in heterodox economics on the subject. Interestingly, there is also precedent in neoclassical economics for exploring reflexivity in the form of multiple selves analysis, though the subjectivist basis of this literature is inadequate to the agency emphasis in the social psychology literature on the subject. What indeed seems
192
Revisiting the issues
particularly interesting about this latter literature is that it describes how individuals are active in engaging conceptions of themselves that have strong social antecedents. Nonetheless, the details of how self-referent behavior might be more fully incorporated into economics remain to be worked out, and my discussion on the subject has mostly been restricted to demonstrating the implicit connections between individual reflexivity and collective intentionality analysis. This link is valuable for throwing light on the embedded individual conception because it tells us that how we relate to others (as outer objects) is tied to how we relate to ourselves (as inner objects), or how we use “we” language in social groups is tied to how we see ourselves as individual members of social groups. These connections may seem to set a rather large agenda for the explanation of individual behavior, but if in fact reflexivity is a key to individual autonomy in socially embedded circumstances, we may have no choice but to try to explain it. 10.4.4
Economic behavior as complex
I regard the application of collective intentionality analysis to economics as the main substantive contribution of this book. What is important about collective intentionality analysis is that it provides a way of understanding embedded individuals specifically as individuals. Thus though collective intentionality analysis is individualist ontologically speaking, it is also holist in being able to account for the influence that groups have on individuals. This analysis works, moreover, through a pair of well-accepted theoretical pathways: the standard idea that behavior is purposeful or intentional, and the explanation of interaction between individuals in terms of reciprocal sets of expectations that they have of one another. At the same time, the consequences of the analysis for economics are fairly radical. Though instrumental rationality theorists are sure to resist the idea that the use of first-person plural speech requires a characterization of behavior in deontological or rationally principled terms, the main motivation for their arguments seems more a matter of trying to preserve a single type of explanation for individual behavior than reflection on the nature of behavior. My view on this, again, is pluralist. Different types of social circumstance appear to involve different types of behavior. Rather than attempt to fit square pegs into round holes, then, I recommend that economic behavior be seen as complex in combining qualitatively different kinds of behavior, instrumental and deontological. Such a strategy, it seems, would be more versatile in fitting explanation to the world as opposed to imposing explanation on the world. 10.4.5
The embedded individual
As argued above, this conception was originally essentially metaphorical in nature. However, that it can now be more carefully articulated in terms of structure–agent thinking, reflexivity, and collective intentionality analysis creates the possibility that
Revisiting the issues
193
the embedded individual conception may increasingly function as a viable understanding of individuals in the future. Thus while we may regard the social science and postmodernist critiques of the traditional view of individuals as persuasive in their diagnosis, we may also disagree about what this calls for. For many critics of the Enlightenment view, the inference has been that we should reduce or abandon our traditional concern with individuals. However, the approach that strikes me as more reasonable is to pursue better accounts of the nature of individual life as socially embedded in an effort to preserve the individual as an object of theoretical and moral concern. This, I believe, is not just a personal view but also one widely held in contemporary human society. Should this be true, however, it would not be enough to ensure continuing concern for the individual in the future. If the individual is to remain a theoretical and moral center to our world, arguments and efforts to produce that outcome will have to be advanced and promoted. The socially embedded individual conception, it seems, offers one basis on which this might occur.
Notes
1 Framing the issues 1 It should be noted that Newton’s own thinking on this score was more complex than what ultimately emerged in popular form as the “triumphant Newtonian model” (Taylor 1989: 171). 2 C. L. Becker puts this particularly well in connection with the vision of Newton: “Obviously the disciples of the Newtonian philosophy had not ceased to worship. They had given another form and a new name to the object of worship: having denatured God, they deified nature” (Becker 1948: 63). 3 Even more interesting is the representative agent of macroeconomics made up of many identical individuals. I return to this in Chapters 2 and 5. 4 Political philosophers, particularly those in the liberal tradition, are concerned with individuation in connection with the normative principle that persons are distinct. Thus, Rawls and Nozick invoke the distinctness of persons in criticizing utilitarianism, which lacks such a principle. For four different formulations of the principle of distinctness in political philosophy, see Graham (2002: 29ff.). 5 Here I agree with Cartwright’s emphasis on natures and capacities of things as fundamental to our understanding of science (Cartwright 1989). 2 The atomistic individual 1 With apologies to Rizvi for the same title (1994a). 2 Backward induction describes the reasoning of rational players who see no reason not to defect from cooperative agreements on the last round of games with multiple play, recognize that in the penultimate round of play each would defect on the last round and thus that they also would defect in the penultimate round, and so on back through all rounds of play to conclude that defection in this first round of play is rational. 3 A mixed strategy is where individuals play a number of pure strategies with certain frequencies to provide a strategic advantage in not being predictable to one another. 3 Reidentification: preferences and human capital 1 An earlier version of the arguments in this chapter appeared in Davis (1995). 2 According to Fox (1988: 12), who relies on the Oxford English Dictionary, Locke was the first to use the term “consciousness” in the English language. The OED refers to book 2, chapter 27 of Locke’s Essay Concerning Human Understanding as the first usage. Locke’s discussion of personal identity first appeared in the second (1694) edition of the Essay. 3 That the soul was an independent substance or substratum—see Locke’s argument that the soul cannot be identified with thinking, because thinking does not occur in sleep (Locke 1975 [1694]: 2.3.10).
196
Notes
4 Investing in music appreciation is an example of what Stigler and Becker call a “beneficial addiction.” In contrast, if the partial derivative of S with respect to a good is negative, it is an example of a “harmful addiction” (Stigler and Becker 1977). This difference does not affect the arguments here. 4 Individuation: multiple selves 1 This way of looking at the matter recalls Hutchison’s reaction to Robbins’ well-known critique of interpersonal utility comparisons. Robbins argued that comparing different individuals’ intensities of satisfaction required making value judgments, but that value judgments had no place in economics (1935 [1932]). Hutchison responded that, if interpersonal comparisons of utility required value judgments, then intrapersonal comparisons of utility were likely to require them as well (1938 [1950]). 2 Alternatively, were Arrow’s impossibility analysis invoked to explain intrapersonal collective choice, a dictatorial authority would be needed to produce a single individual “social” choice function. 3 Thus, George (2001) argues that an advantage of the metapreference approach is that standard theory assumptions can be retained at the level of second-order preferences. 4 A related problem is that metapreference rankings need not be complete, so that human decision making is psychologically complex, possibly self-contradictory, and may thus depend upon “satisficing”-type notions (Lutz 1993: 148–9). 5 This is all bound up with the question of how Wittgenstein’s private language argument applies to the interpretation of subjective preferences (cf. Davis 1988). 6 As seen above, an individual who revises preferences upon observing the ill effects on individual welfare of time preference possesses endogenously changing preferences. 7 A third sort of change, wishful thinking, involves changes in belief or perception that subsequently accommodate unpreferred changes in an individual’s opportunity sets. George (2001) takes a different tack on “behind the back” sorts of change in preferences in arguing that market failure occurs when the market creates undesired preferences. This sort of case is intermediate between Elster’s two drives and desires cases in that the market substitutes for individual drives in changing preferences, and the preferences that emerge conflict with the individual’s metadesires. 8 See Cowen (1993) for references to the traditional welfare literature and further discussion of the problems associated with the ideal of consumer sovereignty in the context of changing preferences. The terms desire and preference are used interchangeably here. 9 “In general, we are dealing with exactly one person—neither more nor less. That person may have some cognitive coordination problems, and some motivational conflicts, but it is his job to sort them out” (Elster 1986: 30). “It thus seems best to refer to multiplicity with such terms as facets or aspects of the person, or the ‘different sides’ of our nature or even to refer to the agent considering a choice ‘from different points of view’, or making different partial judgments” (Steedman and Krause 1986: 201). “I do not claim that we are in fact composed of multiple distinct selves, each of which forms an integrated unit over time and has separate dispositions or values from the other selves of the same individual” (Kavka 1991: 148). 5 After the fall: the machinery of choice 1 For example, “I dispense with the notion that people fully understand the structure of the games they play, that they have a coherent model of others’ behavior, that they can make rational calculations of infinite complexity, and that all of this is common knowledge. Instead I postulate a world in which people base their decisions on limited data, use simple predictive models, and sometimes do foolish things. Over time, such simple adaptive learning processes can converge to quite complex equilibrium patterns of behavior” (Young 1998: xi).
Notes
197
2 A more difficult question, to be taken up below in Section 5.3, is whether functional states are really properly termed mental states in the first place. That is, does the functionalist view capture the constitutive features of the human mind? 3 The term “intensional” is used to refer to the meaning or connotation of an expression describing individuals’ mental lives, whereas the term “intentional” is rather used to refer to individuals’ beliefs, desires, and so on. 4 See Sent (1998) for the connections between Simon, Sargent, and Lucas regarding the role of artificial intelligence in models. 5 See Hoover (1995) for the argument that Simon was also a precursor of Lucas with respect to the Lucas critique. 6 Some of the characteristics of consciousness that Searle considers (Searle 1980: 128–41) are: consciousness is manifested in a limited number of sensory modalities, conscious states generally come to us in unified sequences, conscious states involve what-it-feels-like aspects, consciousness often has a figure–ground Gestalt structure, familiarity is a standard feature of conscious states, there is an “overflow” quality to what we are immediately conscious of, there is a center–periphery character to consciousness, there are boundary conditions to conscious states, mood pervades consciousness, and there is always a pleasure/unpleasure dimension to consciousness. 7 Thus, neoclassical subjectivity can be fairly characterized in Nagel’s terms as “the view from nowhere” (Nagel 1986). 6 The embedded individual 1 “Structuration” theory is different from structuralism in linguistics, social anthropology, and psychology, which investigates the relationships between elements of a system apart from the question of agency. 2 Kaplan distinguishes four primary forms of self-referent behavior (self-cognition, self-evaluation, self-feeling, and self-protective/enhancing), but they each have essentially the same structure. 3 Note also how this view differs from the metapreferences framework. The relation of secondorder preferences to first-order preferences is affective, but self-referent behavior involves individuals’ active engagement with their self-conceptions. 4 These different strategies also arise in connection with other standard (and related) topics of investigation in the social psychology literature such as “role playing” and “social comparisons” which similarly concern how others’ opinions influence individuals’ self-appraisal. 5 Lawson’s situated rationality is different from Popper’s situational analysis or situational logic. Situational analysis/logic is a methodological approach drawn from economic analysis that involves reconstruction of a problem situation faced by an agent to explain the agent’s action as a reasonable response to the situation (Caldwell 1991). Situated rationality in Lawson’s sense similarly explains action in terms of situation, though this is understood more in terms of the space for action in a framework of social rules and positions. 6 I used this framework in my own analysis of Keynes’ concept of convention as a system of interdependent belief expectations (Davis 1994). 7 Also see Bruni (2000) and Rizvi (2002), who rely on Smith to redevelop thinking about the individual. Akerlof and Kranton (2000) examine how individuals’ sense of self-worth affects their choices, but do so in a utility framework. 7 Individuation: collective intentionality 1 Locke was Shaftesbury’s tutor. 2 Gilbert explains “we” language in a similar way: “A person X’s full-blooded use of ‘we’ in ‘Shall we do A?’ with respect to Y, Z, and himself, is appropriate if and only if it expresses his recognition of the fact that he and the others are jointly ready to share in doing A in relevant circumstances” (Gilbert 1989: 199). Gilbert holds that individuals’ use of “we” language constitutes a “plural subject” (Gilbert 1989: 199ff.).
198
Notes
3 Tuomela draws on an account of mutual belief that has become fairly standard among philosophers and relies on the idea of a hierarchical set of beliefs iterated across individuals (Tuomela 1995: 41ff.). See Shwayder (1965: 257) and Lewis (1969: 52ff.) for early formulations. 4 For one example of how such an explanation might be produced, see Minkler (1999), where a “commitment function” is added to a standard utility function representation of individual behavior. The individual is said to engage in a two-step iterative procedure, with the first step corresponding to a response to group requirements and the second step corresponding to an instrumentally rational maximization of utility. 5 However, the private language argument still applies to ordinalism (cf. Davis 1988). 6 Here I put aside for the moment the argument of Chapter 4. 8 Reidentification: capabilities 1 A related issue that arises in connection with Sen’s interpretation of this framework in terms of functionings and capabilities (see below) is whether he should identify an “objective” set of functionings for social policy analysis. Nussbaum makes this criticism from the perspective of her preferred Aristotelian approach to capabilities which identifies one unique set of functionings and capabilities as conducive to the good life (Nussbaum 1993). Sen’s response is that leaving his framework more open ended makes it potentially more serviceable to a wider range of applications (Sen 1993: 47). 2 I have applied this particular framework to the need for health care (Davis 2000). 3 However, with the rise of behavioral economics, this state of affairs may be changing (cf. Rabin 1998). 4 Alternatively, they criticize a different ideal conception whose realization was sought by others, namely, that women should specialize in housework. 5 Sen’s thinking and the embedded individual also share this understanding with neo-Austrian economics, which regards individuals as active by virtue of possessing creative powers of imagination. 9 Before the fall: value in economics 1 It goes further to argue that all language—not just that involved in forming we-intentions—is value laden by virtue of the same compulsion to observe what one can and ought to say in connection with any established usage. Some might consider this argument trivially true, and therefore of little interest in connection with the topic of the relationship between facts and values. By contrast, the normative character of we-intentions goes directly to the fact–value distinction. 2 This is the “strong” Pareto criterion. The “weak” criterion requires that all individuals be satisfied that the state of affairs is preferable (Arrow 1963 [1951]: 36, 96). 3 Thus, Sen distinguishes between seeing utility as an object of value in itself and utility as a valuational device that is used to evaluate other objects of value (Sen 1987: 6–7).
References
Albert, A. and Ramstad, Y. (1998) “The social psychological underpinnings of Commons’s institutional economics II: The concordance of George Herbert Mead’s social self and John R. Commons’s will,” Journal of Economic Issues 31(4): 1–46. Akerlof, G. (1970) “The market for ‘lemons’: quality uncertainty and the market mechanism,” Quarterly Journal of Economics 84: 488–500. —— and Kranton, R. (2000) “Economics and identity,” Quarterly Journal of Economics 115(3): 715–53. Allais, M. (1953) “Le comportement de l’homme rationnel devant le risque: critique des postulats et axiomes de l’école americaine,” Econometrica 21: 503–56. Althusser, L. (1969) For Marx, trans. B. Brewster, New York: Pantheon Books. Amariglio, J. (1988) “The body, economic discourse, and power,” History of Political Economy 20(4): 583–613. —— (1990) “Economics as a postmodern discourse,” in W. Samuels (ed.) Economics as Discourse, Boston and Dordrecht: Kluwer, 15–46. Anscombe, G. (1979 [1963]) Intention, 2nd edn, Oxford: Blackwell. Archer, M. (1995) Realist Social Theory: The Morphogenetic Approach, Cambridge: Cambridge University Press. Aristotle (1962) Nichomachean Ethics, trans. M. Oswald, Indianapolis: Bobbs-Merrill. Armstrong, D. (1968) A Materialist Theory of Mind, London: Routledge and Kegan Paul. —— (1980) The Nature of Mind, Sydney: University of Queensland Press. Arrow, K. (1963 [1951]) Social Choice and Individual Values, 2nd edn, New Haven and London: Yale University Press. —— (1959) “Towards a theory of price adjustments,” in M. Abramovitz et al. (eds) The Allocation of Economic Resources, Stanford: Stanford University Press. —— (1968) “Economic equilibrium,” in International Encyclopedia of the Social Sciences, vol. 4, London: Macmillan, 376–89. —— (1984) The Economics of Information, Cambridge, MA: Belknap Press of Harvard University. Aumann, R. (1976) “Agreeing to disagree,” Annals of Statistics 4: 1236–9. Ayer, A. (1946 [1936]) Language, Truth and Logic, 2nd edn, New York: Dover. Bacharach, M. (1999) “Interactive team reasoning: a contribution to the theory of cooperation,” Research in Economics 53: 117–47. Backhouse, R. (1996) “Economics is a historical process,” in S. Medema and W. Samuels (eds) Foundations of Research in Economics: How Do Economists Do Economics? Cheltenham: Edward Elgar, 7–17.
200
References
—— (2003) “The stabilization of price theory, 1920–1955,” in W. Samuels, J. Biddle, and J. Davis (eds) The Blackwell Companion to the History of Economic Thought, Oxford: Blackwell. Baudrillard, J. (1983) Simulations, New York: Semiotext(e). Becker, C. (1948) The Heavenly City of the Eighteenth-century Philosophers, New Haven: Yale University Press. Becker, G. (1971 [1957]) The Economics of Discrimination, 2nd edn, Chicago: University of Chicago Press. —— (1965) “A theory of the allocation of time,” Economic Journal 65: 493–517. —— (1976a) “Altruism, egoism, and genetic fitness: economics and sociobiology,” Journal of Economic Literature 14: 817–26. —— (1976b) The Economic Approach to Human Behaviour, Chicago: University of Chicago Press. —— (1996) Accounting for Tastes, Cambridge, MA: Harvard University Press. Benhabib, S. (1996) “From identity politics to social feminism. A plea for the nineties,” in D. Trend (ed.) Radical Democracy: Identity, Citizenship and the State, London: Routledge, 28–41. —— (1999) “Civil society and the politics of identity and difference in a global context,” in N. Smelser and J. Alexander (eds) Diversity and Its Discontents, Princeton: Princeton University Press, 293–312. Ben-Ner, A. and Putterman, L. (1998) “Values and institutions in economic analysis,” in A. Ben-Ner and L. Putterman (eds) Economics, Values, and Organization, Cambridge: Cambridge University Press, 3–69. Bentham, J. (1970 [1789]) An Introduction to the Principles of Morals and Legislation, London: Athlone Press. Bhaskar, R. (1989 [1979]) The Possibility of Naturalism: A Philosophic Critique of the Contemporary Human Sciences, 2nd edn, Brighton: Harvester. Blaug, M. (1992 [1980]) The Methodology of Economics: Or How Economists Explain, 2nd edn, Cambridge: Cambridge University Press. —— (1998) “The positive–normative distinction,” in J. Davis, D. Hands, and U. Mäki (eds) Handbook of Economic Methodology, Cheltenham: Edward Elgar, 370–4. Block, N. and Fodor, J. (1972) “What psychological states are not,” Philosophical Review 81: 159–81. Boumans, M. (1998) “Lucas and artificial worlds,” in J. Davis (ed.) New Economics and Its History, Durham: Duke University Press, 63–88. —— (1999) “Built-in justification,” in M. Morrison and M. Morgan (eds) Models as Mediators: Perspectives on Natural and Social Science, Cambridge: Cambridge University Press, 66–96. —— and Morgan, M. (2001) “Ceteris paribus conditions: materiality and the application of economic theories,” Journal of Economic Methodology 8(1): 11–26. Bratman, M. (1993) “Shared intention,” Ethics 104: 97–113. —— (1999) Faces of Intention, Cambridge: Cambridge University Press. Bridgman, P. (1927) The Logic of Modern Physics, New York: Macmillan. Bruni, L. (2000) “Ego facing alter: how economists have depicted human interactions,” Annals of Public and Cooperative Economics 71(2): 285–313. Bunge, M. (1980) The Mind–Body Problem: A Psychological Approach, Oxford: Pergamon. Butler, J. (1896 [1736]) “The analogy of religion,” in W. Gladstone (ed.) The Works of Joseph Butler, 2 vols, Oxford: Clarendon. Caldwell, B. (1991) “Clarifying Popper,” Journal of Economic Literature 29: 1–33. Cartwright, D. and Zander, A. (1968) Group Dynamics Research and Theory, New York: Harper and Row. Cartwright, N. (1983) How the Laws of Physics Lie, Oxford: Clarendon Press.
References
201
—— (1989) Nature’s Capacities and their Measurement, Oxford: Clarendon Press. Cassel, G. (1932 [1923]) The Theory of the Social Economy, trans. S. Barron, New York: Harcourt, Brace and World. Chomsky, N. (1986) Knowledge of Language: Its Nature, Origin and Use, New York and Philadelphia: Praeger Special Studies. Churchland, P. M. (1984) Matter and Consciousness: A Contemporary Introduction to the Philosophy of Mind, Cambridge, MA: MIT Press. Coase, R. (1937) “The nature of the firm,” Economica 4: 386–405. Coates, J. (1996) The Claims of Common Sense: Moore, Wittgenstein, Keynes and the Social Sciences, Cambridge: Cambridge University Press. Collard, D. (1978) Altruism and Economy, Oxford: Martin Robertson. Cooley, C. (1902) Human Nature and the Social Order, New York: Scribner. Cottrell, A. (1995) “Intentionality and economics,” Economics and Philosophy 11: 159–76. Cowen, T. (1993) “The scope and limits of preference sovereignty,” Economics and Philosophy 9: 253–69. Craib, I. (1992) Anthony Giddens, London: Routledge. Cullenberg, S., Amariglio, J., and Ruccio, D. (eds) (2001) Postmodernism, Economics and Knowledge, London: Routledge. Dannefer, D. (1984) “Adult development and social theory: a paradigmatic reappraisal,” American Sociological Review 49: 100–16. Danner, P. (2002) The Economic Person, Lanham, MD: Rowman and Littlefield. Davidson, D. (1980) Essays on Actions and Events, Oxford: Clarendon Press. —— (1980) “How is weakness of the will possible?” in D. Davidson (ed.) Essays on Actions and Events, Oxford: Oxford University Press. —— (1986) “Deception and Division,” in J. Elster (ed.) The Multiple Self, Cambridge: Cambridge University Press, 79–92. Davis, J. (1988) “Sraffa, Wittgenstein and neoclassical economics,” Cambridge Journal of Economics 12: 29–36. —— (1990) “Cooter and Rappaport on the normative,” Economics and Philosophy 6: 139–46. —— (1994) Keynes’s Philosophical Development, Cambridge: Cambridge University Press. —— (1995) “Personal identity and standard economic theory,” Journal of Economic Methodology 2(1): 35–52. —— (2000) “Public policy and the lack of health insurance coverage,” Health Care Analysis 8(1): 55–64. —— (2001) “Agent identity in economics,” in U. Mäki (ed.) The Economic World View, Cambridge: Cambridge University Press, 114–31. —— (2002a) “Collective intentionality and the agency–structure model,” in E. Fullbrook (ed.) Intersubjectivity in Economics: Agents and Structures, London: Routledge, 11–27. —— (2002b) “The Emperor’s clothes,” Journal of the History of Economic Thought 24: 141–54. —— (2002c) “Using Sen’s real opportunities capabilities concept to explain personal identity in Folbre’s ‘structures of constraint’ analysis,” Review of Political Economy 14(4): 481–96. —— (2003) “Economic methodology since Kuhn,” in W. Samuels, J. Biddle, and J. Davis (eds) The Blackwell Companion to the History of Economic Thought, Oxford: Blackwell. Debreu, G. (1974) “Excess demand functions,” Journal of Mathematical Economics 1: 15–23. Demos, D. (1985) “The measurement of self-esteem: refining our methods,” Journal of Personal Psychology 48: 1490–1502. —— (1992) “The self-concept over time: research issues and directions,” American Review of Sociology 18: 303–26.
202
References
Dennett, D. (1976) “Conditions of personhood,” in A. Rorty (ed.) The Identities of Persons, Berkeley: University of California Press, 175–96. —— (1978) Brainstorms: Philosophical Essays on Mind and Psychology, Cambridge, MA: MIT Press. —— (1987) The Intentional Stance, Cambridge, MA: MIT Press. Dolfsma, W. (2002a) “Mediated Preferences – How Institutions Affect Consumption,” Journal of Economic Issues, 36(2): 449–57. —— (2002b) “The mountain of experience: how people learn in a complex, evolving environment,” International Journal of Social Economics, 29(8): 675–84. Duesenberry, J. (1960) “Comments,” in Demographic and Economic Change in Developed Countries, Washington, DC: National Bureau of Economic Research, 231–4. Dupuy, J.-P. (1988) “L’individu libéral, cet inconnu: Adam Smith à Friedrich Hayek,” in C. Audard, J.-P. Dupuy, and R. Sève (eds) Individu et Justice Sociale, Paris: Le Seuil, 73–125. —— (1992) Introduction aux Sciences Sociales: Logique des Phénomènes Collectifs, Paris: Ellipses. —— (1994) “The self-deconstruction of convention,” SubStance 74: 86–98. —— (2002) “Market, imitation and tradition: Hayek vs Keynes,” in E. Fullbrook (ed.) Intersubjectivity in Economics: Agents and Structures, London: Routledge, 139–58. Durkheim, E. (1938 [1901]) The Rules of Sociological Method, 8th edn, trans. S. Solovay and J. Mueller, G. Catlin (ed.), Chicago: University of Chicago Press. Elder, G. (1991) “The life course,” in E. Borgotta and M. Borgotta (eds) The Encyclopedia of Sociology, New York: Macmillan. Elster, J. (1979) Ulysses and the Sirens, Cambridge: Cambridge University Press. —— (1982) “Marxism, functionalism and game theory,” Theory and Society 11(4): 453–82. —— (1983) Sour Grapes: Studies in the Subversion of Rationality, Cambridge: Cambridge University Press. —— (ed.) (1986) The Multiple Self, Cambridge: Cambridge University Press. Etzioni, A. (1988) The Moral Dimension: Toward A New Economics, New York: Free Press. Evensky, J. (1993) “Adam Smith on the human foundation of a successful liberal society,” History of Political Economy 25(3): 395–412. Ferber, M. and Birnbaum, B. (1977) “The ‘new home economics:’ retrospect and prospect,” Journal of Consumer Research 4(1): 19–28. Field, H. (1978) “Mental representation,” Erkenntnis 13: 9–61. Fisher, M. (1989) “Games economists play: a noncooperative view,” RAND Journal of Economics 20(1): 113–24. Fodor, J. (1975) The Language of Thought, New York: Thomas Y. Crowell. Folbre, N. (1982) “Exploitation comes home: a critique of the Marxian theory of family labor,” Cambridge Journal of Economics 6(4): 317–29. —— (1994) Who Pays for the Kids? Gender and the Structures of Constraint, London: Routledge. Fontaine, P. (1997) “Identification and economic behavior: sympathy and empathy in historical perspective,” Economics and Philosophy 13(2): 261–80. —— (2000) “Making use of the past: theorists and historians on the economics of altruism,” European Journal of the History of Economic Thought 7(3): 407–22. —— (2001) “The changing place of empathy in welfare economics,” History of Political Economy 33(3): 387–409. Foster, J. and Sen, A. (1997 [1973]) “On economic inequality after a quarter of a century,” in J. Foster and A. Sen (eds) On Economic Inequality, enlarged edn, Oxford: Oxford University Press, 107–219. Fox, C. (1988) Locke and the Scriblerians, Berkeley: University of California. Frankfurt, H. (1971) “Freedom of the will and the concept of a person,” Journal of Philosophy 68: 5–20.
References
203
Fraser, N. (1997) Justice Interruptus: Critical Reflections on the “Postsocialist” Condition, London: Routledge. —— (2000) “Recognition without ethics?” in M. Garber, B. Hanssen, and R. Walkowitz (eds) The Turn to Ethics, London: Routledge, 95–126. Frazier, C. (1997) Cold Mountain, New York: Vintage. Friedman, M. (1949) “The Marshallian demand curve,” Journal of Political Economy 57: 463–95. —— (1953) Essays in Positive Economics, Chicago: University of Chicago Press. Fudenberg, D. and Tirole, J. (1989) “Noncooperative game theory for industrial organization: an introduction and overview,” in R. Schmalensee and R. Willig (eds) Handbook of Industrial Organization, vol. 1, Amsterdam: North-Holland, 259–327. Fullbrook, E. (1996) “The metaphysics of consumer desire and the French Subjectivists,” International Advances in Economic Research 2(3): 287–94. —— (1997) “Post-modernising Homo economicus,” in S. Earnshaw (ed.) Just Postmodernism, Amsterdam and Atlanta: Rodolfi Press, 67–87. —— (ed.) (2002) Intersubjectivity in Economics: Agents and Structures, London: Routledge. Galbraith, J. (1969 [1958]) The Affluent Society, 2nd edn rev., Boston: Houghton Mifflin. Gecas, V. (1982) “The self-concept,” American Review of Sociology 8: 1–33. —— (1996) “Self-concept,” in A. Kuper and J. Kuper (eds) The Social Science Encyclopedia, 2nd edn, London: Routledge. George, D. (2001) Preference Pollution: How Markets Create Desires We Dislike, Ann Arbor: University of Michigan. Gibbard, A. and Varian, H. (1978) “Economic models,” Journal of Philosophy 75(11): 664–77. Giddens, A. (1976) Central Problems in Social Theory, Berkeley: University of California Press. —— (1984) The Constitution of Society, Cambridge: Policy Press. —— (1991) Modernity and Self-identity: Self and Society in the Late Modern Age, Stanford: Stanford University Press. Giere, R. (1988) Explaining Science: A Cognitive Approach, Chicago: University of Chicago Press. Gilbert, M. (1989) On Social Facts, London: Routledge. —— (1996) Living Together: Rationality, Sociality, and Obligation, Lanham, MD: Rowman & Littlefield. Giocoli, N. (2003) Modeling Rational Agents: From Interwar Economics to Early Modern Game Theory, Cheltenham: Elgar. Goffman, I. (1959) The Presentation of Self in Everyday Life, New York: Doubleday. Graham, (2002) Practical Reasoning in a Social World, Cambridge: Cambridge University Press. Grandmont, J. (1992) “Transformations of the commodity space, behavioral heterogeneity, and the aggregation problem,” Journal of Economic Theory 57(1): 1–35. Granovetter, M. (1985) “Economic action and social structure: the problem of embeddedness,” American Journal of Sociology 91(3): 481–510. Grice, H. (1975) “Method in philosophical psychology (from the banal to the bizarre),” Proceedings and Addresses of the American Philosophical Association, vol. 48, November 1975: 23–53. Grossman, G. (1972) The Demand for Health: A Theoretical and Empirical Investigation, New York: National Bureau of Economic Research. Guala, F. (2000) “The logic of normative falsification: rationality and experiments in decision theory,” Journal of Economic Methodology 7(1): 59–94. Hands, D. (2001) Reflection Without Rules: Economic Methodology and Contemporary Science Theory, Cambridge: Cambridge University Press. ——(2002) “Operationalisms in economics,” paper given at 6th Annual Conference of European Society for the History of Economic Thought, University of Crete, Rethymnon, Greece, 14–17 March (unpublished).
204
References
—— and Mirowski, P. (1997) “Harold Hotelling and the neoclassical dream,” in R. Backhouse, D. Hausman, U. Mäki, and A. Salanti (eds) Economics and Methodology: Crossing Boundaries, London: Macmillan, 322–97. —— (1998) “A paradox of budgets: the postwar stabilization of American neoclassical demand theory,” in M. Morgan and M. Rutherford (eds) From Interwar Pluralism to Postwar Neoclassicism, Durham: Duke University Press, 260–92. Haraway, D. (1991) Simians, Cyborgs, and Women: The Reinvention of Nature, London: Free Association Books. Harding, S. (1986) The Science Question in Feminism, Ithaca: Cornell University Press. Hargreaves Heap, S. (1989) Rationality in Economics, Oxford: Blackwell. —— (2001) “Expressive rationality: is self-worth just another kind of preference?” in U. Mäki (ed.) The Economic World View, Cambridge: Cambridge University Press, 98–113. —— and Varoufakis, Y. (1995) Game Theory: A Critical Introduction, London: Routledge. Harman, G. (1977) “How to use propositions,” American Philosophical Quarterly 14: 173–6. Harsanyi, J. (1955) “Cardinal welfare, individualistic ethics and interpersonal comparisons of utility,” Journal of Political Economy 63: 309–21. —— and Selten, R. (1988) A General Theory of Equilibrium Selection in Games, Cambridge, MA: MIT Press. Hausman, D. (1992) The Inexact and Separate Science of Economics, Cambridge: Cambridge University Press. Hayek, F. (1952) The Counter-revolution in Science: Studies on the Abuse of Reason, Glencoe, IL: Free Press. Hempel, C. (1965) Aspects of Scientific Explanation, New York: Free Press. Hesse, M. (1966) Models and Analogies in Science, Notre Dame: University of Notre Dame Press. Hicks, J. and Allen, R. (1934) “A reconsideration of the theory of value,” Economica 1: 52–76, 196–219. Hirschleifer, J. (1976) “Shakespeare vs. Becker on altruism: the importance of having the last word,” Journal of Economic Literature 14: 500–2. Hodgson, G. (1988) Economics and Institutions: A Manifesto for a Modern Institutional Economics, Cambridge: Polity. —— (1996) “Towards a worthwhile economics,” in S. Medema and W. Samuels (eds) Foundations of Research in Economics: How Do Economists Do Economics? Cheltenham: Edward Elgar, 103–21. —— (1999) “Structures and institutions: reflections on institutionalism, structuration theory and critical realism,” University of Hertfordshire working paper. —— (2000) “What is the essence of institutional economics?” Journal of Economic Issues 34(2): 317–29. —— (2002) “Reconstitutive downward causation: social structure and the development of individual agency,” in E. Fullbrook (ed.) Intersubjectivity in Economics: Agents and Structures, London: Routledge, 159–80. —— (2003) “The hidden persuaders,” Cambridge Journal of Economics. Hollis, M. (1998) Trust within Reason, Cambridge: Cambridge University Press. —— and Sugden, R. (1993) “Rationality in action,” Mind 102(405): 1–35. Hoover, K. (1995) “Facts and artifacts: calibration and the empirical assessment of realbusiness-cycle models,” Oxford Economic Papers 47: 24–44. Hume, D. (1888 [1739]) A Treatise of Human Nature, L. Selby-Bigge (ed.), Oxford: Clarendon. Hutchison, T. (1938, 1950) The Significance of the Basic Postulates of Economic Theory, New York: Augustus M. Kelley. Jackendoff, R. (1987) Consciousness and the Computational Mind, Cambridge, MA: MIT Press.
References
205
Jameson, F. (1984). “Postmodernism, or the cultural logic of late capitalism,” New Left Review 146: 53–92. Janssen, M. (1993) Microfoundations: A Critical Inquiry, London: Routledge. Johnson-Laird, P. (1988) Mental Models: Toward a Cognitive Science of Language, Inference, and Consciousness, Cambridge, MA: MIT Press. Kant, I. (1948 [1785]) Groundwork of the Metaphysic of Morals, trans. H. Paton, New York: Harper & Row. Kaplan, H. (1986) Social Psychology of Self-referent Behavior, New York: Plenum. Kavka, G. (1991) “Is individual choice less problematic than collective choice?” Economics and Philosophy 7(1): 143–65. Keynes, J. (1936) The General Theory of Employment, Interest and Money, London: Macmillan. Kincaid, H. (1996) Philosophical Foundation of the Social Sciences: Analysing Controversies in Social Research, Cambridge: Cambridge University Press. Kirman, A. (1989) “The intrinsic limits of modern economic theory: the Emperor has no clothes,” Economic Journal 99: Conference, 126–39. —— (1992) “Whom or what does the representative individual represent?” Journal of Economic Perspectives 6: 117–36. Kitcher, P. (1993) The Advancement of Science: Science without Legend, Objectivity without Illusions, Oxford: Oxford University Press. Knight, F. (1944) “Realism and relevance in the theory of demand,” Journal of Political Economy 52: 289–318. Krugman, P. (1996) “How to be a crazy economist,” in S. Medema and W. Samuels (eds) Foundations of Research in Economics: How Do Economists Do Economics? Cheltenham: Edward Elgar, 131–41. Lagueux, M. (1999) “Do metaphors affect economic theory?” Economics and Philosophy 15(1): 1–22. Lancaster, K. (1966) “A new approach to consumer theory,” Journal of Political Economy 74: 132–57. Lasch, C. (1979). The Culture of Narcissism: American Life in an Age of Diminishing Expectation, New York: Norton. Lawson, T. (1992) “Realism, closed systems and expectation,” Research in the History of Economic Thought and Methodology 10: 149–69. —— (1997) Economics and Reality, London: Routledge. Lester, R. (1946) “Shortcomings of marginal analysis for wage–employment problems,” American Economic Review 36: 62–82. —— (1947) “Marginal, minimum wages, and labor markets,” American Economic Review: 135–48. Levy, T. (2002) “The theory of conventions and a new theory of the firm,” in E. Fullbrook (ed.) Intersubjectivity in Economics, London: Routledge, 254–72. Lewis, D. (1969) Convention: A Philosophical Study, Cambridge, MA: Harvard University Press. —— (1972) “Psychological and theoretical identification,” Australasian Journal of Philosophy 50: 249–58. Locke, J. (1975 [1694]) An Essay Concerning Human Understanding, P. H. Nidditch (ed.), Oxford: Clarendon Press. Lucas, R. (1977) “Understanding business cycles,” in K. Brunner and A. Melzer (eds) Stabilization of the Domestic and International Economy, Carnegie–Rochester Conference Series on Public Policy, vol. 5, Amsterdam: North-Holland, 7–29. —— (1980) “Methods and problems in business cycle theory,” Journal of Money, Credit and Banking 12(2): 696–713. —— (1981) Studies in Business Cycle Theory, Oxford: Blackwell. —— (1986) “Adaptive behavior and economic theory,” Journal of Business 59(4): 401–26.
206
References
—— (1987) Models of Business Cycles, Oxford: Blackwell. —— (1988) “On the mechanics of economic development,” Journal of Monetary Economics 22: 3–42. Lutz, M. (1993) “The utility of multiple utility: a comment on Brennan,” Economics and Philosophy 9(1): 145–54. —— (1999) Economics for the Common Good, London: Routledge. Maas, H. (2001) “Mechanical reasoning: William Stanley Jevons and the making of modern economics,” Ph.D. dissertation, University of Amsterdam. Mäki, U. (1989) “On the problem of realism in economics,” Ricerche Economiche 43: 176–98. —— (1992) “Friedman and realism,” Research in the History of Economic Thought 10: 1–36. —— (1993) “Social theories of science and the fate of institutionalism,” in U. Mäki, B. Gustafsson, and C. Knudson (eds) Rationality, Institutions and Economic Methodology, London: Routledge, 3–42. —— (1994) “Isolation, idealization and truth in economics,” Idealization in Economics, special issue, Poznan Studies in the Philosophy of the Sciences and the Humanities 38: 147–68. Malinvaud, E. (1972) Lectures on Microeconomic Theory, Amsterdam: North-Holland. Mantel, R. (1974) “On the characterization of aggregate excess demand,” Journal of Economic Theory 7(3): 348–53. —— (1976) “Homothetic preferences and community excess demand functions,” Journal of Economic Theory 12(2): 197–201. Margolis, H. (1982) Selfishness, Altruism and Rationality, Cambridge: Cambridge University Press. —— (1990) “Dual utilities and rational choice,” in J. J. Mansbridge (ed.) Beyond Self-interest, Chicago: University of Chicago Press. Marshall, A. (1920 [1890]) The Principles of Economics, London: Macmillan. Marx, K. (1973 [1852]) “The eighteenth Brumaire of Louis Bonaparte,” trans. D. Fernbach, in D. Fernbach (ed.) Karl Marx, Surveys from Exile; Political Writings Vol. 2, Harmondsworth: Penguin/NLR, 143–249. Mayhew, A. (1987) “Culture: core concept under attack,” Journal of Economic Issues 21(2): 587–603. —— (2000) “Comments on papers presented at the ‘Institutional Economics at the Millennium: Its Past and Future’ Session, January 2000,” Journal of Economic Issues 34(2): 331–3. Maynard Smith, J. (1982) Evolution and the Theory of Games, Cambridge: Cambridge University Press. McCloskey, D. (1990) If You’re So Smart, Chicago: University of Chicago Press. Mead, G. (1934) Mind, Self, and Society, Chicago: University of Chicago Press. Mehta, J. (1999) “Look at me look at you,” in R. Garnett (ed.) What Do Economists Know? London: Routledge. Minkler, L. (1999) “The problem with utility: toward a non-consequentialist/utility theory synthesis,” Review of Social Economy 52(1): 4–24. Mirowski, P. (2002) Machine Dreams: Economics Becomes a Cyborg Science, Cambridge: Cambridge University Press. —— and Hands, D. (1998) “A paradox of budgets: the postwar stabilization of American neoclassical demand theory,” History of Political Economy 30: 260–92. Moore, G. (1903) Principia Ethica, Cambridge: Cambridge University Press. Moore, H. (1914) Economic Cycles: Theory, Law and Cause, New York: Macmillan. Morgan, M. and Rutherford, M. (eds) (1998) From Interwar Pluralism to Postwar Neoclassicism, Durham: Duke University Press. Morrison, M. and Morgan, M. (eds) (1999) Models as Mediators: Perspectives on Natural and Social Science, Cambridge: Cambridge University Press.
References
207
Nagel, J. (2000). “False faces: ethnic identity, authenticity, and fraud in Native American discourse and politics,” in J. Davis (ed.) Identity and Social Change, New Brunswick, NJ: Transaction, 81–106. Nagel, T. (1986) The View from Nowhere, Oxford and New York: Oxford University Press. Nelson, J. (1996) Feminism, Objectivity and Economics, London: Routledge. Nietzsche, F. (1968 [1889]) “Twilight of the idols,” in The Portable Nietzsche, trans. W. Kaufmann, New York: Penguin. Noonan, H. (1993) Personal Identity, Aldershot: Dartmouth. Nussbaum, M. (1993) “Non-relative virtues: an Aristotelian approach,” in M. Nussbaum and A. Sen (eds) The Quality of Life, Oxford: Oxford University Press, 242–69. O’Boyle, E. (1994) “Homo socio-economicus: foundational to social economics and social economy,” Review of Social Economy 52(3): 286–313. —— (1998) Personalist Economics: Moral Convictions, Economic Realities, and Social Action, Boston: Kluwer. Orléan, A. (1989) “Mimetic contagion and speculative bubbles,” Theory and Decision 27: 63–92. —— (1990) “Le role des influences interpersonnelles dans la formation des cours boursiers,” Revue Économique 5: 839–68. —— (1992) “Contagion des opinions et fonctionnement des marchés financiers,” Revue Économique 4: 685–98. Packard, V. (1957) The Hidden Persuaders, London: Longmans, Green. Pareto, V. (1971 [1909]) Manual of Political Economy, trans. A. Schweier, New York: Augustus M. Kelley. Parfit, D. (1986) Reasons and Person, Oxford: Oxford University Press. Patinkin, D. (1965) Money, Interest, and Prices: An Integration of Monetary and Value Theory, New York: Harper & Row. Peacocke, A. and Gillett, G. (eds) (1987) Persons and Personality, Oxford: Blackwell. Pickering, A. (1995) The Mangle of Practice, Chicago: University of Chicago Press. Pigou, A. (1932 [1920]) The Economics of Welfare, 4th edn, London: Macmillan. Place, U. (1956) “Is consciousness a brain process?” British Journal of Psychology 47: 44–50. Polanyi, K. (1968) “Anthropology and economic theory,” in M. Fried (ed.) Readings in Anthropology, New York: Thomas Y. Crowell, 215–38. Pollak, R. (1968) “Consistent planning,” Review of Economic Studies 35: 201–8. Popper, K. (1945) The Open Society and Its Enemies, London: G. Routledge & Sons. Pressman, S. and Summerfield, G. (2000) “The economic contributions of Amartya Sen,” Review of Political Economy 12(1): 89–113. Putnam, H. (1967) “The Mental Life of Some Machines,” in H. Castaneda (ed.) Intentionality, Minds, and Perception, Detroit: Wayne State University Press. Pylyshyn, Z. (1984) Computation and Cognition: Toward a Foundation for Cognitive Science, Cambridge, MA: MIT Press. Quine, W. (1960) Word and Object, Cambridge, MA: MIT Press. —— (1969) Ontological Relativity and Other Essays, New York: Columbia University Press. Rabin, M. (1998) “Psychology and economics,” Journal of Economic Literature 36: 11–46. Reder, M. (1982) “Chicago economics: permanence and change,” Journal of Political Economy 35: 1–38. Reuten, G. (2003) “Marxian economics,” in W. Samuels, J. Biddle, and J. Davis (eds) The Blackwell Companion to the History of Economic Thought, Oxford: Blackwell. Revue Économique (1989) “L’économie des conventions,” 40(2). Riesman, D., Glazer, N., and Denney, R. (1950) The Lonely Crowd: A Study of the Changing American Culture, New Haven: Yale University Press.
208
References
Rizvi, S. (1991) “Specialisation and the existence problem in general-equilibrium theory,” Contributions to Political Economy 10: 1–10. —— (1994a) “Game theory to the rescue?” Contributions to Political Economy, 13: 1–28. —— (1994b) “The microfoundations project in general equilibrium theory,” Cambridge Journal of Economics 18: 357–77. —— (2002) “Adam Smith’s sympathy: towards a normative economics,” in E. Fullbrook (ed.) Intersubjectivity in Economics: Agents and Structures, London: Routledge, 241–53. Robbins, L. (1935 [1932]) An Essay on the Nature and Significance of Economic Science, 2nd edn, London: Macmillan. —— (1938) “Interpersonal comparisons of utility: a comment,” Economic Journal 48: 635–41. Rorty, R. (1979) Philosophy and the Mirror of Nature, Princeton: Princeton University Press. Rosenberg, M. (1979) Conceiving the Self, New York: Basic. Runde, J. (1998) “Assessing causal economic explanations,” Oxford Economic Papers 50: 151–72. Rutherford, M. (1994) Institutions in Economics: The Old and New Institutionalism, Cambridge: Cambridge University Press. —— (2000) “Institutionalism between the wars,” Journal of Economic Issues 34(2): 291–303. Samuelson, P. (1938) “A note on the pure theory of consumer behavior,” Economica 5: 344–56. —— (1947) Foundations of Economic Analysis, Cambridge, MA: Harvard University Press. —— (1948) “Consumption theory in terms of revealed preference,” Economica 15: 243–53. —— (1950) “The problem of integrability in utility theory,” Economica 17: 355–85. Sartre, J.-P. (1956) Being and Nothingness: An Essay on Phenomenological Ontology, trans. H. Barnes, New York: Philosophical Library. Scanlon, T. (2001) “Symposium on Amartya Sen’s philosophy: 3 Sen and consequentialism,” Economics and Philosophy 17(1): 39–50. Schelling, T. (1960) Strategy of Conflict, Oxford: Oxford University Press. —— (1978) “Egonomics, or the art of self-management,” American Economic Review: Papers and Proceedings 68: 290–4. —— (1980) “The intimate contest for self-command,” Public Interest 60: 94–118. —— (1984) “Self-command in practice, in policy and in a theory of rational choice,” American Economic Review 74: 1–11. Schick, F. (1991) Understanding Action, Cambridge: Cambridge University Press. Schultz, H. (1928) “Rational economics,” American Economic Review 18: 643–8. —— (1938) The Theory and Measurement of Demand, Chicago: University of Chicago Press. Schumpeter, J. (1954) History of Economic Analysis, New York: Oxford. Searle, J. (1980) “Minds, brains, and programs,” Behavioral and Brain Sciences 3: 450–6. —— (1990) “Collective intentions and actions,” in P. Cohen, J. Morgan, and M.E. Pollack (eds) Intentions in Communication, Cambridge, MA: MIT Press. —— (1992) The Rediscovery of the Mind, Cambridge, MA: MIT Press. —— (1995) The Construction of Social Reality, New York: Free Press. Segal, G. (1991) “Review of Garfield, J., Belief in Psychology,” Philosophical Review C3: 463–6. Seiz, J. (2003) “Postwar heterodox economics: feminist,” in J. Biddle, J. Davis, and W. Samuels (eds) The Blackwell Companion to the History of Economic Thought, Oxford: Blackwell. Sellars, W. (1968) Science and Metaphysics, London: Routledge and Kegan Paul. Sen, A. (1970) Collective Choice and Social Welfare, San Francisco: Holden-Day. —— (1977) “Rational fools,” Philosophy and Public Affairs 6: 317–44. —— (1980) “Equality of What?” in S. McMurrin (ed.), Tanner Lectures on Human Values, Cambridge: Cambridge University Press. —— (1982) “Rights and agency,” Philosophy and Public Affairs 11: 3–39.
References
209
—— (1985) Commodities and Capabilities, Amsterdam: North-Holland. —— (1985) “Well-being, agency and freedom: the Dewey Lectures, 1984,” Journal of Philosophy 82: 169–221. —— (1987) On Ethics and Economics, Oxford: Blackwell. —— (1993) “Capability and well-being,” in M. Nussbaum and A. Sen (eds) The Quality of Life, Oxford: Oxford University Press, 30–53. —— (1999) Development as Freedom, New York: Knopf. —— (2001) “Symposium on Amartya Sen’s philosophy: 4 Reply,” Economics and Philosophy 17(1): 51–66. Sent, E.-M. (1998) The Evolving Rationality of Rational Expectations, Cambridge: Cambridge University Press. —— (2001) “Sent simulating Simon simulating scientists,” Studies in History and Philosophy of Science 32(3): 479–500. Shaftesbury, A. (1963 [1900]) Characteristics of Men, Manners, Opinions, Times, J. Robertson (ed.), reprint edn, Gloucester, MA: Peter Smith. Shrauger, J. and Schoeneman, T. (1979) “Symbolic interactionist view of the self-concept: through the looking glass darkly,” Psychological Bulletin 86: 549–73. Shubik, M. (1953) “The role of game theory in economics,” Kyklos 6(21): 21–34. Shwayder, D. (1965) The Stratification of Behaviour, London: Routledge and Kegan Paul. Simon, H. (1957) Models of Man, New York: Wiley. —— (1969) The Sciences of the Artificial, Cambridge, MA: MIT Press. —— (1983) Reason in Human Affairs, Stanford: Stanford University Press. ——(1998) “Psychology and economics,” in J. Davis, D. W. Hands, and U. Mäki (eds) The Handbook of Economic Methodology, Cheltenham: Edward Elgar. Slutsky, E. (1953 [1915]) “On the theory of the budget of the consumer,” in G. Stigler and K. Boulding (eds) Readings in Price Theory, London: Allen and Unwin, 27–56. Smart, J. J. C. (1959) “Sensations and brain processes,” Philosophical Review 68: 141–56. Smith, A. (1976 [1759]) The Theory of Moral Sentiments, D. Raphael and A. Macfie (eds), Oxford: Clarendon Press. —— (1976 [1776]) An Inquiry into the Nature and Causes of the Wealth of Nations, R. Campbell, A. Skinner, and W. Todd (eds), Oxford: Clarendon Press. Sonnenschein, H. (1972) “Market excess demand functions,” Econometrica 40(3): 549–63. —— (1973) “Do Walras identity and continuity characterize the class of excess demand functions?” Journal of Economic Theory 6(4): 345–54. ——(1985) “Comment,” in K. J. Arrow and S. Honkapohja (eds) Frontiers of Economics, Oxford: Blackwell. Steedman, I. and Krause, U. (1986) “Goethe’s Faust, Arrow’s possibility theorem and the individual decision-taker,” in J. Elster (ed.) The Multiple Self, Cambridge: Cambridge University Press, 197–231. Stevenson, C. (1937) “The emotive meaning of ethical terms,” Mind 46: 10–31. Stigler, G. (1950) “The development of utility theory,” Journal of Political Economy 58: 307–27, 373–96. —— (1952) The Theory of Price, rev. edn, New York: Macmillan. —— and Becker, G. (1977) “De gustibus non est disputandum,” American Economic Review 67: 76–90. Strawson, P. (1959) Individuals: An Essay in Descriptive Metaphysics, London: Methuen. Strotz, R. (1955–6) “Myopia and inconsistency in dynamic utility maximisation,” Review of Economic Studies 23: 165–80. Sugden, R. (1986) The Economics of Rights, Co-operation and Welfare, Oxford: Blackwell.
210
References
—— (1989) “Spontaneous order,” Journal of Economic Perspectives 3: 85–97. —— (1991) “Rational choice: a survey of contributions from economics and philosophy,” Economic Journal 101: 751–85. —— (1998) “Normative expectations,” in A. Ben-Ner and L. Putterman (eds) Economics, Values, and Organization, Cambridge: Cambridge University Press, 73–100. —— (1999) “Review of Trust within Reason,” Economics and Philosophy 15(1): 127–36. —— (2000) “Team preferences,” Economics and Philosophy 16(2): 175–204. Swinburne, R. (1987) “The structure of the soul,” in A. Peacocke and G. Gillett (eds) Persons and Personality, Oxford: Blackwell. Taylor, C. (1989) Sources of the Self: The Making of the Modern Identity, Cambridge, MA: Harvard. —— (1992) “The politics of recognition,” in A. Gutmann (ed.) Multiculturalism and “The Politics of Recognition,” Princeton: Princeton University Press. Thaler, R. and Shefrin, M. (1981) “An economic theory of self-control,” Journal of Political Economy 89: 392–406. Thévenot, L. (1989) “Equilibre et rationalité dans un univers complexe,” Revue Économique 40(2): 147–97. Tool, M. (1979) The Discretionary Economy, Boulder, CO: Westview Press. Tuomela, R. (1991) “We will do it: an analysis of group intentions,” Philosophy and Phenomenological Research 51: 249–77. —— (1995) The Importance of Us: A Philosophical Study of Basic Social Notions, Stanford: Stanford University Press. Turing, A. (1950) “Computing machinery and intelligence,” Mind 59: 433–60. van Fraassen, B. (1980) The Scientific Image, Oxford: Clarendon. van Gelder, T. (1995) “What might cognition be, if not computation?” Journal of Philosophy 92: 345–81. Varian, H. (1996) “What use is economic theory,” in S. Medema and W. Samuels (eds) Foundations of Research in Economics: How Do Economists Do Economics? Cheltenham: Edward Elgar, 238–47. Veblen, T. (1899) The Theory of the Leisure Class: An Economic Study in the Evolution of Institutions, New York: Macmillan. von Neumann, J. and Morgenstern, O. (1944) Theory of Games and Economic Behavior, Princeton: Princeton University Press. Vromen, J. (1995) Economic Evolution: On Inquiry into the Foundations of New Institutional Economics, London: Routledge. Weintraub, E. (1979) Microfoundations: The Compatibility of Microeconomics and Macroeconomics, Cambridge: Cambridge University Press. —— (1985) General Equilibrium Analysis: Studies in Appraisal, Cambridge: Cambridge University Press. —— (1990) “Methodology doesn’t matter, but history of thought might,” in S. Honkapohja (ed.) The State of Macroeconomics, Oxford: Blackwell, 263–79. Wicksell, K. (1934–5 [1918]) Lectures on Political Economy, trans. E. Classen, London: G. Routledge & Sons. Wilber, C. (1999) “Value judgments and world views,” in P. O’Hara (ed.) Encyclopedia of Political Economy, vol. 2, London: Routledge, 1219–22. Williamson, O. (1985) The Economic Institutions of Capitalism: Firms, Markets and Relational Contracting, New York: Free Press. Wittgenstein, L. (1958 [1953]) Philosophical Investigations, 2nd edn rev., Oxford: Blackwell. Wong, S. (1978) The Foundations of Paul Samuelson’s Revealed Preference Theory: A Study in the Method of Rational Reconstruction, London: Routledge and Kegan Paul. Young, P. (1998) Individual Strategy and Social Structure, Princeton: Princeton University Press.
Index
ability 58–9, 144 agency 57–8, 153; class system and 109; freedom 174, 179–80; principal–agent views of the unity of the individual 70–1; structure-agency models of individual and society 111–14, 136–9 aggregation rule 69, 70 Akerlof, G. 48 Albert, A. 119 alienation 8, 110 Allais, M. 48 Allen, R. 28, 29 Althusser, l 110 altruism 31–2 Amariglio, J. 9 Anscombe, G. 132 Archer, M. 112 Aristotle 72 Armstrong, D. 87 Arrow, K. 34, 37, 68, 75, 90, 91–3, 98 artificial economies 98–100 artificial intelligence 87–8, 98 “as if” intentionality 94–6 asocial individuals 120 atomistic individual 23–44, 151, 156, 164–6, 188, 189–90; autonomy and 33–5; evolution of conception of 27–32; game theory and 38–42; individuation and 147–9; Locke’s legacy 24–7; methodological individualism and 35–6; reductionism and 36–8; requirements of atomism 32–5; welfarism and utilitarianism and 171–4; what kind of individual 42–4 Aumann, R. 39 Austrian economics 16 autonomy 33–5 axiomatic preference 30–1 Ayer, A. 29 Backhouse, R. 29, 89 Baudrillard, J. 9
Becker, G. 30, 32, 36, 45, 50, 53, 55, 56, 60–1, 77, 79, 80, 150 behaviorism 85 Benhabib, S. 123, 181 Bentham, Jeremy 27, 64, 66–7, 69, 70 Bhaskar, R. 112 Blaug, M. 35, 168 Block, N. 85 Boumans, M. 15, 98, 99 bounded rationality 82, 97 Bratman, M. 134 Bridgman, P. 29, 93 Bunge, M. 119 Butler, J. 45, 50, 51, 53–4 capabilities 151–66; embedded individual conception as an ideal conception 160–3; reidentification of embedded individuals and 155–60; Sen and 152–5, 157–9, 163–4 cardinal utility theory 27–8 Cartesian dualism 2–6 Cartwright, D. 133 Cartwright, N. 15, 184, 185 Cassell, G. 29 causal-historical theory of reference 15, 16 central state materialism 84, 87 character formation rule 69, 70 Chicago chool 29–30, 46, 49, 54, 163 choice theory 1–2, 25–6, 48–9, 97; altruism and 31–2; cardinal utility theory 27–8; machinery of choice 81–103; rationality in 26, 27; social choice 75–6, 91–2, see also preferences Chomsky, N. 83 Christianity 3, 4–5 Churchland, P.M. 85 class system 109–10, 169 classical economics 182–3 cleansed preference 78
212
Index
Coase, R. 48 Coates, J. 90 cognitive science 82, 83, 101, 107; economics as 88–100; identity theories of the mind 84–5 Collard, D. 32, 40 collective intentionality 130–49, 177–8; fact–value distinction and 170–1; individuation in the atomistic and socially embedded individual conception 147–9; individuation of embedded individuals 143–7; intention 132–3; revised view of individual economic behavior 139–43; social groups 133–4; structure–agent framework and 136–9; Tuomela and 134–6 collectivism 108 competition 25 computers 82, 83, 97, 101; mind as 84–7; theory of artificial intelligence 87–8, 98 conflict: identity and 8 consciousness 50, 51 consequentialism 171, 172–3; deonticvalue inclusive consequential reasoning 173–4 constraint: structures of 156–7 consumer sovereignty 173 consumerism 8 continence principle 74 conventions 41, 124, 141 Cooley, C. 119 cooperation 40 Cottrell, A. 96 Cowen, T. 78 Cowles Commission 31 Craib, I. 112 critical realism 121–3 Cullenberg, S. 9 Dannefer, D. 116 Danner, P. 120 Davidson, D. 74, 79, 80, 132 Davis, J. 16, 29, 157, 186, 191 Debreu, G. 37, 90 demand: household 95–6; time allocation and 55–6; utility theory and 29, 30 Demos, D. 115, 116 Dennett, D. 75, 76, 80, 88, 96 deontic-value inclusive consequential reasoning 173–4 deontological rationality 139 depth models 9 Descartes, René 2, 3, 4, 5, 9, 10, 11, 84, 165, 166 description theory 15
discrimination 181 disempowerment 10 dualism 2–6, 24, 107 Duesenberry, J. 48 Dupuy, J.-P. 10, 124 Durkheim, Emile 110–11, 113 economics: as cognitive science 88–100; complexity of economic behavior 192; heterodox 16–18, 191; individual in 1–2, 11–17, 187–8; methodology 15, 93; ontology in 185–7; orthodox 16–18, 190–1; socially embedded individual in 117–29; value in 167–83 Edgeworth, F. Y. 28, 32 education 58, 59–60 egonomics 67, 77 Einstein, Albert 36 Elder, G. 117 eliminative materialism 85 Elster, J. 36, 72–3, 77–8, 79, 80 embedded individual 107–29, 151, 156, 164–6, 188–9, 192–3; as an ideal conception 160–3; individuation 130–49; legacy of Marx and Durkheim 109–11; normative reasoning and 174–80; reidentification 155–60; self-referent behavior and individual self-concept 114–17; socially embedded individual concept 117–29, 156–7, 159, 174–80; structure–agency models of individual and society 111–14, 136–9 emotivist thinking 29 empathy 126, 127 endogenously changing preferences 77–9 endowments 48 equilibrium: equilibrium price 88; general equilibrium models 33, 37, 90, 98–9, 100 ethnicity 8 Etzioni, A. 134 Evensky, J. 182 evolutionary game theory 100–1, 141 evolutionary psychology 119 evolutionary stable strategies (ESS) 41 existentialism 8, 169 exploitation 10 extentionalist language 91, 93 externalist definitions of the individual 169–70 facts and values 168–71 fasting 155 feminism 10, 123–4, 163 Field, H. 88 firms 70, 71
Index Fisher, M. 41 Fodor, J. 85 Folbre, N. 123–4, 150, 157 folk psychology 85 Fontaine, P. 31, 32, 65, 125, 126–7 formalism 88, 89 Foster, J. 154 Foucault, Michel 9 Frankfurt, H. 75, 76, 80 Fraser, N. 123, 181 Frazier, C. 150, 184, 188 freedom 152–5, 163; agency 174, 179–80 Frege–Russell description theory 15 French Intersubjectivist school 10, 124–5 Friedman, Milton 30, 35, 94–6, 98, 161 Fudenberg, D. 41 Fullbrook, E. 10, 124–5 functionalism 88, 92–3, 96, 101; views of mental states 82, 86–7 Galbraith, J. 47 game theory 82; atomistic individual and 38–42; evolutionary 100–1, 141 Gecas, V. 115 general equilibrium models 33, 37, 90, 98–9, 100 George, D. 75, 120, 121 Gibbard, A. 15 Giddens, A. 112, 113 Giere, R. 15 Gilbert, M. 133, 134, 140, 147 Gillett, G. 63, 81 Goethe, Johann 63 Goffman, I. 116 Grandmont, J. 37–8 Granovetter, M. 113, 156, 159, 180 Grice, H. 86 Grossman, G. 56 groups: social 133–4 Guala, F. 31 Hands, D. 29, 30, 93, 94, 170 Haraway, D. 11, 81, 89 Harding, S. 8 Hargreaves Heap, S. 39, 41, 125–6 Harman, G. 88 Harsanyi, J. 39, 40, 65 Hausman, D. 66 Hayek, F. 34, 35 Hempel, C. 74, 79 Hesse, M. 15 heterodox economics 16–18, 191 Hicks, J. 28, 29 hierarchical organizations of preferences 74–7 Hirschleifer, J. 32 historical materialism 109
213
Hobbes, Thomas 27 Hodgson, G. 47, 89, 112, 118, 119 holism: markets and 34; methodological holism 17, 110–11, 118 Hollis, M. 40–2, 64 human capital theory 55, 57, 58, 60 Human Development Index (HDI) 154 Hume, David 50, 51–2, 53–4, 55, 57, 63, 64, 67, 141, 174–5 “I” language 132, 134, 142, 177 ideal conceptions 160–3 identity 7, 12–13, 186; identity criteria in economics 13–14; identity theories ofthe mind 84–5; as normative concern 180–2; personal 187–8; reidentification 12–13, 45–62, 151–66; self-referent behavior and 114–17; social identification 7, 8 impossibility theorem 68, 69 indifference curve 28 individualism: methodological 17, 35–6 individuation 13, 61–2; atomistic individual and 147–9; collective intentionality 130–49; of embedded individuals 143–9; multiple selves 64–80 information 48, 68 institutional economics 118–19; new institutional economics 101, 118 instrumental rationality 27, 139, 140, 141 instrumentalism 95 intensional expressions 91 intentionality: “as if” 94–6; collective 130–49, 170–1, 177–8 internal preference structures 69–77 internalist definitions of the individual 169 International Society Theory Consortium 112 intersubjectivist economics 10, 124–5 intuitionism 29 invisible hand metaphor 24–5 irrationality: weakness of will and 72 Jackendoff, R. 101 Jameson, F. 9 Janssen, M. 37 Jevons, W. S. 28 Johnson-Laird, P. 87 Kant, Immanuel 176 Kaplan, H. 114 Kavka, G. 67–8, 69, 75, 76 Keynes, John Maynard 10, 124 Keynesian economics 37, 98 Kincaid, H. 17, 36, 37 Kirman, A. 37, 99 Kitcher, P. 17
214
Index
Knight, F. 30 Krause, U. 66, 67, 68–9, 75 Krugman, P. 88 Lagueux, M. 189 Lancaster, K. 55 language 113, 185; we/I 132, 134, 142, 145, 177, 192 Lasch, C. 8 Lawson, T. 48, 95, 121–3, 138 learning 119 Lester, R. 94, 95 Levy, T. 125 Lewis, D. 39, 86, 87, 175, 177 Locke, John 3, 4–5, 7, 9, 10, 11, 24–7, 30, 32, 33, 42, 45, 49–50, 51, 57, 107, 166, 182 Lucas, R. 35, 98–100 Lutz, M. 120–1 Maas, H. 28 McCloskey, D. 15 machinery of choice 81–103; after the fall 100–3; economics as cognitive science 88–100; mind as computer 84–8 Mäki, U. 15, 60, 95, 170 Malinvaud, E. 78 Mantel, R. 37 marginalism 26, 27–8, 94 Margolis, H. 66 markets: general equilibrium models 33; holism and 34; invisible hand metaphor 24–5 Marshall, A. 29 Marx, Karl 8, 109–10, 163 materialism: central state materialism 84, 87; eliminative materialism 85; historical materialism 109 mathematical economics 89 Mayhew, A. 118 Maynard Smith, J. 41 Mead, G. 119 Mehta, J. 10 memories 51, 52 metaontological analysis 185–6 metapreferences 74–7 methodological holism 17, 110–11, 118 methodological individualism 17; atomistic individual and 35–6 methodology of economics 15, 93 mind 82, 84–8; “as if” intentionality 94–6; as computational processes implemented in different hardware 90–3; functionalist views of mental states 82, 86–7; identity theories 84–5; rationality as symbol processing 96–8; theory of artificial
intelligence 87–8, 98; as unconscious processes 93–4, 101–2 Mirowski, P. 11, 29, 30, 31, 38, 81, 91 modernism: contemporary critiques 6–11; individual in 2–11 Moore, G. E. 29, 175, 176 Moore, Henry 29 morality and ethics: altruism and 31–2; emotivist thinking in 29; moral obligations in organizations and institutions 177–8; moral sentiments in organizations and institutions 174–7; multiple utility functions and 65–6 Morgan, M. 15, 81 Morgenstern, O. 31, 38 Morrison, M. 15 multiple selves 64–80; endogenously changing preferences 77–9; internal preference structures 69–77; single versus multiple utility functions 65–9; unity or plurality of selves 79–80 music 56, 58–60 myopia 72 Nagel, J. 8 Nagel, T. 10, 23, 24 narcissistic personality 8 naturalistic fallacy 29, 175, 176 nature: mechanistic view of 4; prescientific view of 3 Nelson, J. 10 neoclassical economics 45–6, 82, 183; Butler–Hume critique applied to neoclassical view 53–5; choice theory 25–6 New Classical Economics 37, 38 new institutional economics 101, 118 Newton, Isaac 2–6, 24, 25, 36 Nietzsche, F. 8–9 Noonan, H. 12, 51 norms 135–6, 141 O’Boyle, E. 120 ontology in economics 185–7 oppression 10 ordinal utility theory 28 organizations: moral obligations in 177–8; moral sentiments in 174–7 Orléan, A. 124 orthodox economics 16–18, 190–1 Packard, V. 47 Pareto, V. 28 Parfit, D. 16, 188 Patinkin, D. 81 patriarchy 180–1 Peacocke, A. 63, 81
Index perceptions 52 personal identity 187–8 Pickering, A. 89 Pigou, A. 29 Place, U. 84 Plato 3, 29, 72 pluralism 81 Polanyi, K. 117 Pollak, R. 73 Popper, Karl 111 postmodernism: critique of modernist conception of the individual 8–11 poverty 154 precommitment strategies 72, 73 predictions 95 preferences: axiomatic 30–1; endogenously changing preferences 77–9; functionalist view of 92–3; hierarchical organizations of 74–7; intensity 92; internal preference structures 69–77; metapreferences 74–7; multiple utility functions and 66; pure preferences view of the individual 46, 47–55; revealed 30–1, 72, 93–4; team preferences 141; transitive 54 Pressman, S. 152 price: equilibrium 88 principal–agent views of the unity of the individual 70–1 psychology 8, 25, 29; evolutionary 119; folk psychology 85; social 115–17, 144, 146, see also mind psychotherapy 8 pure preferences view of the individual 46, 47–9; Butler–Hume critique applied to neoclassical view 53–5; individuation and 64; Locke and 49–50; Locke’s critics 50–2 Putnam, 87 Pylyshyn, Z. 87 Quine, W. 1, 13, 85, 90–1, 92, 98, 185 Rabin, M. 31 Ramsey sentences 86–7 Ramstad, Y. 119 realism: critical 121–3 reason and rationality 4, 54, 64; bounded 82, 97; choice and 26, 27, 40; deontological rationality 139; imperfect 72; instrumental rationality 27, 139, 140, 141; reflected rationality theory 125–7; situated 138; as symbol processing 96–8 recognition: politics of 123 Reder, M. 49 reductionism: atomistic individual and 36–8; game theory and 39–40; identity theories of the mind and 84
215
reference: theory of 15–16 reflected rationality theory 125–7 reflexivity 114, 191–2 regression analysis 29 reidentification 12–13, 45–62, 151–66, 187–8; Butler–Hume critique applied to neoclassical view 53–5; capabilities 151–66; embedded individual 155–60; individuation and 61–2; Locke and 49–50; Locke’s critics 50–2; pure preferences view of the individual 46, 47–55; time allocation conception of the individual 46, 55–61, 156 religion 3, 4–5 representation 15 Rethinking Marxism 9 Reuten, G. 109 revealed preference 30–1, 72, 93–4 Riesman, D. 7 Rizvi, S. 37, 38, 41 Robbins, L. 28, 29, 182–3 Rorty, R. 3 Rosenberg, M. 115, 146 rules 135–6, 141; following 82, 97–8, 100–1 Runde, J. 60 Rutherford, M. 81, 118, 162 Samuelson, P. 30, 93–4, 98 Sartre, Jean-Paul 169 satisficing 97 Scanlon, T. 171–3 scarcity 48 Schelling, 40, 67, 68, 70, 72, 73, 77 Schick, F. 74 Schoeneman, T. 116 Schultz, H. 29 Schumpeter, Joseph 35, 36 science: history of 2 Searle, J. 101, 102, 134 second-best solutions: theory of 72 Segal, G. 88 self-control 72–4 self-esteem 115, 116 self-imposing tasks 145–7 self-interest 25, 40, 65–6 self-referent behavior: individual self-concept and 114–17; intention and 144–5 selfishness 31–2 Sellars, W. 132 Selten, R. 40 semantics 91 Sen, A. 48, 65, 66, 74–5, 150, 152–5, 157–9, 163–4, 167, 171, 172, 173, 179–80, 182 Sent, E.-M. 31, 97 Shaftesbury, A. 130 Shakespeare, William 23
216
Index
Shefrin, M. 70, 79 Shrauger, J. 116 Shubik, M. 38 Shwayder, D. 13, 14, 185 signifiers 9 Simon, H. 23, 48, 96–8, 99, 100 situated rationality 138 Slutsky, E. 28, 29 Smart, J. J. C. 84 Smith, Adam 24–6, 71, 107, 125, 182, 183 social choice 75–6, 91–2 social economics 119–21 social groups 133–4 social identification 7, 8 social psychology 115–17, 144, 146 social science: critique of modernist conception of the individual 7–8 social theory: structure–agency models of individual and society 111–14, 136–9 social welfare approach to preferences 78 socially embedded individual concept 117–29, 156–7, 159; normative reasoning and 174–80 society 34; structure–agency models of individual and society 111–14, 136–9 sociological individualism 36 Sonnenschein, H. 37, 38 sour grapes 77 specialization 71 standpoint epistemology 8 starving 155 Steedman, I. 66, 67, 68–9, 75 stereotypes 10 Stevenson, C. 29 Stigler, G. 28, 30, 50, 53, 56, 60 Strawson, P. 16 Strotz, R. 73 structuration theory 112 structure–agency models of individual and society 111–14, 136–9 Sugden, R. 41, 42, 64, 140, 141, 160, 175, 176, 177 sum-ranking 171–2 Summerfield, G. 152 Swinburne, R. 10 symbol processing: rationality as 96–8 sympathy 126, 127, 175 syntax 102 Tarski, Alfred 91
Taylor, C. 1, 3, 6, 32, 123, 147, 165, 181 team preferences 141 Thaler, R. 70, 79 therapy 8 Thévenot, L. 124 time allocation conception of the individual 46, 55–61, 150, 156; problems 58–61 time-inconsistency 72, 73 Tirole, J. 41 Tool, M. 117 transaction costs 48 transitive preferences 54 Tuomela, R. 134–6, 138, 140, 141 Turing, Alan 83, 87, 100 Turing machines 87, 100 unconscious processes 93–4, 101–2 utilitarianism 29, 171–4 utility theory 26, 27–30, 45, 53; single versus multiple utility functions 65–9 value in economics 167–83; classical world 182–3; facts and values 168–71; individual identity as normative concern 180–2; normative reasoning and socially embedded individual conception 174–80; welfarism, utilitarianism and atomistic individual conception 171–4 van Fraassen, B. 15 van Gelder, T. 82 Varian, H. 15, 88 Varoufakis, Y. 39, 41 Veblen, Thorstein 47 von Neumann, J. 31, 38 Vromen, J. 32, 95 Walras, Leon 88 “we” language 132, 134, 142, 145, 177, 192 weakness of will 72–4 Weintraub, E. 37, 91, 186 welfare economics 152–3 welfarism 171–4 Wicksell, K. 29 Wilber, C. 168 Williamson, O. 71 Wittgenstein, Ludwig 6, 113, 148 Wong, S. 30 Zander, A. 133