Difficult choices: autonomy and the liberal state

This project concerns a deeply contested moral ideal: autonomy. To be autonomous is to have authority over one's self and to govern one's life on the basis of value commitments one deems important. One of the dominant views of liberalism - antiperfectionist comprehensive liberalism - distinguishes itself from other liberal views because it grants unique privilege to the ideal of autonomy in personal and political life. Will Kymlicka is one of the most prominent defenders of antiperfectionist comprehensive liberalism and while he appeals to the ideal of autonomy in his liberalism, he is not clear about which theory of autonomy he is appealing to. As a result, his theory of autonomy is sketchy and incomplete. As an autonomy theorist, I think that there is more Kymlicka can say to elaborate on the view of autonomy operating in his theory of liberalism. Kymlicka has not explained whether his view is a procedural, or substantive, or a socio-relational view and these exhaust the kinds of views of autonomy in the literature, so Kymlicka's view must be one of these. It is important for Kymlicka to be clear on the view of autonomy he incorporates because each theory has its own motivating assumptions and standards for what counts as an autonomous choice. In addition, it is important for Kymlicka to be clear about the theory of autonomy in his liberalism, because, in some cases, the standards for autonomy may be inconsistent with his liberal commitments. In this project, I argue that Kymlicka incorporates a socio-relational view of autonomy in his liberalism.

DIFFICULT CHOICES: AUTONOMY AND THE LIBERAL STATE by Jennifer Anne Warriner A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Philosophy The University of Utah May 2014 Copyright © Jennifer Anne Warriner 2014 All Rights Reserved Th e Uni v e r s i t y o f Ut a h Gr a dua t e S cho o l STATEMENT OF DISSERTATION APPROVAL The dissertation of Jennifer Anne Warriner has been approved by the following supervisory committee members: Cynthia A. Stark , Chair 12/11/2012 Date Approved Leslie Francis , Member 12/11/2012 Date Approved Chrisoula Andreou , Member 12/11/2012 Date Approved Ella Myers , Member 12/11/2012 Date Approved Marina Oshana , Member 12/11/2012 Date Approved and by Stephen M. Downes , Chair of the Department of Philosophy and by David B. Kieda, Dean of The Graduate School. ABSTRACT This project concerns a deeply contested moral ideal: autonomy. To be autonomous is to have authority over one's self and to govern one's life on the basis of value commitments one deems important. One of the dominant views of liberalism - antiperfectionist comprehensive liberalism - distinguishes itself from other liberal views because it grants unique privilege to the ideal of autonomy in personal and political life. Will Kymlicka is one of the most prominent defenders of antiperfectionist comprehensive liberalism and while he appeals to the ideal of autonomy in his liberalism, he is not clear about which theory of autonomy he is appealing to. As a result, his theory of autonomy is sketchy and incomplete. As an autonomy theorist, I think that there is more Kymlicka can say to elaborate on the view of autonomy operating in his theory of liberalism. Kymlicka has not explained whether his view is a procedural, or substantive, or a socio-relational view and these exhaust the kinds of views of autonomy in the literature, so Kymlicka's view must be one of these. It is important for Kymlicka to be clear on the view of autonomy he incorporates because each theory has its own motivating assumptions and standards for what counts as an autonomous choice. In addition, it is important for Kymlicka to be clear about the theory of autonomy in his liberalism, because, in some cases, the standards for autonomy may be inconsistent with his liberal commitments. In this project, I argue that Kymlicka incorporates a socio-relational view of autonomy in his liberalism. This is dedicated to my grandmothers, both of whom taught me about the importance of autonomy. TABLE OF CONTENTS ABSTRACT ....................................................................................................................... iii ACKNOWLEDGEMENTS .............................................................................................. vii Chapters 1: LIBERALISM AND AUTONOMY ...............................................................................1 Introduction .....................................................................................................................1 Antiperfectionist Comprehensive Liberalism .................................................................4 Chapter Outlines ............................................................................................................11 2: PROCEDURALIST THEORIES OF AUTONOMY ....................................................18 Introduction ...................................................................................................................18 Autonomy and the Social World ...................................................................................19 Christman on Proceduralism .........................................................................................24 Kymlicka and Proceduralism ........................................................................................32 Concluding Remarks .....................................................................................................38 3: STRONG AND WEAK SUBSTANTIVE THEORIES OF AUTONOMY ..................39 Introduction ...................................................................................................................39 Strong Substantivism and Proceduralism on Oppressive Values ..................................42 Weak Substantive Autonomy ........................................................................................50 Weak Substantive Autonomy: A New Way? ................................................................55 Conclusion .....................................................................................................................66 4: KYMLICKA AND SOCIO-RELATIONAL AUTONOMY ........................................67 Introduction ...................................................................................................................67 Socio-Relational Autonomy ..........................................................................................68 Oshana and Proceduralism ............................................................................................72 Oshana and Strong Substantivism .................................................................................76 The Conditions for Socio-Relational Autonomy ...........................................................79 Oshana and Content-Neutrality .....................................................................................83 Kymlicka and the Ideal of Autonomy ...........................................................................89 Conclusion ...................................................................................................................103 vi 5: SOCIO-RELATIONAL AUTONOMY AND NEUTRALITY ..................................104 Introduction .................................................................................................................104 The Concept of Neutrality and Liberal Theories .........................................................107 Kymlicka and Neutrality: The Response to Quong .....................................................119 Conclusion ...................................................................................................................126 6: CONCLUDING REMARKS .......................................................................................127 BIBLIOGRAPHY ............................................................................................................135 ACKNOWLEDGEMENTS Writing a dissertation is a lengthy endeavour and I wish to thank those individuals who helped me to complete this project. First, Prof. Cynthia Stark deserves special mention for not only agreeing to supervise my project, but for providing invaluable and constructive criticism throughout its evolution. This final product would not be the same without her contribution. I wish also to thank Prof. Leslie Francis, Prof. Chrisoula Andreou, Prof. Ella Myers, and Prof. Marina Oshana, all of whom served on my committee and provided valuable feedback for improving the overall project. Any errors in this project are mine alone. I wish to thank the University of Utah's Philosophy Department for its contribution to my development as a philosopher. In particular, I wish to thank Prof. Stephen Downes, Prof. Lex Newman, and Prof. Jim Tabery for their support. In addition, I wish to thank my fellow graduate students, Monika Piotrowska, Matt Mosdell, Meg Bowman, all of whom provided valuable support during my PhD, both in and out of the department. I offer special thanks to Mike Wilson, for reading drafts of chapters and for being available for numerous phone calls. I offer special thanks also to Myrto Mylopoulous, for knowing exactly what to say. My most profound debt is to Glen Oikawa. Glen not only suffered through my anxieties and absences, his love and support made this dissertation possible. CHAPTER 1 LIBERALISM AND AUTONOMY Introduction This project concerns a deeply contested moral ideal: autonomy. To be autonomous is to have authority over one's self and to govern one's life on the basis of value commitments one deems important.1 Autonomy is a concept that plays a prominent role in several subfields of philosophy, including political philosophy, where there is a debate over the role of autonomy in liberalism. It's no secret that all liberal theorists regard autonomy as an important moral ideal. Indeed, we can say that the "basic organizing idea" of liberalism is the following, namely "the fundamental value of an individual's rationally and autonomously pursuing or embracing those things she judges to be worthwhile."2 In other words, liberal theorists think that it is valuable for individuals to be autonomous, to shape their lives freely and as they see fit. Liberal theorists, however, disagree over what this commitment to autonomy means for liberal theorizing.3 This disagreement emerges over two key questions. First, liberal thinkers 1 When I refer to "value commitments," I intend this is to include religious or philosophical beliefs, moral ideals, moral principles, and character traits. 2 John Christman, Social and Political Philosophy: A Contemporary Introduction, New York: Routledge, 2002, 96. 3 I follow Jonathan Quong's formulation of these two debates. See Liberalism Without Perfection, (Oxford: Oxford University Press, 2011), especially chapter 1, 15-16. 2 disagree over whether a necessary feature of liberalism is a claim about the good life. In particular, is liberal theory itself based upon a particular ideal such as the autonomous life? Call this the character question.4 Second, liberal thinkers disagree over the role of the liberal state in promoting the good life to its citizens. More precisely, should the liberal state promote the autonomous life? Call this the practice question.5 While there are four main liberal theories, each of which provides different answers to these two questions, my project will focus upon one theory in particular: antiperfectionist comprehensive liberalism (henceforth APCL). APCL distinguishes itself from other liberal views because it answers the character and practice questions by appeal to the ideal of personal autonomy. Defenders of APCL claim that 1) liberal theory is committed to an ideal of personal autonomy as an intrinsically valuable way of life and that 2) the state should remain neutral and refrain from the active promotion of valuable ways of living over other reasonable alternatives, in the interest of protecting personal autonomy. One of the most prominent defenses of APCL in the literature is offered by Will Kymlicka and I focus exclusively on his view in this project.6 Here's why. While Kymlicka appeals to the ideal of autonomy in his liberalism, he is not clear about which theory of autonomy he is appealing to. As a result, his theory of autonomy is sketchy and 4 Quong identifies this as the first question "about the fundamental character of liberal philosophy," which he poses in the following way: "Must liberal philosophy be based in some particular ideal of what constitutes a valuable or worthwhile life, or other metaphysical beliefs?" (Liberalism Without Perfection, 15). 5 Quong identifies this as the second question "about the practice of liberal states," (16), which he poses as follows: "Is it permissible for a liberal state to promote or discourage some activities, ideals, or ways of life on grounds relating to their inherent or intrinsic value, or on the basis of some other metaphysical claims?" (Liberalism Without Perfection, 15). 6 Kymlicka's first articulation of his view is in Liberalism, Community, and Culture, Oxford: Oxford University Press, 1989. See also Multicultural Citizenship, New York: Oxford University Press, 1995. 3 incomplete. As an autonomy theorist, I argue that Kymlicka must be clear about the theory of autonomy he has in mind when he incorporates the ideal in his liberalism. After all, there are several different theories of autonomy, each with its own motivating assumptions and standards for what counts as an autonomous choice. In addition, it is important to be clear on the theory of autonomy one incorporates in his view of liberalism, because, in some cases, the standards for autonomy may be inconsistent with one's liberal commitments. So, the primary aim of my project is to explain what theory of autonomy is operating implicitly in Kymlicka's liberalism. I explain each of these theories of autonomy briefly in this chapter, but I can say now that there are four main accounts in the literature: proceduralist, strong substantivist, weak substantivist, and socio-relational. These exhaust the kinds of views of autonomy in the literature, so Kymlicka's view must be one of these. The central claim of my project is this: I argue that Kymlicka's view of autonomy in his liberalism is socio-relational. I think if we are clear on the standards for autonomy associated with Kymlicka's liberalism, then this will serve two important ends. First, my project is bringing to the surface something important for Kymlicka's view: the account of autonomy he incorporates in his theory of liberalism is socio-relational. Two, identifying the standards for autonomy associated with Kymlicka's account of liberalism will provide a novel way for him to respond to his critics. I devote this chapter to elaborating on the fundamental aspects of Kymlicka's liberalism, after which I provide summaries of the ensuing chapters. 4 Antiperfectionist Comprehensive Liberalism I begin this section with a disclaimer: I am proceeding on the admittedly controversial assumption that liberal theorists ought to endorse Kymlicka's view of liberalism. I acknowledge that critics of Kymlicka's liberalism have raised serious concerns about the viability of the position.7 I leave a full-scale defense of Kymlicka's particular approach to liberal theory for another time, but let me say this briefly. I think part of the reason why theorists are critical of Kymlicka's account of liberalism stems from misunderstandings about the view itself. In particular, I think that theorists misunderstand the role that autonomy plays in Kymlicka's liberalism and we can trace this confusion in part to the incomplete account of autonomy Kymlicka incorporates within his view. Accordingly, this provides some of the motivation behind my project. A second disclaimer: while I discuss only Kymlicka's defense of APCL in this project, it would be a mistake to think that APCL is "one single theory."8 While Kymlicka offers one prominent defense of the view, there are others, including John Locke, Theory of Justice-era John Rawls, and Ronald Dworkin.9 Nevertheless, however one defends the theory, any account of APCL grant unique privilege to the ideal of autonomy in personal and political life. We see evidence of this in how APCL theorists 7 Quong in particular has raised one of the most recent critiques of APCL in general. See Liberalism Without Perfection, 22-26. 8 Quong, Liberalism Without Perfection, 24. 9 John Locke, Two Treatises of Government, Ian Shapiro, ed., New Haven, CT: Yale University Press, 2003; John Rawls, A Theory of Justice, Cambridge, MA: Harvard University Press, 1971; Ronald Dworkin, Sovereign Virtue: The Theory and Practice of Equality, Cambridge, MA: Harvard University Press, 2002. Following Quong, I include Dworkin as an APCL theorist, even though his account of liberalism is couched in a defense of equality, not autonomy per se, because his "‘challenge model' of the good life...plays a similar role to autonomy," (Quong, Liberalism Without Perfection, 19, fn. 14.) 5 respond to the character and practice questions. Let's consider how Kymlicka responds to these key questions. First, the character question: Is liberalism based on a view about what constitutes a good or flourishing life? As I noted, APCL theorists usually refer to personal autonomy when answering this question. However, Kymlicka offers a unique answer to the character question. Even though Kymlicka is an APCL theorist, the ideal of personal autonomy plays only an indirect role in his formulation of liberalism.10 To be sure, Kymlicka thinks that autonomy is a valuable ideal, but he doesn't regard autonomy as valuable for its own sake.11 Rather, Kymlicka claims that autonomy is instrumentally valuable because it contributes to our essential interest in leading an objectively good life. According to Kymlicka, it is this essential interest - and not autonomy as such - that "forms the basis of liberal political theory."12 What is an objectively good life on Kymlicka's view? Kymlicka doesn't offer a particular account of the good because he thinks there is a wide range of objectively good lives that individuals can lead. But, Kymlicka does claim that for any objectively good life, it must be led from the inside.13 According to Kymlicka, "no life goes better by being led from the outside according to values the person doesn't endorse. My life only 10 Kymlicka, Liberalism, Community, and Culture, 10-13. 11 As Kymlicka notes in Liberalism, Community, and Culture, some theorists (notably Rawls) suggest that our essential interest (or ‘higher order interest', in Rawls's terminology) "is in our capacity to form and revise our rational plans of life," (11). In other words, Rawls is saying that our essential interest lies in being autonomous. But, Kymlicka (following Ronald Dworkin) claims that "this puts the cart before the horse," (12). Here, Kymlicka quotes Dworkin: "Our higher-order interest is not an interest in exercising a capacity because we find that we have it...but rather we develop and train capacities of the sort that [they] describe because we have a certain interest," namely the interest in leading a life that is good (12, emphasis mine). 12 Kymlicka, Liberalism, Community, and Culture, 13. 13 Kymlicka, Liberalism, Community, and Culture, 12. 6 goes better if I'm leading it from the inside, according to my beliefs about values."14 Liberal theorists refer to this as the endorsement constraint and it is important to note that Kymlicka regards it as a fundamental aspect of his liberalism for the following reason. While each of us has an essential interest in leading an objectively good life, it is also possible that we are mistaken about what constitutes this kind of life. It is easy to understand how we might make this error. Identifying the beliefs that we think matter most is a difficult, sometimes agonizing, task. We have limited reasoning faculties and sometimes we reason incorrectly, so the value commitments we hold may not lead us to live an objectively good life. Sometimes, we simply change our minds, so the value commitments we identify today may not be right for us in the future. We must make difficult decisions, yet we know we are fallible. However, even if we are (or can be) in the wrong about what the good life consists in, other people cannot step in and correct our errors, in hopes of making our life go better. On Kymlicka's view, the endorsement constraint prevents us from intervening in this way. A person who isn't leading his life from the inside, i.e. who is coerced by external factors to hold or reject certain beliefs, will not be leading a good life. A person must lead his life from the inside, even if he is leading his life according to objectively bad values. So, when does a person lead his life from the inside? According to Kymlicka, one of the necessary conditions for leading life from the inside is that one is autonomous. In other words, a person must be able to self-govern on the basis of what he thinks are valuable pursuits and moral ideals, even if these pursuits and ideals are "wrong" from some better perspective. 14 Kymlicka, Liberalism, Community, and Culture, 12. 7 How, then, does Kymlicka understand the ideal of autonomy? According to Kymlicka, to be autonomous, one must have certain capacities.15 In particular, one must have the ability to arrive at her values in the proper way and the ability to rationally revise one's good. Furthermore, to be autonomous, individuals require certain rights and liberties, and cultural conditions (e.g. freedom of the press, freedom of association) in order to exercise her capacities for autonomy.16 For example, a liberal state grants political and civil rights to individuals, which rights provide them formal guarantees that they can (say) pursue whatever religious or philosophical beliefs they wish without persecution. Let's put all of these claims together. As an APCL theorist, Kymlicka claims that liberalism is based upon the idea that each of us has an essential interest in leading an objectively good life. For any objectively good life, it is led from the inside, because a person's life only goes well when she leads it from the inside, according to values she deems important from her own perspective. One necessary condition for leading life from the inside is that one is autonomous. Thus, being autonomous is a necessary condition for leading an objectively good life. However, while autonomy plays only an indirect role in Kymlicka's liberalism, it is an important feature of his overall view. As I stated earlier, it is my claim in this project that Kymlicka's account of autonomy is incomplete. As an autonomy theorist, I think that there is more Kymlicka can say to elaborate on the view of autonomy operating in his theory of liberalism. Kymlicka hasn't explained whether his view is procedural, 15 Kymlicka, Liberalism, Community, and Culture, 13. 16 Kymlicka, Liberalism, Community, and Culture, 13. 8 strong substantivist, weak substantivist, or socio-relational and these exhaust the kinds of views of autonomy in the literature, so Kymlicka's view must be one of these. We need to know what kind of view of autonomy Kymlicka is advocating because there are different standards for different views, and the standards for autonomy tell us what conditions must obtain in order that a person counts as autonomous. Moreover, the standards associated with some views of autonomy are inconsistent with certain liberal commitments, so it is important for Kymlicka to be clear about what theory of autonomy he is endorsing in his liberalism. So far, I've explained Kymlicka's answer to the character question and now I turn to examining his answer to the practice question. What role, if any, should the state play in promoting the good life to its citizens? Some liberal theorists say that the state should be perfectionist.17 Perfectionism is the claim that humans have a particular nature and that, given this nature, there are goods any human must have, in order to lead an excellent life. For example, one might think that the capacity for rationality is an essential part of human nature and that it is a perfectionist good for individuals to develop their rational capacities.18 Along the same lines, a perfectionist liberal view holds that the state is justified in identifying and advancing through its policies and laws objectively good human ways of living, regardless of whether individuals recognize or endorse that way of 17 Some of perfectionist comprehensive liberalism's most prominent defender are Joseph Raz , The Morality of Freedom, Oxford: Clarendon Press, 1988; William Galston, Liberal Purposes: Good, Virtues, and Diversity in the Liberal State, Cambridge: Cambridge University Press, 1991; Stephen Wall, Liberalism, Perfectionism, and Restraint, Cambridge: Cambridge University Press, 1998. 18 Steven Wall, "Perfectionism in Moral and Political Philosophy," Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/perfectionism-moral/ , accessed June 6, 2013. 9 living as good, though within the boundaries of justice.19 Defenders of perfectionist liberalism claim that the state plays an active and essential role in evaluating and promoting particular ways of living over others (and penalizing those who favour less worthy ways of life). In contrast, Kymlicka thinks that the liberal state should maintain a neutral position between objectively good ways of living endorsed by individuals. This means that the state is not justified in actively promoting certain ways of living over others to citizens or in justifying its day-to-day policies and laws by appeal to certain conceptions of the good. According to Kymlicka, the role of the state is to provide and regulate a neutral framework in which citizens can pursue their good. Negatively, this means that the state cannot evaluate or make judgments about citizens' conceptions of the good (provided that these conceptions are "justice-respecting").20 If a person's conception of the good is "justice-respecting" - even if one's life is characterized by (say) vulgarity or misanthropy - then it is not the place of the state to weigh in on whether that conception really is worthy of pursuit or whether the person pursing it merits resources.21Positively, the state is responsible for protecting individuals' capacity for 19 John Christman, Social and Political Philosophy: A Contemporary Introduction, 104. In other words, perfectionist liberal states cannot force or coerce individuals into accepting one set of beliefs or values over another. Indeed, given that liberals from Locke onward have argued against the effectiveness of forcing citizens to profess or reject a particular faith, it seems that the PCL state should avoid using coercion. But, some critics wonder whether it is possible for the perfectionist liberal state to refrain from coercing citizens to accept a particular way of life, even if perfectionist liberalism is constrained by liberal ideals. See Alfonso J. Damico, "What's Wrong with Liberal Perfectionism," Polity, Vol. 29, No. 3 (Spring 1997), 397-420, 399. 20 The term "justice-respecting" is Kymlicka's, Contemporary Political Philosophy: An Introduction, Second Edition, p. 217. I say more about this in Chapter 5. 21 Of course, even if some people are leading immoral but "justice-respecting" lives, it is open to others around her to "remonstrate with these people, reason with them, entreat them, persuade them, and if that does not work, avoid their company. But it is not a reason for compelling them or taking political action to visit upon them any evil or disadvantage if they will not mend their ways," Jeremy Waldron, 10 autonomy, so that they can determine for themselves what counts as a worthy conception of the good, without state intrusion, and the state has a positive duty to provide equally to citizens the proper means and resources to pursue whatever their "justice-respecting" conceptions of the good.22 How does Kymlicka defend neutrality? It will be instructive to identify, in order to set aside, one possible argumentative strategy. While some liberal thinkers claim that neutrality is (or ought to be) defended by appeal to sceptical considerations,23 Kymlicka rejects this line of thinking. He does not advocate neutrality because we are unable to know whether certain conceptions of the good are more valuable than others.24 Instead, Kymlicka offers three different lines of argument to defend neutrality. I consider these in depth in Chapter 5, but I'll state them now. First, Kymlicka argues that perfectionist state action violates the endorsement constraint and it is self-defeating to violate it. Second, he argues that perfectionist state action interferes with individuals' autonomy, in particular with the ability to rationally revise one's good. Lastly, Kymlicka argues that perfectionist actions distort the cultural marketplace of ideas. Based on these considerations against state perfectionism, Kymlicka concludes that neutrality is preferable, i.e. a state which doesn't actively promote certain ways of living over others or justify its policies and laws on the basis of certain conceptions of the good. "Autonomy and Perfectionism in Raz's Morality of Freedom," Southern California Law Review, Vol. 62, 1989, 1098-1152, 1133-1134. 22 Kymlicka, Contemporary Political Philosophy: An Introduction, Second Edition, 217-8. 23 See Brian Barry, Justice as Impartiality, New York: Oxford University Press, 1989, 169-182. 24 As Kymlicka points out, "scepticism does not in fact support self-determination. If people cannot make mistakes in their choices, then neither can governments. If all ways of life are equally valuable, then no one can complain when the government chooses a particular way of life for the community. Hence scepticism leaves the issue unresolved," (Contemporary Political Philosophy: An Introduction, Second Edition, 214). 11 Kymlicka faces two kinds of criticisms directed toward his stance on state neutrality. First, political liberals like Jonathan Quong argue that liberal views such as Kymlicka's are inconsistent.25 How can Kymlicka claim that the liberal state ought to be neutral and justify neutrality by appeal to a normative ideal like autonomy? Isn't neutrality clearly violated by this move? The second line of criticism comes from perfectionist comprehensive liberals like Joseph Raz, Thomas Hurka, and Richard Arneson and they argue that Kymlicka's position on neutrality is incoherent. If Kymlicka claims that all individuals have an essential interest in leading a good life, why should the state remain neutral? That is, why shouldn't the state take steps to actively ensure that citizens really are leading good lives? In this project, I think that by getting clear on Kymlicka's theory of autonomy and the role it plays in his liberalism, Kymlicka will be in a better position to respond to his critics. Chapter Outlines As I stated at the outset, the central claim of this project is that Kymlicka's view of autonomy in his liberalism is socio-relational. What is notable about a socio-relational view of autonomy is that it claims a person's autonomy is largely constituted by her social conditions, although certain psychological states are also necessary for autonomy. As we shall see, Kymlicka's view of autonomy must endorse these particular kinds of standards for autonomy as well. Put differently, I will argue that Kymlicka cannot endorse any view of autonomy except for a socio-relational view. In Chapter 2, I argue that Kymlicka's view cannot be proceduralist. Roughly, a proceduralist theory of autonomy claims that a person is autonomous if she subjects (or 25 Quong, Liberalism Without Perfection, 22-26. 12 would subject) her value commitments to the "right" process of critical reflection.26 Put another way, proceduralist views say that if a person's psychological states meet certain standards, then this is both necessary and sufficient for autonomy. While proceduralist theorists disagree over how to understand this process, such theories have been the dominant view in the literature since the 1970s.27 Part of the appeal of proceduralist theories is that they are content-neutral. A content-neutral view of autonomy does not require autonomous individuals to hold particular kinds of value commitments. What 26 Some of the most prominent defenders of proceduralism are John Christman, Gerald Dworkin, Marilyn Friedman, Diana T. Meyers, and Andrea Westlund. For Christman, see "Autonomy and Personal History, Canadian Journal of Philosophy 20 (1990) 1-24; "Liberalism and Individual Positive Freedom," Ethics 101 (1991), 343-359; "Relational Autonomy, Liberal Individualism, and the Social Constitution of Selves," Philosophical Studies 117 (2004): 143-164; "Procedural Autonomy and Liberal Legitimacy," in James Stacey Taylor (ed.) Personal Autonomy: New Essays on Autonomy and Its Role in Contemporary Moral Philosophy, New York: Cambridge University Press, 2005; "Autonomy, Self-Knowledge, and Liberal Legitimacy," in Joel Anderson and John Christman (eds.), Autonomy and the Challenges to Liberalism, Cambridge: Cambridge University Press, 2005; The Politics of Persons. Individual Autonomy and Socio-historical Selves, Cambridge: Cambridge University Press, 2009. For Dworkin, see The Theory and Practice of Autonomy, New York: Cambridge University Press, 1988. For Friedman, see M. "Autonomy and Social Relationships: Rethinking the Feminist Critique," in Diana Tietjens Meyers (ed.) Feminists Rethink the Self Boulder, CO: Westview, 1997; Autonomy, Gender, Politics New York: Oxford University Press, 2003. For Meyers, "Personal Autonomy and the Paradox of Feminine Socialization," Journal of Philosophy 84 (1987): 619:628; Self, Society and Personal Choice, New York: Columbia University Press, 1989; "Feminism and Women's Autonomy: The Challenge of Female Genital Cutting. Metaphilosophy 31 (2000): 469-491; "Intersectional Identity and the Authentic Self. Opposites Attract!" in Catriona Mackenzie and Natalie Stoljar (eds.) Relational Autonomy: Feminist Perspectives on Autonomy, Agency and the Social Self, New York: Oxford University Press, 2000; Gender in the Mirror: Cultural Imagery and Women's Agency. New York: Oxford University Press, 2002; "Decentralizing Autonomy: Five Faces of Selfhood," in Joel Anderson and John Christman (eds.), Autonomy and the Challenges to Liberalism, Cambridge: Cambridge University Press, 2005. For Westlund, see "Rethinking Relational Autonomy," Hypatia 24 (2009): 26-49. 27 See Harry Frankfurt "Freedom of the Will and the Concept of a Person," The Journal of Philosophy, Vol. 68, No. 1, January 1971, 5-20; Gerald Dworkin "Acting Freely," Nous, 4 (November), 1972, 367-383 ; Gary Watson, "Free Agency," Journal of Philosophy, 72 (8), 1975, 205- 20. 13 matters for autonomy is that a person engages in the right sort of critical process. This means that individuals can hold a wide range of value commitments and count as autonomous. Proceduralist views are attractive because they set only minimal standards for autonomy related to a person's critical abilities; as a result, individuals have great latitude with respect to the kinds of values they can hold and qualify as autonomous. However, because proceduralist views claim that psychological standards are necessary and sufficient for autonomy, this means that a person's social conditions don't play a role in establishing whether a person is autonomous. Put another way, defenders of proceduralism are claiming that social conditions are conceptually irrelevant for determining whether a person qualifies as autonomous. However, I will argue that Kymlicka doesn't regard the social conditions as conceptually irrelevant for autonomy. I will argue that Kymlicka see social conditions as (at least partly) constitutive of a person's autonomy. If this line of thinking is correct, then his view of autonomy cannot be proceduralist. If Kymlicka's view of autonomy incorporates social conditions as constitutive for autonomy, one might think that he is offering a strong substantivist account of autonomy.28 While theorists offer different interpretations of these views, we can say roughly that strong substantive accounts hold that critical reflection is necessary but not sufficient for autonomy. Defenders of these views introduce substantive constraints on 28 Among those who defend strong substantive views are Susan Babbitt, "Feminism and Objective Interests: The Role of Transformation Experiences in Rational Deliberation," in Feminist Epistemologies, Linda Alcoff and Elizabeth Potter eds., New York: Routledge, 1993; Natalie Stoljar, "Autonomy and the Feminist Intuition," in Relational Autonomy: Feminist Perspectives on Autonomy, Agency, and the Social Self, Catriona Mackenzie and Natalie Stoljar, eds., New York: Oxford University Press, 2001, 94-111; Susan Wolf, "Asymmetrical Freedom," The Journal of Philosophy 77 (1980), 151-166. Some autonomy theorists argue that Thomas Hill is a strong substantivist theorist, on the basis of claims he makes in Autonomy and Self-Respect Cambridge: Cambridge University Press, 1991. However, in Chapter 3, I question whether this is the case. 14 the content of the value commitments an autonomous agent is permitted to endorse. On this view, a person must engage in critical reflection and her choices and preferences are subject to certain normative constraints. In particular, her choices and preferences must "hook up" in the proper way with an objective feature of the world, e.g. that she is a moral equal. However, I argue in Chapter 3 that because strong substantivist places restrictions on the kinds of choices and preferences that autonomous agents can hold, this is inconsistent with one aspect of Kymlicka's view, namely his commitment to the endorsement constraint. After rejecting strong substantivist standards as inconsistent with Kymlicka's liberalism, I consider next a relatively recent view introduced in the literature, namely weak substantivism.29 On this view, a person is autonomous if she critically reflects upon and endorses her value commitments in the proper way and holds the right kind of psychological attitudes toward her agency. Weak substantive accounts are supposed to be preferable to strong substantive and proceduralist views because they allow agents to hold a wide range of value commitments (even oppressive ones) while incorporating normatively robust standards. In other words, weak substantive theories carve a middle path between proceduralist and strong substantive views, retaining the positive aspects of these views while avoiding the problematic aspects. I argue that weak substantivism fails 29 For weak substantive views, see Paul Benson, "Feminist Intuitions and the Normative Substance of Autonomy," in Personal Autonomy: New Essays on Personal Autonomy and its Role in Contemporary Moral Philosophy, ed. James Stacey Taylor, New York: Cambridge University Press, 2005, 128-142; Paul Benson, "Taking Ownership: Authority and Voice in Autonomous Agency," in Autonomy and the Challenges to Liberalism: New Essays, John Christman and Joel Anderson, eds., Cambridge: Cambridge University Press, 2005; Sigurdur Kristinsson, "The Limits of Neutrality: Toward a Weakly Substantive Account of Autonomy," Canadian Journal of Philosophy, 30 (2) June 2000, 257-286. 15 to provide a meaningful alternative theory of autonomy. My argument will proceed in two steps. First, I will argue that weak substantive standards are too weak to avoid the objection raised by proceduralist views of autonomy. As a result, I suggest that weak substantive views must incorporate more normatively rigorous standards in order to meet the objection. However, I will argue that the introduction of these standards will collapse weak substantivism into strong substantivism. This is because both views will count the very same cases as autonomous and nonautonomous. In other words, I am arguing that there is a distinction without a difference between weak substantive and strong substantive theories. From there, I present the second part of my argument. Because strong substantive standards are inconsistent with an important aspect of Kymlicka's liberalism and because weak substantive standards collapse into strong substantive standards, Kymlicka's view of autonomy in his liberalism is not weak substantive. Given that proceduralism, strong substantivism, weak substantivism, and socio-relational views exhaust the kinds of views autonomy in the literature and if Kymlicka's view is not proceduralist, nor strong substantivist, nor weak substantivist, then his view of autonomy must be socio-relational. I defend this claim in Chapter 4 and to make my argument, I appeal to Marina Oshana's account of socio-relational autonomy.30 According to this view, autonomy is largely constituted by the social conditions in which a person is embedded. A person is autonomous when her social conditions allow her to exercise de facto authority over her life. According to Oshana, a socio-relational view of autonomy is attractive because it avoids the problems that plague the other dominant views of autonomy and is more in line with our considered intuitions about what it is to 30 Marina Oshana, Personal Autonomy in Society, Burlington, VT: Ashgate Publishing, 2006. 16 be autonomous. However, I argue that Oshana's account of autonomy imposes restrictions on the kinds of value commitments an individual can hold and qualify as autonomous. In other words, Oshana's view of autonomy is not content-neutral. If this line of thinking is correct, then this poses a serious problem, not only for her view, but for the central claim of my project as well. To avoid these problems, I argue that we can modify Oshana's socio-relational view by incorporating content-neutrality. While this will weaken the standards for autonomy, I will argue that this move allows Oshana to avoid serious objections to her view and allows me to preserve my claim that Kymlicka's account of autonomy is socio-relational. However, one long-standing criticism against Kymlicka's particular approach to liberalism is that it cannot consistently maintain a commitment to neutrality.31 Here's why. As we have seen, defenders of APCL (such as Kymlicka) say that liberalism is based upon an idea of the good life. But, critics claim that once we based liberalism upon a claim about what a flourishing life consists in, the state cannot help but to act for perfectionist reasons, i.e. to act to ensure that individuals really lead flourishing lives. If the state cannot refrain from acting in this way, in what sense, then, is the liberal state neutral? This problem seems to be exacerbated if I am right that Kymlicka incorporates a socio-relational theory of autonomy in his liberalism. By endorsing a socio-relational view of autonomy, Kymlicka is making a claim not only about the goodness of autonomous living, but about the kinds of social relations in which citizens ought to live. It seems transparently clear that Kymlicka is making perfectionist claims within his liberalism. Moreover, if these social relations are found to be lacking in the sense that 31 Quong, Liberalism Without Perfection, 22-26. 17 they prevent autonomous living, it seems the state is justified in stepping in to enact policies and laws that address the state of these social conditions. In doing so, however, the state would be appealing to controversial reasons to justify these laws and policies. In other words, Kymlicka's view is inconsistent because it endorses both perfectionism and neutrality. In Chapter 5, I respond to this concern. First, I argue that we must be clear on precisely what "neutrality" is and what kinds of actions that a neutral state may take without violating its commitment to neutrality. While some theorists have characterized the neutral liberal state as "indifferent" toward its citizens, I think this way of describing the state is not only incorrect, it is misleading.32 Instead, we should understand neutrality as consisting of two commitments: neutrality of aim and neutrality of justification. Furthermore, I will argue that while Kymlicka's critics are correct to say that his liberalism incorporates both perfectionist and antiperfectionist aspects, there is no inconsistency. 32 Chandran Kukathas, "Liberalism and Multiculturalism: The Politics of Indifference," Political Theory, Vol. 26, No. 5, October 1998, 686-699. CHAPTER 2 PROCEDURALIST THEORIES OF AUTONOMY Introduction I said earlier that, in rough terms, proceduralists regard autonomy as a psychological feature of an agent and thus claim that the standards or criteria for autonomy are internal or psychological in nature. Many autonomy theorists maintain that proceduralism represents the most reasonable and most intuitive view of autonomy, even if they disagree over how to cash out these standards. What is especially notable about such views, however, is the following, namely that proceduralists argue that internal or psychological standards are both necessary and sufficient for autonomy. According to proceduralists, our evaluations about a person's autonomy depend solely upon the kinds of psychological states that result (or could result) from engaging in the right sort of critical reflection. No other considerations play a role in deciding whether a person is autonomous, including the kinds of social conditions in which a person lives. In this chapter, I will elaborate on John Christman's proceduralist account, which he refers to as an historical approach to autonomy. Indeed, throughout this project, I focus on Christman's account for the following reasons. For one, autonomy theorists regard it as one of the most influential accounts of proceduralism in the literature. For another, Christman has developed and refined his view partly in response to the problems facing other proceduralist conceptions of autonomy. In what follows, then, I understand 19 Christman's view as representative of proceduralism, while also acknowledging that other theorists may conceptualize the view in slightly different terms. My goal in this chapter is to argue that Kymlicka's view of autonomy he employs in his theory of liberalism cannot be proceduralist. This is because Kymlicka and proceduralist theories are at odds over the role of social conditions in a theory of autonomy. As I will discuss in this chapter, while proceduralist theorists like Christman regard social conditions as only causally related to any reasonable theory of autonomy, Kymlicka views social conditions as part of the defining conditions for autonomy. Given that there is this deep disagreement between proceduralism and Kymlicka's view of autonomy, his view cannot be proceduralist. To establish my claim, I examine Christman's account of proceduralism in "Christman on Proceduralism" and Kymlicka's view of autonomy in "Kymlicka and Proceduralism." However, before this, I think it will be instructive to examine briefly the debate among autonomy theorists about the role of social conditions in a theory of autonomy. Autonomy and the Social World In the 1980s, theorizing about autonomy took a "social turn," wherein theorists began to take seriously the idea that social factors ought to play a central role when we think about and conceptualize autonomy.33 The "social turn" was fruitful for theorists, for it introduced new concepts and engendered new ways of thinking about autonomy and the autonomous agent. My aim here is not to rehearse them all but I want to consider two 33 Theorists generally regard Jennifer Nedelsky as the first theorist who argued for this claim. See "Reconceiving Autonomy: Sources, Thoughts, and Possibilities," Yale Journal of Law and Feminism, Vol. 1, 1989, 7-36. 20 important claims made by autonomy theorists. In turn, these remarks will shed light on understanding why proceduralists conceptualize autonomy in the way that they do. First, theorists proposed a new understanding of the individual who was capable of autonomy. With the "social turn," autonomy theorists challenged the historically prevalent, traditional conception of the solitary, self-sufficient autonomous individual as a myth. In its place, theorists introduced the "social self" and an important part of this shift was to make apparent both the capacities and limitations of persons. Contrary to the traditional conception of the self, individuals capable of autonomy were not "social atoms," independent from social influences and relations with others, but rather, as persons who are embedded within particular networks of social relations and who are (to a large extent) shaped by their social environment. Traditional conceptions of the self envisioned the rational and emotional capabilities of individuals as boundless: individuals could engage in radical self-creation, rationally selecting their ends and value commitments34 from a wide-range before them and were seen as capable of ignoring or setting aside entirely interpersonal relations, if one deemed it necessary to achieve one's ends. In contrast, autonomy theorists now conceptualize the self more modestly, in that they do not regard individuals as capable of engaging in that kind of radical self-creation. According to the "social self" thesis, individuals do not "make themselves" ex nihilo by self-determining their ends and value commitments, but rather, they are constructed by their interactions with other individuals in the social sphere. Moreover, theorists pointed out that individuals do not in any robust sense "choose" their value commitments from a vast array of possibilities, but rather that individuals "find" or "recognize" the values that 34 A person's value commitments can refer to her philosophical or religious beliefs, character traits, or preferences. 21 matter to them as a result of being embedded with social relations. Furthermore, theorists claimed that the self cannot be separated from the social, historical, and physical location in which it is embedded. In other words, individuals could not escape the social forms and relations that influenced their self-development. Yet, most theorists caution against interpreting this claim to suggest that external forces totally determine individual selves or that changing particular aspects of oneself is impossible. Rather, it is because selves are embedded in this manner that makes it possible for selves to become selves in the first place or to alter or transform identity-conferring aspects of the self. That is, being embedded within a particular social, historical, and physical location provides a frame of reference for individual selves and supplies the "resources" needed when an individual imagines herself otherwise. However, while this embeddedness plays a crucial role for knowing oneself and knowing the world, a significant limitation also emerges. The social self, unlike the traditional conception of the self, has limited rational and reflective abilities because it is embedded. Agents cannot, as it were, fully "step back" from their own particular location, to reflect uninfluenced by their own perspective. One's critical reflections will be piecemeal, partial, and incomplete and this includes reflections upon one's own psychological states. Aspects of one's identity may not be immediately accessible to one via introspection and one's perception and assessment of one's psychological states may not be veridical.35 35 Iris Marion Young observes: "Because the subject is not a unity, it cannot be present to itself, know itself. I do not always know what I mean, need, want, desire, because meanings, needs, and desires do not arise from an origin in some transparent ego." See Justice and the Politics of Difference, Princeton: Princeton University Press, 1991, 231-2. 22 Along with the new concept of the individual-as-socially-constructed, theorists also turned their attention to the nature of autonomy itself. Some theorists suggested that if the "social turn" shifted our understanding of the agent, then our concept of autonomy ought to be re-thought as well. In thinking about the concept of the person, theorists asked: how do social conditions influence individual self-development? Now, in thinking about the concept of autonomy itself, these theorists asked: how do external social conditions relate to the concept of autonomy? By social conditions, I refer to the wide range of phenomena that govern the relations between individuals as well as determine the kinds of ends individuals might pursue. Social conditions might refer to economic systems as well as religious traditions, social norms and institutions, and cultural and artistic practices. Among autonomy theorists, the debate over social conditions concerns the role that social conditions play when we conceptualize autonomy. Do social conditions play a background role in autonomy? If so, then we needn't reference social conditions when deciding whether a person is autonomous. Or is it the case that social conditions are part of the definition of autonomy, such that if these conditions were deficient or absent altogether, then individuals would fail to count as autonomous?36 According to the proceduralist, social conditions are part of the background conditions for autonomy, rather than part of the definition of autonomy. Another way to understand the proceduralist position is to say that social conditions are causally necessary for autonomy as opposed to conceptually necessary, and here is the 36 Both a strong substantivist theory of autonomy and a socio-relational theory make this claim. I discuss these views in Chapters 3 and 4, respectively. 23 difference.37 The proceduralist would identify education or an adequate range of options as important background conditions for autonomy because the former can contribute to or impede the development of a person's capacity for autonomy while the latter can contribute to effective critical reflection on one's value commitments. However, in the final analysis to determine whether one is autonomous, proceduralists consider only one's psychological states, and not the particular background features of one's social environment. On the proceduralist view, the explanation for why a person counts as autonomous has nothing to do with external social conditions. How, then, does the proceduralist defend the claim that conditions focused on an agent's critical reflections (rather than her external conditions) are necessary and sufficient for autonomy? Defenders of proceduralism claim that the view is motivated by two, different intuitions about autonomy. The first intuition is this: autonomy seems to require a large space of noninterference in which individuals can identify and pursue their good as they see fit. Proceduralist theorists argue that one important way to capture this intuition is to exclude social conditions in the definition of autonomy. After all, requiring that autonomous agents live in particular kinds of social conditions reduces the space of noninterference because some conditions will be deemed as incompatible with autonomous living. However, if we identify relatively thin conditions for autonomy, this provides to individuals the greatest possible space to self-govern, or to identify and to live according to the beliefs they deem valuable. 37 Holger Baumann makes the same distinction in "Reconsidering Relational Autonomy: Personal Autonomy for Socially Embedded and Temporally Extended Selves," Analyse and Kritik, 30 (2008), 445- 468, 447-448. 24 The second intuition that motivates proceduralism is this: value commitments "are valid for a person when she can autonomously come to see their import."38 Here's what I take the proceduralist to be saying: unless I decide and endorse my values from my own perspective, I will not see the significance of these values for myself. We can put the point another way: when someone or something external to my perspective determines the values I should endorse, I cannot see the validity of these values for myself. According to its defenders, proceduralist views capture this intuition in another important feature of the view: proceduralist theorists place no normative restrictions on the kinds of value commitments autonomous agents can endorse or the kinds of choices they can make. This is why autonomy theorists refer to proceduralist accounts as content-neural. There is no "right" set of beliefs or value framework that an individual must endorse to be autonomous. Instead, what matters for a person's autonomy is that she engages (or could engage) in the proper critical processes. Proceduralism denies that it has a complete perspective, decided in advance, on the kinds of lives an autonomous agent might lead but rather, leaves it up to agents to determine how they might live. Proceduralism simply identifies a decision-procedure for autonomy, rather than a prescription for the good life, and this ensures that individuals are living according to values they endorse from their own perspective. Christman on Proceduralism So far, I've made only general remarks about the proceduralist view of autonomy and the intuitions behind it. It is my claim that Kymlicka's view of autonomy is not proceduralist, but before I make this argument, I want to expand upon Christman's 38 Christman, "Procedural Autonomy and Liberal Legitimacy," 281. 25 particular conception of proceduralism. However, one might ask: why consider only Christman's view, especially given that there are other proceduralist views in the literature? As I mentioned earlier, Christman has refined his view over a long period of time, so autonomy theorists regard it as a sophisticated view that bears out the two motivating intuitions I mentioned above.39 Nevertheless, whether I consider one account of proceduralism or all of them, my overarching point would remain the same: Kymlicka's view of autonomy can't be proceduralist and this is true of any account of proceduralism. So, in making my argument about Kymlicka's view of autonomy, my argument is not that there is something about Christman's view in particular which explains why Kymlicka's view is not proceduralist. If this was my claim, then we could simply appeal to another proceduralist view in the literature. As I mentioned earlier, proceduralists view autonomy as an internal, psychological phenomenon. To put the proceduralist position somewhat crudely, autonomy is a function of "what is happening in your mind," rather than "what is happening in your external environment." We can clarify what the proceduralist position by considering the following scenario.40 Jane lives within a community which is characterized by gendered social hierarchies and which subscribes to very traditional gender roles as a matter of religious faith. In Jane's community, women are strictly wives and mothers, while men are pursue careers outside the home. While women are expected to manage the households and raise children without complaint, they are also expected to "submit" to their husband's authority in all important economic, social, and sexual 39 Christman first elaborates on his historical view in 1990 in "Autonomy and Personal History." 40 I take this example from the Quiverfull community. See Kathryn Joyce, Quiverfull: Inside the Christian Patriarchy Movement, Boston: Beacon Press, 2009. 26 decisions, as well as decisions about how many children they will have. Women are required to be subordinate to male authority because women in Jane's community are regarded as inferior to men, both socially and spiritually. Certainly, some of us, including proceduralists, may recoil at these kinds of social relations, not only because they are deeply inegalitarian, they also seem oppressive. Nevertheless, Jane is fully committed to her community and her faith, and she thinks that the social arrangements in which she lives are the best kind of social relations to live in. Is Jane autonomous with respect to her value commitment to female subordination? On Christman's proceduralist account, this answer would be "yes," providing that Jane meets (or could meet) the following conditions. If Jane were to reflect critically upon her value commitment, in light of the historical processes by which it arose, and would not feel alienated from it, then she is autonomous.41 If a person's reflections would produce a sense of alienation, then she is nonautonomous with respect to that commitment.42 Like other proceduralist accounts, Christman's view emphasizes reflective competency and agential authenticity as necessary components of autonomy. Competency conditions ensure that a person is self-governing while authenticity 41 Christman's most recent defense of his view is in The Politics of Persons. As Christman notes, his earlier formulation of his view of autonomy claimed that a person is autonomous relative to a value commitment provided that she approves of or would not resist the historical processes by which it arose. Some of Christman's critics, however, argued that a person's attitude toward the historical processes that give rise to her value commitments isn't as relevant for autonomy as her attitude toward the commitments themselves. For example, Al Mele argues that a person could be raised to hold particular religious commitments in a very restrictive social environment and accept her commitments while rejecting the kind of upbringing to which she was subject (Autonomous Agency: From Self-Control to Autonomy, New York: Oxford University Press, 1995, 138-139). Christman acknowledges this point and this is why he changes his view of autonomy to place more importance on reflection upon the value itself given its historical development, rather than merely the development itself. 42 Notice that I say she is nonautonomous with respect to that commitment. It's important to make this kind of distinction because even if a person isn't autonomous with respect to one value commitment, this doesn't render her nonautonomous as such. 27 conditions ensure that the governing is "her own." So, on the one hand, the conditions for competency "pick out those characteristics by which a person effectively makes competent decisions," and these include "rationality, self-control, freedom from psychosis and other pathologies, access to minimally accurate information, motivational effectiveness and the like."43 On the other hand, the conditions for authenticity refer to "requirements that the person's values and decisions are truly her own; these most often include the condition that persons reflect on their personal characteristics and identify with (or at least not feel deeply alienated from) them."44 But, unlike some other proceduralist accounts, Christman's view is an historical approach to autonomy. I think it will to be helpful at this point to make some brief observations about Christman's reasons for adopting an historical account of autonomy in the first place. Many proceduralist theorists have defended what is referred to as a "hierarchical" account of autonomy. On this view, a person must engage in second-order reflection on her first-order value commitments claim and endorse them or identify with them, in order to be autonomous.45 Hierarchical views have been very influential in the literature, largely in part to their intuitive appeal. If autonomy is self-governance on the basis of values that are "one's own," then hierarchical views seem to provide a reasonable procedure for identifying and endorsing one's values. Moreover, hierarchical views are 43 Christman, "Autonomy, Self-Knowledge, and Liberal Legitimacy," 333. 44 Christman, "Autonomy, Self-Knowledge, and Liberal Legitimacy," 333. 45 The most well-known advocate of this type of view is Harry Frankfurt in "Freedom of the Will and the Concept of a Person." For example, suppose that I have a (first-order) preference to attend school and earn a degree. Frankfurt's hierarchical theory of autonomy would say that I am autonomous with respect to this preference provided that I reflect upon it at the second-order level - "Do I want to want to attend school?" If, after reflecting, I decide that I want to want this and that I endorse this preference as my own, then I am autonomous with respect to it. 28 content-neutral. What matters for autonomy is that individuals submit their values to the proper critical evaluations, and not the substantive content of these values. Nevertheless, even critics sympathetic to proceduralism have raised several objections against hierarchical views, but rather than rehearse all of them, I consider two here because they will shed further light on Christman's motivations behind his historical view of proceduralism. First, a problem with hierarchical accounts is that they cannot account for the following intuition, namely that if a person endorses her value commitments as a result of oppressive socialization or manipulation, then she is not autonomous.46 For example, suppose that all of my value commitments were implanted in me two days ago by a malevolent neurosurgeon. Suppose further that I reflected upon and endorsed my first-order value commitments at the second-order level. On a hierarchical view, I would count as autonomous, but this is counterintuitive. I may have reflected on and endorsed "my" commitments, but given that I hold my values simply on the basis of what the neurosurgeon decided, it seems odd to say that these are my values. Similarly, suppose I have been oppressively socialized to hold certain beliefs about the rightness of gender hierarchy. Even if I reflect upon my first-order commitments at the second-order level and thus count as autonomous, many theorists think this is counterintuitive. This is because individuals who are oppressively socialized are trained to identify with the values and beliefs that arise from their socialization. As such, it's not clear whether an oppressively socialized individual's reflections are "her own" or the product of her oppressive upbringing. Christman agrees with these criticisms of hierarchical accounts, 46 Christman, The Politics of Persons, 137. 29 but thinks that incorporating historical conditions - that is, conditions concerning the historical process by which an individual acquires her values - will render hierarchical accounts more intuitively appealing.47 On Christman's view, the historical constraint works to ensure that individuals who are oppressively socialized or manipulated by malevolent neurosurgeons do not count as autonomous by considering the process behind the formation of their beliefs. Second, critics of hierarchical views have taken issue with the idea of "identify with." As Christman notes, "the notion of ‘identification' is problematically ambiguous between acknowledgement and endorsement."48 On the one hand, I might aim to be a more generous person and identify with this trait, but my failure to be generous "should not disqualify [me] from being autonomous."49 On the other hand, I might identify with my preference for smoking cigarettes as "part of who I am" but identifying with this particular preference does not seem to establish my autonomy. Moreover, identification is too rigorous to allow for the possibility that a person might feel ambivalent toward her value commitments. As Christman points out, some of us have "elements of the self" which we would not reject but which we do not regard as "ideal from our point of view."50 Put another way, some of us have traits that we do not identify with fully but acknowledge as "part of who we are." Nevertheless, Christman argues that ambivalence shouldn't make a person nonautonomous, so Christman aims to identify a conception of 47 Christman, The Politics of Persons, 137. 48 Christman, The Politics of Persons, 143. 49 Christman, The Politics of Persons, 143. 50 Christman, The Politics of Persons, 143. 30 autonomy that "capture[s] this reality."51 These considerations lead him to think that nonalienation is the proper test for determining whether a person is autonomous because nonalienation doesn't require that individuals identify whole-heartedly with their value commitment. Alienation, in the sense Christman intends, picks out a reaction to one's value commitment in question. This reaction isn't ambivalence or indifference but rather that one feels constrained by a value commitment and wants to reject it.52 Notice that Christman conceptualizes alienation as a necessary "combination of judgment and affective reaction."53 This distinguishes Christman's view in another way from standard accounts of proceduralism, which he claims are "overly cognitive...stressing the ability to make rational, detached, and calculative judgments about the acceptability of a trait."54 So, a person must critically reflect on her commitment but her judgment about its acceptability has an affective aspect to it. There are three other important aspects of Christman's view of reflection besides the cognitive and affective components. First, as we have already noted above, reflection is counterfactual: an individual isn't required to reflect in order be autonomous. Rather it must be true of her that if she did reflect on her value commitment she would endorse it. Second, it is piecemeal because an individual does not have to reflect upon all of her 51 Christman, The Politics of Persons, 143. 52 Christman, The Politics of Persons, 143. Christman emphasizes a page later that "the key element of such alienation is this resistance, the anxious sense that the fact in question is constraining..." (144). 53 Christman, The Politics of Persons, 144. 54 Christman, The Politics of Persons, 144. 31 value commitments, but rather only the commitment in question.55 Third, an individual's analysis of the value commitment in light of its historical development is (in Christman's terms) "fully subjective."56 The appraisal, in other words, depends upon her own attitude toward the value commitment, rather than relying upon judgments or values that are independent of her perspective. What is important to notice about all of these conditions, however, is that they make no reference to "what is happening in the social environment." Rather, the conditions for autonomy on a proceduralist view make reference only to an individual's psychological processes and states. What this means is that proceduralists think that it is possible even for Jane to be autonomous, even though she lives within highly restrictive, traditional social arrangements, because autonomy from the perspective of proceduralism is not about adopting the "right" value commitments or living in the "right" social relations.57 Rather, autonomy is about engaging in the proper critical reflection and adopting the right sort of attitudes toward one's value commitments. To be clear, however, Christman acknowledges that social conditions might interfere with the development or the exercise of a person's capacity for autonomy.58 Given this, Christman also places further constraints on the competency conditions for 55 The motivation for the "piecemeal" reflection requirement stems from theorists' criticisms that it is neither possible nor desirable for an individual to critically reflect upon all of one's value commitments, particularly those that one regards as self-defining and unrevisable. See Michael Sandel, Liberalism and the Limits of Justice, New York: Cambridge University Press, 1982. 56 Christman, The Politics of Persons, 142. 57 The idea that it is possible for any individual to be autonomous, provided she reflects in the right way and regardless of her social conditions, might strike some as counterintuitive and I consider this line of thought in Chapter 3. 58 Christman, The Politics of Persons, 146. 32 autonomy, i.e. those related to the characteristics necessary for effective decision-making. Christman stipulates further that a person's reflections cannot be "the product of social and psychological conditions that prevent adequate appraisal of oneself."59 This means that one must be free from certain factors (e.g. blinding rage, drug or alcohol addiction, depression) and one must have the ability and the freedom to evaluate aspects of one's personality and social conditions, which requires that one has "minimal education" and "exposure to alternatives."60 Thus, if an agent is critically reflecting upon her value commitments, but she is under the psychological sway of a powerful cult leader, she would not count as autonomous on Christman's account, even if she felt no alienation toward her commitments. Nevertheless, the inclusion of these constraints on the competency conditions for autonomy does not change the proceduralist theorist's underlying assumption, namely that, in the final analysis, what establishes whether a person is autonomous is a function of her psychological states and not her social conditions. Proceduralist theorists such as Christman deny that social conditions are part of the defining conditions for autonomy, but acknowledge that social conditions can interfere with the development or exercise of a person's autonomy. For proceduralists, then, social conditions play only a causal role in the debate over autonomy. Kymlicka and Proceduralism In this section, I argue that the view of autonomy Kymlicka incorporates in his theory of liberalism is not a proceduralist view. To make my case, I begin by considering the following argument. At first glance, one might think that Kymlicka's view of 59 Christman, The Politics of Persons, 146-147. 60 Christman, The Politics of Persons, 147. 33 autonomy is proceduralist because Kymlicka's account and proceduralism share two important motivating assumptions. As we saw above, proceduralism is motivated by the intuition that a person's values are valid for her only when she endorses them from her own perspective. This is why defenders of proceduralism maintain that any reasonable theory of autonomy should be content-neutral. Autonomy shouldn't require that individuals endorse (or reject) particular values in order to be autonomous because individuals will not recognize the import of values deemed "right" by an external perspective. Along the same lines, at the heart of Kymlicka's view of liberalism is the idea that each of us has an essential interest in leading a good life.61 But, as Kymlicka argues, I live a good life only when I am leading it on the basis of beliefs I identify as valuable. Liberal thinkers refer to this as the "endorsement constraint": there is no way of living that is good for me unless I endorse it from my own perspective.62 This is why autonomy is crucial for Kymlicka's formulation of liberalism because being autonomous helps individuals identify what is valuable. Kymlicka also recognizes, however, that it is possible for individuals to be mistaken about what is valuable and "no one wants to lead a life based on false beliefs."63 But, even if a person can be mistaken in this way, it does not follow that his life will go better if he is compelled to live his life according to values he doesn't endorse. Rather, his life goes better only when he is "leading it from the inside, according to [his] beliefs 61 Kymlicka, Liberalism, Community, and Culture, 10. 62 Ronald Dworkin states that "my life cannot be better for me in virtue of some feature or component I think has no value," Sovereign Virtue: The Theory and Practice of Equality, 268. 63 Kymlicka, Multicultural Citizenship, 84. 34 about value."64 So, on Kymlicka's view, individuals not only must have the critical and reflective capacities necessary to determine one's values in order to be autonomous, they must also have the ability to rationally revise their beliefs about what is valuable.65 According to Kymlicka, in order for a person to have this ability, he must have access to information about his options, as well as a minimal level of education to be able to think about his options in an intelligent and informed manner. On Kymlicka's view, then, autonomy requires that individuals have the space to figure out whether their current values are the values they want to hold and to change their minds and pursue different values. So, I think it's reasonable to say that Kymlicka shares the proceduralist intuition identified above: autonomy requires a wide space of noninterference in which individuals can identify and pursue their values. Nevertheless, even though both Kymlicka and the proceduralist accept the intuition that autonomy requires a wide space of noninterference, they disagree over how to cash out this intuition. Put another way, they differ over how to create this space. As we have seen, Christman argues that this wide space is created by making the conditions for autonomy as thin as possible. For Christman, this means excluding social conditions from the definition of autonomy. On Christman's view, to impose normative restrictions on the kinds of social conditions in which autonomous individuals must live only serves to narrow the space of noninterference and thus may prevent some individuals from pursuing their good as they see fit. For example, if we say that a person's social conditions cannot be characterized by relations of domination and 64 Kymlicka, Liberalism, Community, and Culture, 12. 65 Kymlicka, Multicultural Citizenship, 84. 35 subordination in order for her to be autonomous, it is difficult for Jane to pursue her beliefs about the rightness of female subordination. In contrast, while Kymlicka seems to share to the intuition that autonomy requires a wide space of noninterference, he seems to think that excluding social conditions from the definition of autonomy will only serve to threaten this space. Put another way, I argue that Kymlicka rejects the idea that social conditions are only causally related to a theory of autonomy and instead thinks that social conditions are partly constitutive of autonomy. This means that if certain social conditions are missing from a person's life, then that person does not qualify, by definition, as autonomous. To make my argument, I want to return briefly to Christman's claim about the role of social conditions in autonomy. As we have seen, proceduralists think that social conditions can cause or impair the development of a person's capacity for autonomy. Proceduralists have been particularly interested in the way in which oppressive social conditions can interfere with a person's capacity for autonomy.66 But, proceduralists also suggest that it is both empirically and logically possible for a person to live in oppressive social conditions and be autonomous, provided she exercises (or can exercise) her capacity for autonomy in the proper way. This is why someone like Jane can count as autonomous on the proceduralist view. While we might question whether this understanding of the relation between oppression and autonomy is sound, I don't consider this debate here.67 For our purposes, the take-home message with respect to proceduralism and oppressive social conditions is this. Because it 66 For informative discussions about this issue, see Meyers, "Feminism and Women's Autonomy: The Challenge of Female Genital Cutting," and Friedman, Autonomy, Gender, Politics. 67 Critics of proceduralism's take on oppression and autonomy include Stoljar, "Autonomy and the Feminist Intuition," and Anita Superson, "Deformed Desires and Informed Desire Tests," Hypatia 20 (4) Fall 2005: 109-126. 36 is both empirically and logically possible on the proceduralist view for a person to live in an oppressive social environment and be autonomous, this means that certain social conditions can be absent altogether. For example, the proceduralist view can allow that a person is both autonomous and at the same time denied certain rights and freedoms, provided she is not denied the opportunity to develop and exercise her capacity for critical reflection. However, I argue that Kymlicka rejects this view. To understand why, recall that, on Kymlicka's view, leading a good life requires leading it from the inside on the basis of values one identifies as important and this requires that one is autonomous. To be autonomous, Kymlicka claims that "individuals must therefore have the resources and liberties needed to lead their lives in accordance with their beliefs about value, without fear of discrimination or punishment."68 Because individuals must have these conditions to be autonomous, liberal states are required to uphold the individual's right to privacy and remain neutral or refuse to promote a particular moral framework.69 In addition, Kymlicka argues that autonomy requires that individuals "be free to question those beliefs, to examine them in light of whatever information, examples, and arguments our culture can provide. Individuals must therefore have the conditions necessary to acquire an awareness of different views about the good life, and an ability to examine these views intelligently."70 Because individuals must have these conditions to be autonomous, the liberal state must grant to individuals various rights and freedoms, such as freedom of association, conscience, and expression and 68 Kymlicka, Multicultural Citizenship, 81, italics mine. For a similar statement, see Liberalism, Community, and Culture, 12-13. 69 I address the issue of neutrality and Kymlicka's view of liberalism in Chapter 5. 70 Kymlicka, Multicultural Citizenship, 81, italics mine. 37 mandates a minimum level of education for all individuals, for these rights and freedoms make it possible for individuals to learn about other ways of living and value frameworks and to assess them critically. So what can we conclude about Kymlicka's statements about what individuals "must" have to be autonomous? I suggest that they provide evidence that he rejects the idea that social conditions are only causally related to autonomy and instead thinks that they are partly constitutive of autonomy. Without these conditions in place, a person is not autonomous by definition on Kymlicka's view. For example, I must have the right to privacy to be autonomous for without this right, I may be persecuted for holding unconventional beliefs. If I am fearful of persecution, I may not follow my preferred beliefs but rather hold whatever beliefs are deemed "proper." But, if I am leading my life on the basis of beliefs that I do not regard as valuable, then I am not autonomous. Along the same lines, I must have freedom of conscience to be autonomous for without this right, I may be prevented from questioning my socially inherited beliefs. If I am not permitted to engage in this kind of questioning, I may be leading my life on the basis of beliefs I do not deem as valuable from my own perspective. If so, I am not leading my life from the inside, according to what I deem as valuable, and so I am not autonomous. If I'm not autonomous, then I am not leading a good life. However, we have seen that proceduralists maintain that individuals may qualify as autonomous in any kind of social environment, even one characterized by oppressive conditions. If so, then it must be the case on the proceduralist view that individuals can live in an environment without the various rights and freedoms Kymlicka discusses in his view and still be autonomous. However, given that Kymlicka denies that a person can be 38 autonomous without these rights and freedoms, then the view of autonomy he incorporates in his view of liberalism cannot be proceduralist. Concluding Remarks In this chapter, I have argued that the view of autonomy Kymlicka incorporates in his theory of liberalism cannot be proceduralist. This is because proceduralism denies that social conditions are part of the defining conditions of autonomy. In other words, the kinds of social conditions in which a person lives do not determine (or partly determine) whether she is autonomous. Only a person's psychological states matter for determining a person's autonomy. Proceduralist theorists claim that a person's social conditions play only a causal role in the development of a person's capacity for autonomy. However, as I have argued, Kymlicka denies that social conditions are only causally related to autonomy. Rather, I made the claim that Kymlicka maintains that social conditions are part of the defining conditions for autonomy and that they must obtain in order for a person to be autonomous. If I am right that Kymlicka views social conditions in this way, then he is not offering a proceduralist view in his liberalism. Instead, he is offering what autonomy theorists refer to as a substantive account of autonomy because it claims that social conditions are partly constitutive of autonomy. However, just as there are conceptual variations in theories of autonomy, there are different substantive views. In the following chapter, I consider what theorists refer to as a strong substantive view of autonomy. CHAPTER 3 STRONG AND WEAK SUBSTANTIVE THEORIES OF AUTONOMY Introduction In the previous chapter, I concluded that Kymlicka's view of liberalism does not incorporate a proceduralist view of autonomy. As I argued, Kymlicka's account of autonomy claims that social conditions are part of the definition of autonomy, yet this is what proceduralist views deny. I claimed in the conclusion of Chapter 2 that because Kymlicka introduces social conditions as part of the defining conditions of autonomy, his view of autonomy must be substantive. However, as I noted previously, there are different interpretations of substantivist accounts of autonomy, so we need to establish which view Kymlicka employs in his theory of liberalism. In this chapter, I have two main goals. First, I elaborate on one important substantivist view in the literature: strong substantivism. I discuss this in more detail shortly, but for now: on this view, a person must engage in the proper critical reflection and her choices and preferences are subject to certain normative constraints. As we will see, Kymlicka's view of autonomy is not a strong substantivist view, because it offers different substantive conditions for autonomy than those offered by the strong substantivist. However, my claim isn't simply that these two views are dissimilar. Rather, I contend that strong substantivism is incompatible with another aspect of Kymlicka's liberalism, namely the endorsement constraint. This 40 constraint, which we considered in Chapter 2, holds that a person's life doesn't go better if it is led from the "outside," according to beliefs he doesn't endorse. Because Kymlicka regards the endorsement constraint as a fundamental feature of his liberalism, he cannot incorporate a theory of autonomy which infringes on this constraint. My second aim in this chapter is to introduce another substantivist theory of autonomy, which theorists refer to as a weak substantive view. I elaborate in greater detail shortly, but for now we can note that such views claim that a person is autonomous if she critically reflects in the right way and she holds the right psychological attitudes toward herself. One might wonder why I'm not devoting a separate chapter to weak substantive views, given that I am doing so for the other main accounts of autonomy. However, my reasons for doing so will become clear when I develop my argument against weak substantivism. According to its defenders, weak substantive views are attractive because they seem to escape the problems raised by proceduralism and strong substantive views and thus provide a more reasonable account of autonomy than its rivals. Here, I argue that weak substantive theories fail to do this. I will argue that the weak substantivist view collapses into strong substantivism and so does not represent a viable alternative view of autonomy. If so, then Kymlicka's account of autonomy in his liberalism cannot be a weak substantivist view, because weak substantivist views are merely strong substantivist views in disguise and I will have shown that Kymlicka's view of autonomy cannot be strong substantivist. All autonomy theorists grapple with what I shall call the "content" question: Should a theory of autonomy place direct restrictions on the content of an agent's value commitments, e.g. her choices, preferences, beliefs, and character traits? Imposing these 41 kinds of restrictions would require autonomous individuals to make (or avoid making) certain choices or to hold (or avoid holding) certain preferences. As I noted above, strong substantivists answer "yes" to this question. Proceduralists, however, answer "no" to the content question. As we noted in Chapter 2, proceduralists are guided by the intuition that a person's values are valid for him only when he endorses them from his own perspective. If a theory of autonomy requires that individuals accept (or reject) a particular value, there is the possibility those individuals don't regard that value as important from their own perspective. So, proceduralists think it is counterintuitive for a theory of autonomy to impose normative constraints on the content of an individual's value commitments. In this chapter, I elaborate upon a second concern proceduralists raise with respect normative content in a theory of autonomy. According to Christman, a theory of autonomy which imposes normative restrictions on the content of individuals' choices and preferences run the risk of justifying paternalistic intervention in their lives. On the face of it, it appears that proceduralist critics raise compelling concerns about placing such restrictions on the content of choices and preferences. Moreover, as we will see, other autonomy theorists who aren't proceduralists also take issue with strong substantivism for imposing these restrictions. Indeed, all other views of autonomy - except for strong substantivism - maintain that it is inappropriate to place direct restrictions on the kinds of value commitments an autonomous individual can endorse. It appears, then, that strong substantivists face an uphill battle in defending the appeal of the view, given that most autonomy theorists reject its central claim. My first task is to elaborate strong substantivism in more detail, while also providing considerations in favor of the view. 42 Strong Substantivism and Proceduralism on Oppressive Values Why think that a theory of autonomy should place direct restrictions on the kinds of value commitments an autonomous agent can endorse? To make a case for this claim, it will be helpful to return to proceduralism, given that this view denies outright that a theory of autonomy should incorporate any normative restrictions on value commitments. This is why autonomy theorists refer to proceduralist views as content-neutral: they take no stand on the kinds of value commitments an autonomous agent must endorse. However, strong substantivist critics of proceduralism maintain that content-neutrality leads to counterintuitive implications.71 To understand what these implications are and why proceduralism leads to them, let's turn to a divisive issue for autonomy theorists: agents who hold value commitments with oppressive content. These represent hard cases because it's not clear whether acting on these commitments diminish a person's autonomy and, if so, how. Can a person make choices on the basis of oppressive value commitments and be autonomous? To understand what I mean by oppressive value commitments, let's return to the example of Jane in the previous chapter. Recall that Jane preferred to submit to her husband's authority in all important matters on the grounds that her religious beliefs dictated this. While autonomy theorists would agree that Jane's value commitment to subordination is oppressive because it deems Jane to be in a morally subordinate position, theorists disagree over whether it is possible for Jane to hold and act on her commitment and count as autonomous. I noted previously that proceduralists like Christman would deem Jane's choice to be subordinate as autonomous, just in case she 71 To be clear, all critics of proceduralism raise this objection against the view, regardless of the theory of autonomy to which they subscribe. But, autonomy theorists offer different ways to get around the counterintuitive implications, depending upon which theory of autonomy they defend. 43 exhibits (or would exhibit) reflective competence and holds (or would hold) the right psychological attitudes toward her preference, were she to reflect on that preference. On proceduralist views, the content of her commitment is irrelevant for establishing whether she is autonomous. Another so-called hard case for autonomy theorists is an agent who prefers not simply subordination, but slavery. Clearly, a commitment to slavery is oppressive because it regards some individuals not simply as morally inferior, but less than human. However, because proceduralists are committed to content-neutrality, they are committed to the view that even a slave can turn out to be autonomous, provided he is reflectively competent and would hold the right psychological attitudes toward his value commitment for slavery, were he to reflect upon it in the proper way. However, because proceduralism makes room for the possibility of autonomous subordinated wives and (especially) autonomous slaves, strong substantivist critics regard the view as deeply counterintuitive. It's important to be clear on where the counterintuitive implication arises for the strong substantivist. The issue here isn't the historical process by which Jane or the slave arrived at their value commitments or the social environment in which they are embedded. Nor is the issue that Jane or the slave are incompetent reasoners or that the content of their value commitments prevents them from critically reflecting in the proper way. (Of course, this might be the case, but, for the strong substantivist, this is not why proceduralism leads to counterintuitive implications.) Rather, strong substantivists argue that proceduralism is counterintuitive because content-neutrality seems to lead to a paradox. On the one hand, proceduralists hold that individuals are autonomous when they identify and make choices on the basis of preferences determined as valuable from their own subjective point of view. Indeed, 44 proceduralists claim that this is their guiding intuition in formulating their theory of autonomy. On the other hand, because proceduralism upholds content-neutrality, it is a consequence of the view that autonomous individuals can prefer, and make choices on the basis of, subordination or slavery. But, these are ways of living in which an individual does not and cannot determine her preferences and values from her own point of view. How can a person be autonomous, i.e. decide her preferences and values from her own perspective, when others decide them for her? Given that proceduralism seems to result in a paradox, strong substantivist theorists suggest that our theorizing about autonomy should be guided by another intuition: that some ways of living are incompatible with autonomy and living as a subordinate wife or a slave seems to be paradigmatic examples of these ways of living. For strong substantivists, this incompatibility between autonomy and some ways of living arises because of the content of the value commitments that undergird these ways of living, and not because of how these commitments were formed, e.g. through oppressive socialization.72 If we are guided by this intuition, then, we are led to formulating a theory of autonomy which places normative restrictions on the content of an agent's preferences 72 I owe this formulation of the strong substantivist view to Natalie Stoljar. See "Feminist Perspectives on Autonomy," Stanford Encyclopedia of Philosophy, section 8 (forthcoming). I should note that Stoljar attributes the argument above to Thomas Hill and suggests in this entry that Hill is a strong substantivist theorist on the basis of his arguments about the Deferential Wife in his paper "Servility and Self-Respect," (Autonomy and Self-Respect, Cambridge: Cambridge University Press, 1991, 15). However, it's not clear to me that this is an accurate interpretation of his view. For one thing, Hill argues in this paper that the Deferential Wife is servile because she fails to respect herself as a moral agent, but he says nothing about whether she lacks autonomy. For another, Hill's own view identifies seven conditions for autonomy, none of which say anything about holding the "right" preferences and making the "right" choices. In fact, given that his seventh condition is that individuals "have ample opportunities to make use of these [other six] conditions in living a life over which they have a high degree of control," I think there is an argument to be made that Hill's view is socio-relational, not strong substantivist. This is because this seventh condition requires that certain social conditions obtain in order for a person to be autonomous. I leave this line of reasoning for another paper. For Hill's account of autonomy, see "Autonomy and Benevolent Lies," in Autonomy and Self-Respect, 25-42, especially 36. 45 (or, put another way, it must reject content-neutrality). This move not only sidesteps the paradox that arises for proceduralist accounts, it also means that subordinate wives and slaves will never turn out to be autonomous on strong substantivist views. It is important to note, however, that the strong substantivist isn't saying that the mere fact that a person holds preferences with the "wrong" content makes her nonautonomous. Rather, strong substantivists are claiming that if a person holds a preference for subordination and she acts on this preference, then she is not autonomous. Here's why. According to defenders of the view, individuals who choose subservience or slavery make "a special kind of moral mistake."73 For example, Jane's moral error stems from the fact that she makes a choice which denies her standing as a moral equal. Strong substantivists say that the reason Jane makes this error is due to the content of Jane's preference for subordination. The content prevents her psychology from "hooking up" or corresponding to an objective feature of the world, namely the fact that she is morally equal. A person is autonomous on the strong substantivist view if she critically reflects in the proper way and avoids holding the "wrong" preferences and making the "wrong" choices. If a person has the "right" preferences, this will ensure that her psychology is "hooking up" to the world in the proper way, and so will ensure that she makes the "right" kinds of choices, e.g. a non-subordinate life. I've argued in the previous chapter that Kymlicka's account of autonomy in his liberalism in not a proceduralist view. Here, I want to argue a strong substantivist view is incompatible with Kymlicka's view. To make my case, let's begin by considering the following line of thought. One might wonder: why can't Kymlicka simply "add" a 73 Stoljar, "Feminist Perspectives on Autonomy," Stanford Encyclopedia of Philosophy, section 8 (forthcoming). 46 normative restriction condition to the set of conditions he identifies as necessary for autonomy? After all, strong substantivism is attractive because it can explain a moral problem with oppression: it creates and upholds relations of domination and subordination between persons thus denying the equal moral standing of persons. Given that liberals like Kymlicka oppose these kinds of relations, perhaps liberals should adopt a strong substantivist view of autonomy. To provide further support for this line of thought, recall Kymlicka's own observations about the foundations of liberalism. On his view, each of us has an essential interest in leading a good life and it is this interest which forms the basis of liberal political theory. There is a difference, however, between leading a life that is good and leading a life that one believes to be good. For Kymlicka, it is possible that anyone can be mistaken about what the good life consists in and this possibility explains why each of us deliberates, even agonizes, over significant decisions.74 We deliberate because we know we are the kinds of beings prone to making errors in reasoning or we know we may come to regret our choices, even if everything goes according to our rationally decided plan. I may be successful at something, but it doesn't follow that I have good reason to continue on this path or think it is a valuable way to live.75 Indeed, Kymlicka points out that we devote a good deal of time and concern ourselves with thinking about our projects and values and this "only makes sense on the assumption that our essential interest is in living a good life, not the life we currently believe to be good...it is important to us that we not lead our lives on the basis 74 As Kymlicka notes, "We may come to see that we've been wasting our lives, pursuing trivial or shallow goals and projects that we had mistakenly considered of great importance. This is the stuff of great novels," (Liberalism, Community, and Culture, 10). 75 Kymlicka: "I may succeed brilliantly at becoming the best pushpin player in the world, but then come to realize that pushpin isn't as valuable as poetry, and regret that I ever embarked on that project," (Liberalism, Community, and Culture, 11). 47 of false beliefs."76 Let's put these thoughts together. There are some ways of living that are objectively better than others and each of us has an essential interest in leading an objectively good life. Because we can be mistaken about our good, this is why Kymlicka argues that it is necessary for us to be autonomous because being autonomous can help us to come to know what is valuable.77 As we have seen, Kymlicka claims that autonomy requires individuals to have 1) the ability to arrive at our beliefs through a process of rational reflection, 2) the ability to rationally revise our beliefs, should we decide they are no longer worthy of holding, and 3) the social conditions necessary for exercising these abilities. However, the question facing us is this: why can't Kymlicka incorporate a fourth condition for autonomy, one which places restrictions on the value commitments individuals can endorse? After all, Kymlicka thinks that there are objectively good ways of living, so why not embrace a theory of autonomy which clearly specifies these ways of living (or, at least, specifies objectively bad ways of living to avoid)? Furthermore, as Kymlicka himself points out, people sometimes make mistakes about what is valuable, choosing conceptions of the good that are not in fact worthwhile. Why, then, can't Kymlicka appeal to a theory of autonomy that precludes people from holding and acting upon the wrong preferences? If the ultimate goal is for individuals to lead good lives, then why not endorse a theory according to which people are autonomous only if they adopt the values and preferences that ensure they lead good lives? 76 Kymlicka, Liberalism, Community, and Culture, 11. 77 The fact that it is possible for us to come to know what is valuable also explains why Kymlicka rejects the claim that liberalism is based upon scepticism about the good. In fact, he regards this line of criticism against liberalism as "a complete misinterpretation," (Liberalism, Community, and Culture, 18). 48 It is my contention that Kymlicka cannot "add" a normative constraint condition to his current theory of autonomy. Here's why. Kymlicka thinks that being autonomous and embracing objectively good values are both necessary for leading a life that is good, but they are not sufficient. For Kymlicka, there is yet another necessary condition for leading a good life: that one leads it from the inside, according to beliefs and preferences one deems important from one's own subjective perspective. As we noted above, Kymlicka refers to this as the endorsement constraint and enshrines it as a fundamental feature of his liberalism.78 According to this constraint, a person's life goes well only if he leads it from the inside.79 Conversely, a person's life doesn't go better if he leads it from the outside, according to values he doesn't endorse or regard as important from his own perspective. Kymlicka accepts the endorsement constraint on the grounds that it is self-defeating to force people to hold or reject certain preferences or beliefs in order that they lead a good life. Consider Kymlicka's own example.80 Even if it is true that going to church is a valuable activity, we do not help individuals to lead objectively good lives by compelling them to go to church. This is because individuals who are forced to go to church will fail to see the good in going to church for themselves; as such, the activity of going to church will have no meaning for them and their lives won't go better.81 People must identify their own reasons for endorsing a preference or belief; other people cannot supply that reason for them. As Kymlicka points out, "value rarely comes in form that 78 Kymlicka, Liberalism, Community, and Culture, 12-13. 79 Kymlicka, Liberalism, Community, and Culture, 12. 80 Kymlicka, Contemporary Political Philosophy: An Introduction, Second Edition, 216. 81 Thomas Hurka puts the point this way: "An activity makes a person's life go better when it is valuable, but its being valuable requires that it be endorsed," "Indirect Perfectionism: Kymlicka on Liberal Neutrality," The Journal of Political Philosophy, Volume 3, Number 1, March 1995, 36-57, 40, fn. 12. 49 can be injected into a person."82 Margaret Moore elaborates on this idea in the following way: Objective values, such as love or friendship or beauty, cannot merely be dispensed to people, X units for each person: the person must feel love or affection to realize the objective value of a fulfilling relationship or appreciate art to realize the objective value of beauty...it is important that the person feel a subjective commitment to these values in order to live a good life.83 So, if Kymlicka is committed to the endorsement constraint, why does this mean he cannot add a normative constraint condition? The normative constraint condition says, "Here are the kinds of preferences and beliefs people must avoid holding in order to qualify as autonomous, regardless of whether they see the value or disvalue in these preferences and beliefs for themselves." In other words, by identifying the "wrong" preferences, the normative constraint condition in effect "dispenses" preferences and beliefs to people. But, on Kymlicka's view, if people have their values "dispensed" to them from an external perspective, i.e. the normative constraint condition, then they are leading their lives from the "outside." If people are leading their lives from the outside, they are not leading them from the inside, according to values they deem important from their own perspective. If people are not leading their lives from the inside, they are not living lives that are good. This is because leading one's life from the inside is a necessary condition for leading a life that is good. So, this is why Kymlicka cannot add a normative constraint condition to his current theory of autonomy: it is incompatible with a key feature of his liberalism, i.e. the endorsement constraint. A normative constraint condition "dispenses" values to people, but Kymlicka denies that individuals lead lives 82 Kymlicka, Liberalism, Community, and Culture, 19, fn. 2. 83 Margaret Moore, "Liberalism and the Good Life," The Review of Politics, Vol. 53, No. 4 (Autumn 1991), 672-690, 677, her emphasis. 50 that are good if they have values "dispensed" to them. If the normative constraint condition is incompatible with the endorsement constraint, Kymlicka cannot add this condition to his current theory of autonomy. As such, Kymlicka's view of autonomy cannot be strong substantivist. Weak Substantive Autonomy Recently, some autonomy theorists have argued for what they refer to as a weak substantive view of autonomy. This view is situated between strong substantivism and proceduralism. Like proceduralists, weak substantivists claim that a person must critically reflect in the proper way in order to count as autonomous. I'll say more shortly about why weak substantivist theorists think a reasonable account of autonomy must go beyond proceduralist standards. But, for now, let's note that defenders of weak substantivism claim further that a necessary condition of autonomy is that an autonomous agent holds the right kind self-regarding attitudes toward herself, e.g. a sense of self-worth as an agent.84 Having the right kind of attitudes ensures that an agent critically reflects in the right way, for these attitudes ensure that an agent is capable of identifying the reasons for her actions.85 Notice, however, that weak substantivist theorists are not saying that these attitudes are causally necessary for autonomy; rather, they are claiming that a defining 84 Benson argues that autonomous agents must have normative competence: they must "recognize and appreciate various norms that apply to their actions," ("Feminist Intuitions and the Normative Substance of Autonomy," 133). 85 As Benson puts it, the "normative standards for agents' authority to construct and potentially answer for their reasons for acting enter into autonomy by way of the attitudes toward their own competence and worth through which agents claim such authority," ("Feminist Intuitions and the Normative Substance of Autonomy," 136.) Sonya Charles makes a similar point, noting that "Like procedural theories, weak substantive theorists rely mainly on an internal process of critical or self-reflection, but argue that the procedure will ‘work' (yield an autonomous decision) only if the person has the proper ‘self-trust' (or other self-reflexive attitudes,)" (411). See "How Should Feminist Autonomy Theorists Respond to the Problem of Internalized Oppression?" Social Theory and Practice 36 (3), July 2010, 409-428. 51 feature of autonomy is that an agent has a sense of self-worth. If a person lacks this attitude, then she is not autonomous, no matter how carefully she has critically reflected. So, like strong substantivism, weak substantivism claims there are external conditions (that is, conditions beyond an agent's psychology) that must obtain in order for a person to count as autonomous (in this case, these conditions refer to facts about an agent's competence.) But, unlike strong substantivism, weak substantivism doesn't place direct restrictions on the content of an agent's value commitments. This means that individuals can, in principle, endorse beliefs with any content, including hierarchical or oppressive value commitments like Jane's, provided they are competent reasoners and exhibit the right kinds of self-regarding attitudes. So, to the "content" question, weak substantivist theorists would answer "no, but..." According to its defenders, weak substantive accounts are attractive for the following reasons. On the one hand, these views allow for greater leeway than strong substantive accounts with respect to the kinds of value commitments an autonomous agent can endorse, including hierarchical or oppressive value commitments. This is important because some individuals genuinely hold such values because while many of us may think such value commitments are morally problematic, weak substantive theorists share with proceduralists the intuition that autonomy requires a large space of noninterference. Individuals must have the practical and theoretical space to identify and pursue their good, even according to values others might regard as morally wrong or repugnant. On the other hand, defenders claim that weak substantive accounts are preferable to proceduralist views because their standards incorporate some substantive content. Put 52 another way, defenders deny that an account of autonomy that incorporates purely proceduralist standards is sufficient. To understand why, it will be helpful at this point to introduce another standard objection against proceduralism. As we have seen, because such views maintain that reflective competence and the right kind of psychological attitudes are necessary and sufficient for autonomy, there is the possibility that deferential wives and slaves will turn out to be autonomous. This strikes some critics as deeply problematic, but the issue in this case is not the content of the preferences. (Recall that this is a concern for strong substantivists, but not for weak substantivist theorists.) The worry here is that it seems reasonable to think that oppressive socialization interferes with a person's capacity for autonomy, so it seems counterintuitive to say that Jane or the slave is autonomous. To be clear, the point here isn't that Jane's social training renders her incapable of critical reasoning. Rather, the concern is that Jane has been socially trained not only to value, but to want to value, subordination. Because Jane's social training has led her to internalize so deeply ideas about the "rightness" of female subordination, her critical endorsement of her values as "her own" may not be indicative of her autonomy as proceduralists claim. In other words, critics have doubts that proceduralist standards alone picks out autonomous agents. Procedural theorists acknowledge that it is possible for certain factors to interfere with a person's critical faculties, such as one's social training, hypnosis, brainwashing, or manipulation by evil neuroscientists, in which case a person would not count as autonomous.86 But, procedural theorists also maintain that it is possible social training will not interfere with a person's reflective capacities, even if she has been socialized to 86 John Christman, "Autonomy, History, and the Subject of Justice," Social Theory and Practice, 33 (1), January 2007, 1-26, 6. 53 value subordination. We can't assume that a person who wants to choose according to subordinate value commitments has compromised critical capacities as a result of her social training. Furthermore, because it is possible for a person to genuinely prefer subordination, a reasonable view of autonomy will make conceptual room for this possibility by positing only minimal standards for autonomy. What is important, then, is to identify an account of critical reflection fine-grained enough to pick out an agent whom it is reasonable to think endorses hierarchical value commitments as her "own" and yet can provide an effective filter to isolate agents who hold similar commitments but of whom it is reasonable to think does not hold those commitments as her own. In response, critics might say the following to the procedural theorist: "You can refine the critical reflective process and introduce further constraints on the process to make it as fine-grained as you wish. But, this move does not address the worry that critics raise. The problem isn't that proceduralists have yet to identify the right critical reflection procedure; rather, the problem is that because the proceduralist standards as such are so minimal, there are people who meet them that seem, in an obvious, intuitive way to not be autonomous (e.g. they defer to their husbands in all decisions). For example, consider on Christman's view that a person must critically reflect in the right way on her value commitments and she must not feel a sense of alienation toward them. Suppose that Jane reflects carefully and thoughtfully upon the historical processes by which her value commitment to subordination arose. Suppose further that she would not resist these processes, even if someone tells her that she has been socially trained to hold value commitments that maintain oppressive relations of domination and subordination between men and women. But, here's the problem. When Jane is reflecting on her value 54 commitment to subordination, she would likely think, "I should be subject to the kind of social training that led me to have this value commitment. This is the way things should be. So, I wouldn't resist the processes by which my commitment to subordination arose and I don't feel alienated toward it." Because Jane has been trained to want to value subordination and accept the "rightness" of gender hierarchy, it is unlikely she would resist the historical process or that she would feel a sense of alienation toward her commitment. So, critics conclude, we can't be sure whether Jane's critical reflections secure her autonomy or whether she is simply reiterating the terms of her subordination. According to defenders of weak substantivism, however, the introduction of self-regarding attitudes as necessary for autonomy gets around this problem raised by proceduralist views. On the one hand, weak substantivist theorists agree with proceduralists that we need to make room for the possibility of agents who endorse hierarchical or subordinate value commitments and also qualify as autonomous. But, in order to be satisfied that such agents are actually autonomous with respect to those commitments, we need to introduce conditions for autonomy beyond an agent's internal psychological states. A person has to have the right self-regarding attitude toward herself, which of course is a psychological state, but it is an objective feature of the world that she actually has this attitude. According to weak substantivists, the introduction of self-regarding attitudes provides an effective way to filter out agents who endorse hierarchical beliefs from agents whose endorsement of the same kinds of beliefs may not be genuine. If a person has a sense of her self-worth as an agent, that is, if she regards herself as 55 worthy of being "the author of her own conduct,"87 then she is autonomous (provided she meets the further reflective competence condition as well.) Weak Substantive Autonomy: A New Way? While proponents are confident that weak substantive views represent the way forward beyond the impasse between procedural and strong substantive views over the "content" question, I am less certain. Here is my strategy in this section. To make my arguments against weak substantivism, I consider two of the so-called "hard cases" for autonomy. We've already discussed Jane and her choice to subordinate herself to male authority for religious reasons. Another kind of hard case is one in which a woman chooses to engage in (what can be construed as) a practice of gender oppression. For example, suppose Fatima chooses to wear a full face and body veil for (what she claims to be) religious reasons and as a marker of social and cultural identity. However, the Islamic practice of veiling has traditionally been (and perhaps continues to be) employed to control female sexuality. So, choosing to wear a veil is a hard case for autonomy theorists because an individual makes a choice on the basis of values linked to gender oppression and it seems reasonable to think that oppressive values have the potential to impair a person's capacity for autonomy (even if it's not clear how this impairment happens.)88 However, even if we grant that this is an oppressive practice, some autonomy theorists insist that we cannot conclude that choosing to engage in these practices 87 Paul Benson, "Free Agency and Self-Worth," Journal of Philosophy 1994 (91): 650-668, 659. 88 Stoljar, "Feminist Perspectives on Autonomy," Section III, Stanford Encyclopedia of Philosophy, (forthcoming). 56 diminishes a person's autonomy.89 As we have seen, proceduralists would say that Jane is (or would be) autonomous provided she is (or would be) reflectively competent and has (or could have) the right psychological attitudes, e.g. nonalienation, and I think they would make similar evaluations about Fatima, provided that she would meet the conditions for autonomy. In contrast, strong substantivist theorists deny that agents who hold the "wrong" preference or make the "wrong" choices cannot be autonomous, even if they are reflectively competent. Furthermore, I take it that strong substantivist theorists would regard Jane's preference to submit to male authority and Fatima's preference to wear a full veil as the "wrong" kinds of preferences. They are the "wrong" kind of preferences because their content leads Jane and Fatima to make a serious moral error: they fail to regard and treat themselves as full moral equals. Put another way, the content of their preferences prevents their psychologies from "hooking up" with or corresponding to an objective feature about the world, i.e. that they are moral equals. Because they make this kind of moral error, strong substantivists would deem Jane and Fatima as nonautonomous. Let's turn now to considering weak substantivism, because (as I have noted above) defenders of the view see themselves as staking a path between proceduralism and strong substantivism. However, it is my contention that weak substantivist views fail to provide a "third" way between proceduralist and strong substantivist views. First, I will argue that weak substantive standards fail to provide an effective filter to pick out those agents who genuinely hold hierarchical beliefs from agents whose support of the same kinds of beliefs may not be genuine. As such, weak substantivism is subject to the same 89 For a similar discussion, see Stoljar, "Feminist Perspectives on Autonomy," Section II, Stanford Encyclopedia of Philosophy, (forthcoming). 57 objection raised against proceduralism. Second, I will argue that weak substantivist standards must incorporate stronger normative substance in order to avoid this objection. But, I will show that this runs the risk of collapsing weak substantive views into strong substantivism. For, once we ramp up weak substantivist requirements, these views and strong substantive views will assess the very same cases as either autonomous or nonautonomous. If my argument is right, this suggests that weak substantive views aren't doing what they are supposed to do, namely making room for a wider range of value commitments that agents can endorse, and still qualify as autonomous, including hierarchical beliefs. If so, we have no reason to prefer weak substantive views over strong substantive views. To begin to make my case, I appeal to Paul Benson's weak substantive account.90 On his account, a person is autonomous when she has proper regard for her authority as an agent. He states that, "occupying a position of authority to speak for one's intentions and acts seems to depend not only on one's objective fitness to play the role of potential answerer, but also on one's regard for one's abilities and social position."91 I'll consider the latter condition first. For a person to have regard for her abilities and social position, she must hold the right kinds of attitudes toward her "competence and worth through which [she] claim[s] such authority."92 Benson explains this requires individuals to have "no serious doubts about their competence to recognize or construct reasons for their actions or about their 90 Benson, "Feminist Intuitions and the Normative Substance of Autonomy," 128-137 and "Taking Ownership: Authority and Voice in Autonomous Agency," 106-117. 91 Benson, "Taking Ownership: Authority and Voice in Autonomous Agency," 110-111. 92 Benson, "Feminist Intuitions and the Normative Substance of Autonomy," 136. 58 authority to speak and answer for their conduct, should others criticize it."93 Moreover, individuals must not have "attitudes [which] manifest marked disengagement or dissociation from their conduct."94 Positively, individuals must hold "reflexive attitudes toward their own agency [which] indicate that they really do take agential ownership of their decisions" and must "treat themselves as fit and worthy to identify adequate grounds for their decisions, to translate those decisions appropriately into conduct, and to answer for themselves should others challenge their reasons."95 In other words, autonomous individuals are those who can offer to others reasons for their actions, which reasons-giving reflects their attitude toward their competence and worth as moral agents. Notice, then, on Benson's weak substantive view, normative authority doesn't obtain when an agent endorses her value commitments or determines whether or not she feels alienated toward them, but rather, when she authorizes her agency.96 Individuals, therefore, lack autonomy because they fail to have the right self-regarding attitudes toward themselves as competent and worthy and so they regard themselves as unworthy to offer reasons for their conduct. To illustrate his point, Benson appeals to Ralph Ellison's novel The Invisible Man, in which the narrator has internalized his "social invisibility" as an African-American man in the Jim Crow era. The protagonist cannot answer for his actions not only because his social conditions prevent him from occupying a position of 93 Benson, "Feminist Intuitions and the Normative Substance of Autonomy," 128. 94 Ibid. 95 Benson, "Feminist Intuitions and the Normative Substance of Autonomy," 129. 96 Benson, "Taking Ownership: Authority and Voice in Autonomous Agency," 107. 59 agential authority, he is unable to regard himself as a competent agent due to internalizing his invisibility.97 Let's now consider the first condition Benson identifies, namely that occupying a position of agential authority depends upon one's "objective fitness" to fulfil that role. Here, I take Benson to mean that whether one is capable of fulfilling the role of answerer depends not on whether a person thinks she is competent but rather on whether she really is competent. So, whether one can take ownership of one's actions partly depends upon external features of the world, so it is a social or relational condition for autonomy.98 According to Benson, "to have the authority<