![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
In this paper I propose and develop a social account of global autonomy. On this view, a person is autonomous simply to the extent to which it is difficult for others to subject her to their wills. I argue that many properties commonly thought necessary for autonomy are in fact properties that tend to increase an agent's immunity to such interpersonal subjection, and that the proposed account is therefore capable of providing theoretical unity to many of the otherwise heterogeneous requirements of autonomy familiar from recent discussions. Specifically, I discuss three such requirements: (i) possession of legally protected status, (ii) a sense of one's own self-worth, and (iii) a capacity for critical reflection. I argue that the proposed account is not only theoretically satisfying but also yields a rich and attractive conception of autonomy.
Notes
1 Note that the issue here is the extent to which autonomy is constitutively social; a view may be purely nonsocial in this sense despite including a thoroughly social account of autonomy attainment, as do a number of these accounts.
2 The following analogy may be helpful. A sedative is (at least in part) something that tends to cause drowsiness. This is a psychological property. But it is one that is typically realized by some base chemical property, such as that of being an opiate or a barbiturate or an alcohol, which is itself nonpsychological. In the same way, I claim that the concept of autonomy is best analysed in terms of the social property of immunity to interpersonal subjection. But this is a disposition that is in turn realized by one or more base properties, such as that of having a capacity for critical rationality or a sense of one's own self-worth, that may themselves be nonsocial.
3 For instance, Westlund [2003: 515] holds that a failure of dialogical critical reflection constitutes a failure of autonomy by way of constituting a loss of the ‘distinct and determinate self’.
4 This may seem to render the account viciously circular, but it merely makes it recursive. The impossibility of A's having already performed an infinite number of actions means that the chain of analysis will not run for ever: eventually we will reach either an action that was not intentionally influenced by another, or intentional influence that was not endorsed.
5 Thanks to an anonymous referee for helping me to see this (by means of the following example).
6 Similarly, the highly unmanipulable person who gets unlucky and finds herself subject to a foreign will on some specific isolated occasion may nevertheless be highly autonomous.
7 Note also that, whereas part of the job of an account of local autonomy is to explain the normative authority of choices typically made by competent adults, there is no similar reason why levels of global autonomy cannot be ascribed to animals and infants. Indeed, it is not unusual for people to say such things as that some intransigent toddler is attempting to assert her autonomy, or that cats tend to be more autonomous than dogs. These sorts of extended global autonomy claims are easily accommodated by the proposed account (and are wholly mysterious on most existing accounts).
8 On the distinction between autonomy and the enjoyment of options, see Dworkin Citation1988: 14, Feinberg Citation1989: 62–8, and Taylor Citation2009: 108–9. Note that options may be restricted by ‘internal’ as well as by ‘external’ obstacles: in either case an agent might remain autonomous whilst nevertheless lacking freedom. Moreover, while Joseph Raz [1986: 372] famously includes ‘an adequate range of options’ as a condition of autonomy, this disagreement is largely terminological: whereas Raz uses ‘autonomy’ broadly to include the enjoyment of options, the proposed account uses ‘autonomy’ narrowly to refer just to something like his ‘independence’ condition (while preferring ‘freedom’ as the broader term). For some argument in favour of preferring this latter terminological framework, see Garnett [2013].
9 Indeed, the account allows that choices in favour of reduced global autonomy may themselves be autonomous. Thus a Western woman may exercise perfect (local) autonomy in choosing to marry a Saudi and relocate to his homeland, despite the fact that this is, given the social and legal status of women in Saudi Arabia, a choice in favour of reduced (global) autonomy. In so far as we are bound to respect one another's autonomous choices, we are bound to respect this woman's choice of life, even though the life she autonomously chooses is one of relative nonautonomy. (For this reason the present account is not vulnerable to the criticisms levelled by Christman [2004], which are better construed as objections to social accounts of local autonomy. See also Westlund [2009].)
10 Self-worth is therefore here understood as a type of self-confidence or self-trust. It is more specific than self-esteem, which involves a more comprehensive self-assessment [Sachs Citation1981]: one may judge oneself to be a bad person in various respects without doubting one's competence as an agent. It is also distinct from self-respect, understood as knowledge of and appropriate concern for one's own moral standing [Hill Jr. 1973]. There is no conceptual reason why a person cannot fail to claim his own moral entitlements while nevertheless regarding himself as a perfectly competent agent. This said, however, there are no doubt many reasons why self-worth, self-esteem and self-respect should tend to wax and wane together.
11 Strictly speaking, the more important autonomy trait is critical reflection itself; possession of the capacity is an autonomy trait only in so far as it raises the likelihood of critical reflection actually occurring.
12 Dworkin [1976] and Taylor [2009] are also, I believe, somewhat vulnerable to this complaint. Raz, who attempts to derive each of his requirements from the single idea of ‘being the author of one's own life’ [1986: 374], may be less so—though it is doubtful whether appeals to the notion of ‘authorship’ are in the end much more helpful than direct appeals to that of autonomy itself.
13 Note that for each property thought to be important for autonomy, the pure social theorist must establish either that it is indeed an autonomy trait, or else that there are independent and principled reasons for excluding it from our account of autonomy. Take, for instance, identification with one's effective motivations. It may well turn out that, in general, it is easier to push around a disunited and conflicted agent than to push around an integrated and wholehearted one, and that identification is therefore an autonomy trait. If it does not, however, then the social theorist will need to join with others such as Dworkin [1988] and Taylor [2009] in providing principled reasons for denying that such identification is necessary for autonomy (see also, in this connection, n. 8, and Garnett [2013]).
14 An earlier version of this paper was presented at the 2011 New Scholars in Bioethics (NSIB) First Annual Symposium, and I would like to thank the participants (Robyn Bluhm, Kirstin Borgerson, Danielle Bromwich, Joseph Millum and Marika Warren) for their valuable comments. Versions were also presented at a number of seminars at Birkbeck College, and I am grateful to colleagues (especially Miranda Fricker, Keith Hossack and Susan James) as well as to our postgraduate students for some very helpful discussions. Finally, I am especially indebted to three anonymous referees for this journal for their thoughtful and excellent critical comments and suggestions.