Discussions of personal autonomy often proceed as if the free-will problem does not exist. Yet an incompatibilist – one who regards determinism as a genuine threat to free will – may wish to argue that an account of autonomy or self-government is severely compromised by the discovery that the self in question is a deterministic product of heredity and environment. Such a discovery would entail, in her eyes, that, even if I am judged autonomous through, say, a capacity for uncoerced and rational review of my deepest commitments (plus the ability to make appropriate adjustments), the failure in a deterministic world to control the origination of my desires and values, elements that explain my commitments, renders my self-governance seriously inauthentic.

Yet compartmentalization is an essential tool for the avoidance of intellectual paralysis. Perhaps then we ought not to worry simultaneously about both the conditions of autonomous decision making and the possibility that determinism will render our results a sham. If autonomy is our concern, we can let the other philosophical fellow raise the specter of enslavement of self arising from the domination exerted by heredity and environment over the elements that manifest our autonomy. Although no one can object to this pillar of intellectual practice, I would like to try to accommodate that peculiarly philosophical mindset that drives us to excess, that demand to be in good faith that, we suppose, fails to be met by a shallow theory that refuses to face the implications of our immersion in a controlling, hostile world.

INTRODUCTION

Autonomy, we suppose, is self-regulation or self-direction. There is a distinct idea that is easily confused with self-direction, namely, selfexpression, self-fulfillment, or self-realization. (I do not mean to suggest that the latter three terms are all synonymous. But in this essay, whatever differences there are among them play no role, so I will use them interchangeably.) Although it will turn out paradoxically that autonomy is neither self-regulation nor self-realization, it is reasonable to suppose that the former is a superior candidate. My teacher of Indian religion, Dr. Subodh Roy, blind from birth, chose not to undergo an operation that would have made him sighted because he believed, perhaps rightly, that the ability to see would interfere with his religious quest. He thereby chose not to realize one of his fundamental human capacities, one whose cultivation has produced some of the finest fruits of civilization. Joseph Raz describes a case in which a man places his life in jeopardy by undertaking a trip to deliver medical aid to a group of people in a distant place. Since he will be unable to secure food for several days, he, in effect, subordinates one of his own basic needs or interests to a goal that he deems more important. There is no reason to believe that, in refusing to express or realize a dimension of self, either Dr. Roy or Raz's philanthropist have failed to act autonomously.

Having pressed for a value-neutral account of freedom, I cannot but be impressed by the insistent demand for a central role for the notion of value or evaluation, if not in the analysis of freedom, then perhaps in that of autonomy of choice. (By definition, scope of autonomy implicates values, but that is not the notion we are currently trying to understand.) Even relativistic interpretations inevitably require that free or autonomous agents be essentially concerned to realize their values, no matter how silly, odious, or bizarre. There is no way to deal satisfactorily with this deeply ingrained view without a careful attempt to understand the nature of value and evaluation.

Values and desires

Even if we are sympathetic with the notion of objective value, it is clear that a person can have values that are not aligned with objective value; a person can be immoral, for example. Our project, then, concerns the difference between an individual's desires and her values.

We might describe this task as a search for the factor that forces a distinction between the two notions. For example, desires come in varying strengths. But since we certainly want to accommodate the thought that one state of affairs or goal is more highly valued or prized than another, we are not forced to distinguish and do not yet comprehend the difference between a strong desire for some goal and the placing of great value on that goal.

Advocates of valuational freedom are reacting against the idea that ascriptions of freedom to a human being are just ascriptions of the capacity to engage in unfettered or unimpeded activity or the absence of sufficient causes. Freedom has a moral as well as a metaphysical dimension. If we ignore the former, there would be no difference between the freedom of a human agent and the freedom of a fly or an electron. Freedom is that state which enables its bearer to be guided through life by values; it permits the agent to embody in action moral and nonmoral values, thereby sanctioning third party judgments of the agent's worth and responsibility for those actions. Although the absence of impediments or the availability of adequate opportunity may be a necessary condition for any being to achieve this state, there are also conditions internal to the agent, including those which set creatures with values apart from others. Although I upheld the centrality of a nonvaluational conception of freedom of action in the last chapter, we will have to look more closely at the claims made on behalf of this normative component for the case of free agency.

This project – the definition of free agency – is complicated by the fact that the same internal change can either enhance or diminish freedom depending upon the environment and the same external change can either enhance or diminish freedom depending upon internal capacities.

This book is about autonomy. In the next three chapters a conception of freedom will be defended according to which autonomy is different from, but dependent upon, freedom. I want first to summarize, without argument or elaboration, the results about freedom so that I can contrast freedom with autonomy and thereby introduce the basic ideas about the latter which guide my thinking.

Positive freedom and autonomy

Positive freedom is comprised of a set of personal traits which are essential or highly useful to the satisfaction of a wide range of activities and decisions, both short- and long-term. It encompasses relevant knowledge, including self-knowledge, and a variety of intellectual and physical competences. (Among intellectual competences, I include capacities for memory, perception, calculation, reasoning, information processing, and the elimination of irrational and inconsistent belief sets.) A free agent is free of debilitating physical infirmities, addictions, and severe stress. (I count a certain level of emotional health as a separate condition of autonomy. The details are presented in Chapters 8 and 9. Personal integrity, including the absence of conditions like severe psychosis, multiple personality, and split brains, is regarded as a condition of positive freedom rather than one of its elements.) Free agents may or may not possess the freedom to do and decide to do what is (deeply) important to them.

Most people believe they have a right to be autonomous, at least in many domains. The value of autonomy may then be construed as the value of conferring or possessing this right – the right to be a certain way and the right not to be interfered with as one cultivates this state – and that depends on the precise nature and scope of this right. The right will not be of much value to a person who finds that he only has a right to exercise autonomy over matters of little moment to him. I do not propose to address the scope of this right for such a discussion would have to rest on a general theory of rights I am not in a position to offer.

Instead, I wish rather to seek to know the value of that state, autonomy, to which one may or may not have a limited or unlimited right. Here, again, the question may be divided. For one may wish to know the value of making a decision autonomously rather than heteronomously, regardless of the specific content of the decision. Or one may want to know the value of incorporating a particular domain into the scope of one's autonomy.

In a sense, we are all free to run four-minute miles. There is no law against it; we all have access to tracks, shorts, and running shoes; we will be neither ostracized nor beaten by our neighbors in the event that we are about to succeed. Insofar as no one is interfering with us, we possess negative freedom. As desirable as this condition is, its shortcomings as a complete account of freedom are conspicuous when we consider how unfair it would be to blame an octogenarian for her failure to reach this goal in spite of her best efforts. In another sense of freedom, positive freedom, the deficiency is internal. In virtue of her inability to run a four-minute mile, she lacks positive freedom to do so and is consequently blameless. With respect to our context, she lacks an essential tool for a display of autonomy as proficiency. We turn first to the idea of ability that is appealed to here.

Ability

Let us begin inside the skin. In one important sense of “free,” to be free to do something, one must be able to do it. Since differences of ability have a lot to do with internal differences among human beings, we may try to characterize ability as an internal state.

Acts of God are not acts of God. They are sudden, unexpected, natural occurrences with a significant impact on human affairs. It is, therefore, not incoherent to contend, as I do, that autonomous agents are not individuals whose actions and decisions are self-directed. I hasten to add that this negative conclusion about autonomy is not just fallout from the rejection of the very project of theorizing about the self. It is rather a result which is forced on us by the convergence of a variety of considerations, including powerful intuitions about particular cases.

Taking etymology seriously, we begin, like everyone else, with the assumption that autonomy is self-direction and embark on a search for the difference between self and other. When William Blier, a member of David Koresh's cult, acquiesced in his own suicide, he was acting from motives, values, and beliefs that were inculcated from without in a way which made him a virtual paradigm of heteronomy. Although Blier's action was inspired by motives that were in a sense his, that were among other things accessible to him in the unique way any person knows her own mental states, we rightly hesitate to describe his behavior as self-directed. His submission to another person – Koresh – was as complete as is possible for a human creature.

The picture of the autonomous agent that has evolved so far incorporates the following elements: (1) Freedom: an autonomous agent possesses a variety of intellectual and physical skills and capacities which confer upon him the further capacity to assess his options competently. His autonomy is enhanced by a strong rational will. (2) Values: an autonomous agent has and acts on values, but it is not necessary that these values be understood as the Evaluative Theory demands. That is, he is not necessarily disposed to evaluate his desires and options from a moral point of view or from the point view of the worth of his motivations. (3) Rationality: an autonomous agent is basically rational. He possesses rational powers and is generally disposed to use these powers in decision making, even though he may retain and occasionally act on the power to flout rational principles. (4) Independence: the procedures and principles of decision making adopted by an autonomous agent are held by that agent on the grounds of their reliability in the process of decision making. Regardless of origin, they are subject to evaluation in terms of criteria that are as objective as possible. Although an autonomous agent may enter relationships which entail (self-imposed) limitations, he does not thereby abandon either his ultimate capacity for or his right to rational review.

We turn finally to the role of values in rational decision making.

It is well known that theorists disagree on the principles a rational agent must invoke when making decisions under uncertainty. One might try to assimilate such situations to ones of decision making under risk and recommend the maximization of expected utility, or one might invoke minimax, maximax, Hurwicz's variant of the latter, or the strategy of minimizing maximum regret. It is commonly supposed that reason per se does not dictate this decision, in which case rational agents are obliged to adopt principles grounded extrarationally. If some such principles are associated with values, then rational agents, all of whom must confront decision making under uncertainty at one time or another, must have values, although these values cannot be derived from the assumption of rationality.

Now the term “value” is used so widely that even these general methodological principles may be said to reflect values. The maximin rule expresses the value of conservatism or caution, the maximax rule that of optimism. If the term “value” is to be stretched this far, it is indeed impossible for any but the most seriously disturbed psychotics to proceed through life without values. The thesis we would then be examining would be reduced to triviality.

As we have just seen, the case of the addict or compulsive initiates a quest for an internal difference between the genuine autonomy of the normal person and the heteronomy of the addict. But there are features of an addict's life having nothing to do with the origin of his addiction which can explain his heteronomy. First of all, an unwilling addict lacks proficiency, the capacity to control his environment in accordance with his desires. However he came to this state, the loss of proficiency is grounded in his current condition, independently of origin. He is a cripple even if he freely chose to become one. More likely, the evolution to addiction was not initiated by a free choice. A fully informed, rational adult would not choose to become an addict, especially since one item of information at his disposal is that he would come seriously to regret the decision later. But here again the heteronomy of the original decision has to do with contemporaneous features of the decision, namely, ignorance or irrationality.

But we cannot just ignore the etiological component. People become addicts out of hopelessness, the need to become oblivious to their desperate condition, the desire to be free of physical and emotional pain, and the anticipation of intensely satisfying experiences.