Both in ethics and in political philosophy, autonomy is often seen as a major value, to be promoted and/or respected both in our interpersonal relationships and in political decisions. A concern with autonomy can ground requirements of informed consent and antipaternalistic strictures on government policies, and has been proposed as a foundation for democratic rule. However, it is not self-evident that autonomy is valuable, and in his new book Ethical Autonomy, Lucas Swaine provides a thought-provoking and original criticism of the value of autonomy, as that concept is often understood. His criticism centers on the claim that the idea of personal autonomy does not contain sufficient limits on the possibility that the autonomous person is capable of performing deeply immoral actions. Building on this criticism, Swaine proposes a theory of what he calls “ethical autonomy,” which is meant to avoid the weaknesses he identifies in traditional understandings of personal autonomy.

The book consists of four chapters. The first and shortest chapter provides a history of the concept of autonomy from antiquity to its role in contemporary democracies. Of particular interest is the author's demonstration, through detailed textual analysis, that it is wrong to claim, as is sometimes done, that in classical antiquity, the term “autonomia” was used only of collective entities, such as states, and not of individuals.

The second chapter lays the groundwork for the book's main argument by elaborating how the concept of personal autonomy, that is, the autonomy of individual persons, can be understood. Swaine views personal autonomy as involving “individual self-rule,” understood as “the ordering of a human person, by that person” (41), and he describes the criteria that must be satisfied in order for a person to count as autonomous in this sense: the agent must engage in critical reflection on her own beliefs and motives, possess sufficient willpower as well as a number of basic capabilities, and have sufficient options for acting. While one might argue with the details of Swaine's account, the picture he paints of the autonomous person resonates well with standard understandings of autonomy in the literature, and I take it that this is intended. Indeed, Swaine's criticism of personal autonomy would be largely irrelevant if he built his criticism on an idiosyncratic understanding of autonomy.

It is in chapter 3, where Swaine criticizes ideals of personal autonomy, that the book's most original ideas are presented. Central to Swaine's criticism is the notion of an “extreme action,” described as “acts that one would correctly describe as deeply morally wrong” (108), with rape, murder, and torture being core examples. Swaine's point is not that personal autonomy allows the performance of such acts—indeed, ideals of autonomy are perfectly compatible with constraints on actions—but that personal autonomy does not involve sufficient constraints against making such actions possible. Swaine elaborates on this point by discussing how the main criteria of autonomy fail to exclude extreme actions. Thus, as part of the task of critical reflection, the autonomous person may reflect on or sympathetically imagine what it would feel like to perform an extreme action, such as torture. Similarly, the ideal of having sufficient willpower to not be hindered from acting by weakness of will may entail developing the willpower needed to perform extreme actions. In one sense, the person capable of choosing to perform an extreme action and then of following through and actually performing that action is more autonomous than the person who is incapable of performing such an action, for lack of willpower, other capabilities, or relevant options. Yet, Swaine argues, developing this capacity is morally problematic both for instrumental reasons—if a person is capable of choosing and performing extreme actions, it is more probable that she will actually perform such an action—and for noninstrumental reasons, as the capability of performing evil is “inherently iniquitous” (133).

The failure of theories of personal autonomy to condemn the active consideration of performing extreme actions or the formation of the capacity to perform such actions is, on Swaine's view, a major problem for such theories. In chapter 4, he therefore proposes an alternative account of autonomy, which he calls “ethical autonomy.” Central to this account is the notion of moral character, which refers to dispositions to behave in morally good ways. Ethical autonomy, Swaine writes, is “individual self-rule modulated by moral character” (159); it thus keeps the main elements of personal autonomy, but the addition of moral character changes the nature of these elements. This explains how ethical autonomy is meant to avoid the pitfalls of personal autonomy. For example, for the person of strong moral character, “there are some things she is simply not willing seriously to consider doing” (164). In that way, the ideal of ethical autonomy, unlike that of personal autonomy, bars an autonomous agent from considering or sympathetically imagining the performance of an extreme action like rape.

The challenge that Swaine raises for personal autonomy is original and important and must be taken seriously by any proponent of personal autonomy. Some objections can, however, be raised both against Swaine's attack on personal autonomy and against his proposed alternative. One objection (which Swaine himself also considers in the book) is that it is not obvious that it is the task of a theory of autonomy to set normative constraints on what an agent may consider doing or make herself capable of doing. Even if we grant that there should be such constraints, the source of these constraints may be found in other values than in the value of autonomy. A more general objection is that it is not always clear exactly which writers or theories are the target of Swaine's criticisms. In his criticisms, Swaine sometimes refers simply to “personal autonomy” or “theories of personal autonomy.” However, since different theories may differ as to why they take autonomy to be valuable and which implications they draw from the value of autonomy, it would be helpful if Swaine said more about which theories he takes to be vulnerable to his criticism.

When it comes to Swaine's positive contribution, I would have liked to see him flesh out in more detail what his theory of ethical autonomy amounts to, giving the reader better grounds for assessing the various claims he makes about the strengths of his account. Indeed, it is not always clear that the different advantages Swaine claims for ethical autonomy cannot also be claimed by a proponent of personal autonomy. For example, Swaine argues that ethical autonomy “coheres with an appreciation of the normative importance [of] overarching coercive frameworks and the institution of functioning legal systems” (180). It seems to me that proponents of personal autonomy too may recognize the need for a coercive legal framework, in order to protect everyone's autonomy equally.

That said, the fact that some points may be disputed or require further elaboration is not a criticism of the book, but rather a testimony to the fact that Ethical Autonomy makes an important contribution to the ongoing discussion of the role of autonomy in society.