Robert Stern’s new collection of essays, Kantian Ethics: Value, Agency, and Obligation, serves as an excellent companion for his most recent monograph, Understanding Moral Obligation: Kant, Hegel, Kierkegaard (2012). Kantian Ethics has two main parts. Part I concentrates on Kant. Part II examines Kantian themes in the work of a wide range of figures, including Hegel, Bradley, Green, James, Darwall and the Danish philosopher K. E. Løgstrup. This range forces a reviewer to be ruthlessly selective. Given the forum for the present review, I shall focus on Stern’s treatment of Kant, taking the main topics in the following order: obligation, agency and value.

Moral obligation is the central topic and guiding thread of Kantian Ethics. Stern takes obligation to be the main focus of Kant’s moral philosophy (p. 1), and he argues that Kant’s account hinges on the contrast between the human will and the holy will. The moral law applies to both, but the holy will is characterized by its morally (and non-morally) good disposition, while the human will is not. This means that the holy will is necessarily disposed to do what is right and refrain from what is wrong. Kant goes so far as to say that the holy will’s ‘volition is of itself necessarily in accord with the law’ since ‘it can be determined only through the representation of the good’ (Groundwork, 4: 414, quoted on p. 17). The human will, by contrast, can be determined by other things, such as the thought of happiness or the prospect of sensible pleasure. In fact, human volition not only is not necessarily in accord with the law, it is positively disposed to act contrary to the law. For this reason, the human agent who grasps what the law requires experiences the law as a form of constraint or ‘necessitation’ (Nötigung). It imposes itself as an imperative. According to Stern, Kant’s view is that the fundamental nature of obligation consists in this ‘felt necessity of morality’ (p. 28). Interestingly, the claim here is that human beings are necessitated by the law precisely because we do not necessarily obey it. On one interpretation, the idea is that being obligated to do x implies that one might not. On Stern’s interpretation, being obligated to do x implies even more than this. It implies ‘not just an awareness of the existence of possible non-moral options, but also the thwarted desire to take them’ (p. 28). According to Stern, in trying to explain the felt necessity of moral obligation Kant is, among other things, trying to answer the transcendental question of how categorical imperatives are possible. Not surprisingly, his answer involves transcendental idealism, just as it does in the parallel case of trying to counter Hume’s scepticism about causal necessity (pp. 17–19, cf. 85–6).

This approach to obligation raises some concerns. One concern is that it does not address the possibility that the ‘felt necessity’ of morality results from upbringing or a quirk of the brain. After all, some people do not experience it, which might suggest that moral ‘necessitation’ is rooted in contingent psychological or sociological factors. Stern does not address this concern, but he dispatches two other objections to Kant’s account. The first objection is that the contrast between the holy and human will conflicts with Kant’s own view of the good will. The holy will cannot be necessitated, so it cannot act ‘out of duty’, which means that it does not qualify as a good will. In reply, Stern argues that Kant’s characterization of the good will pertains to the specific ‘subjective limitations and hindrances’ under which a human being can be good (Groundwork, 4: 397). This does not rule out other ways of being good. Given our ignorance of the divine, we certainly cannot conclude that a holy will cannot be good because it cannot be good in the way that we can (pp. 24–5). Stern also addresses the concern that the holy–human contrast is ‘symptomatic of the sort of dualistic picture which his critics in the idealist and romantic traditions imputed to Kant’ (p. 25). Such critics sought to fit obligation into a more harmonious or unified picture of the self. Stern argues that this alternative approach, despite its appeal, loses sight of Kant’s explanandum, namely, the ‘felt necessity’ of obligation. Many Kantians will find this reply persuasive, but it brings us back to a version of the first concern, which Stern does not address. Perhaps the moral law would not feel like an imposition if we were properly habituated or if our culture were better calibrated to the good.

Agency is Stern’s second main topic. He approaches it from a number of directions. Two essays explore the link between morality and agency by focusing on the claim that ‘ought implies can’ (OIC). Chapter 6 distinguishes between ‘strong’ and ‘weak’ versions of the claim and argues that Kant employs the latter but not the former. The strong version argues from empirical claims about human capacities to normative claims about the content of morality (and sometimes epistemology). The strong version thus uses OIC to set limits on the normative. Weak versions of OIC, by contrast, argue from normative claims to claims about human capacities. Stern examines a wide range of the passages where Kant employs OIC, and shows that these are all weak versions of the thesis (pp. 115–21). Kant begins with an a priori account of the moral law and then uses OIC ‘to determine what we are capable of qua human beings, insofar as we fall under this law’ (p. 121). His interpretation of Kant’s view is convincing, but it seems odd to label it ‘weak’. Surely it is the bolder and more audacious version of OIC since it claims to infer something about real human capacities from prescriptions of morality that are determined independently of any reflection on those capacities. For some people, this will seem no more plausible than the idea that we could invent a game and then infer from its rules that human beings are actually capable of playing it. Stern does not address this concern. How confident should Kantians be that empirically based doubts about freedom can be resisted with an OIC-based inference from a priori morality? Readers might also wish Stern would say more about the relationship between strong and weak uses of OIC, and about how Kantians should respond to the strong version he ascribes to James Griffin, who uses OIC to call for a more modest or ‘realistic’ moral philosophy. This minor quibble notwithstanding, Stern’s discussion makes an essential contribution to the literature on OIC.

In Chapter 7 Stern zooms out to examine OIC in the context of rival accounts of obligation. He argues that the plausibility of OIC depends on what obligation is. Stern examines four different kinds of accounts – Divine Command, Natural Law, Kantian and Social Command accounts – to show how well or poorly each is able to support OIC. Rather than defending any one of them, he carefully describes the trade-offs of all four. One significant result of his discussion is the claim that OIC is most strongly supported by the ‘appraisive’ rather than ‘directive’ function of moral obligation (pp. 132–5). His suggestion is that our deep commitment to OIC shows up in the way we morally blame people for failing to fulfil their obligations. OIC expresses a belief about the form of control that we use to distinguish between moral and non-moral forms of blame (and praise). When we blame a person for failing to fulfil a moral obligation, we ascribe an ability to control his actions. But when we have good reason to think such control was lacking, we tend to describe his failure as an instance of foolishness, imprudence, gracelessness or some other non-moral breach of conduct (p. 133). OIC thus underlies one way we distinguish the moral from the non-moral.

Chapter 12 approaches the topic of free will via comparison of Kant and William James. Though James disavows the influence of Kant, it is hard not to be struck by similarities between his ‘The Will to Believe’ and ‘The Dilemma of Determinism’ and Kant’s views about freedom and the ‘primacy of the practical’. Some readers think they have a kind of non-evidentialism in common. Stern agrees that James’s and Kant’s views are similar but he rejects the non-evidentialist reading. He claims that both philosophers subscribe to a ‘form of evidentialism [that] gives priority to practical reason’ (p. 217). The issue here is whether Kant and James think that we are entitled to believe that, for example, human beings are free even though we lack evidence for this belief. Stern argues that both philosophers point to a form of practical evidence that supports this belief. In James’s case, this evidence is supplied by feelings of regret experienced in the wake of misdeeds. Such feelings are part of the ‘moral rationality’ that, along with ‘mechanical’ and ‘logical’ rationality, directs our efforts to ‘cast the world in a more rational shape’ (James, quoted on pp. 212–13). Since it is rational to maximize the coherence of one’s system of beliefs (p. 214), and because the facts themselves are silent on the truth of determinism, we are entitled to the belief that fits best with feelings of regret – namely, the belief in indeterminism and free will. In Kant’s case, practical evidence is provided by the moral law, which is given to us as a ‘fact of reason’. One might think that Kant’s position is that the fact of reason entitles us to go beyond all the evidence. But Stern argues that Kant’s argument treats the fact of reason as a piece of evidence, which, since ‘ought implies can’, justifies the belief that one is free. After all, Kant calls the law the ratio cognoscendi of freedom (pp. 208–9). Kant’s use of OIC in this part of the Critique of Practical Reason – and at 4: 455–6 of the Groundwork (pp. 210–11) – suggests a ‘practical form of evidentialism’ (p. 211). For both James and Kant, the belief in freedom is warranted by a distinctly practical piece of evidence. Neither tries to establish that it is safely beyond rational justification.

Value is the third main topic of the book. Stern takes Kant to be a moral realist, and he argues for several claims about this realism. First, realism is a component of Kant’s ‘hybrid’ account of obligation. In a nutshell, this means that Kant is a realist about moral value and an anti-realist about obligation. Value does not depend on human attitudes or the practical point of view but obligation does. It would be wrong to let one’s talents rust, for example, even if no one recognized this fact. But if no one ever felt duty-bound to perfect himself, then there would be no obligation to do so. Obligation or duty is a function of the way morality appears to finite creatures such as ourselves – namely, as a law that imposes imperatives that command categorically. Understanding Moral Obligation presents this account in greater detail, but Chapter 1 of Kantian Ethics offers an excellent statement of the basic idea. Second, Stern argues that realism is compatible with autonomy; the latter does not require constructivism, as so many philosophers assume (Chapter 2). Third, he argues that Kant is not very troubled by forms of scepticism that allegedly require constructivism and that realism is well suited to meet the forms of scepticism that do concern Kant (Chapters 3–5). Fourth, Kant’s argument for the humanity formulation (FH) of the categorical imperative demonstrates a commitment to realism rather than constructivism (Chapter 5).

Debates about Kant’s meta-ethical commitments often focus on the argument for FH, so Chapter 5 is especially central to Stern’s view. In it, he makes a compelling case against Korsgaard’s influential account of FH. Stern asks and answers two main questions. First, does Korsgaard’s constructivist reading fit Kant’s text? Second, is the actual argument for FH motivated by anti-scepticism? Stern answers ‘no’ to both questions. According to Korsgaard, Kant provides a formal characterization of valuing and then shows how unconditional value is conferred on humanity by valuing as such. This amounts to a regress argument for the claim that ‘the value of humanity itself is implicit in every human choice’ (quoted on p. 94). According to Stern, Kant’s argument on 4: 427–9 of the Groundwork looks nothing like Korsgaard’s. Kant tries to show that categorical imperatives must be grounded in objective ends, and that the only thing capable of being an objective end is a rational being. His argument arrives at this conclusion by eliminating two rival candidates: objects of inclination and non-rational beings produced by nature (pp. 97–8). It builds on the assumption, ‘stat[ed] dogmatically at the outset’, that such beings have, in Kant’s words, ‘absolute worth’ (p. 97). This is not the conclusion of a regress argument. It is the starting point of an argument for the claim that only rational beings are ends in themselves, and therefore capable of being the sort of ends that could ground categorical imperatives. If Stern’s reading is correct, we can see why Korsgaard and others are wrong to think the argument for FH is motivated by anti-scepticism. At this point in the Groundwork Kant takes the categorical nature of morality for granted (pp. 101–2). His argument cannot defeat the suspicion that ‘morality is no more than a cover for hidden exercises of power, interest, and corruption’ (p. 92), but it is not designed to. Instead, Kant’s argument is designed to show that our acceptance of morality’s categorical nature commits us to the belief that its commands are grounded in the absolute worth of the only thing that could be an end in itself – namely, humanity. Kant therefore fits Korsgaard’s description of the realist, whose belief in ‘intrinsically normative entities’ is rooted in a confidence about morality’s authority (quoted on p. 103). Ultimately, he does want to address the worry that it is an illusion. The main concern for Kant is that we are unable to be necessitated by categorical imperatives because we lack free will. He replies to this concern with transcendental idealism, not with the argument for FH (p. 105; cf. Chapter 4).

Taken as a whole, Chapters 3–5 dig deeply into Korsgaard’s constructivist interpretation of Kant and the related question of scepticism. Stern’s treatment of scepticism is characteristically sophisticated and his treatment of Korsgaard is admirably sympathetic, given his opposition to her position. I highly recommend these outstanding chapters. In fact, I highly recommend the collection as a whole, which should appeal to a wide variety of readers. It complements Understanding Moral Obligation nicely – clarifying and enriching Stern’s view of Kant, but also extending it to address an impressive range of philosophers and issues.