Critique

It is requisite to explore the infrastructure of Kant's moral law as it sets the background for Wood's analysis. Section one of the Groundwork introduces Kant's famous argument for the absolute primacy of the good will (G 393, 401) and the necessity to act ‘from duty’ (G 397–8) that follows from it. An action is willed because, Kant assumes, the agent knows it is the right thing to do, irrespective of the desirability of empirically contingent objects or goals. Willing is a capacity intrinsic to an agent who is rational. Yet humans are Janus-faced beings whose rationality is unsettled by the undercurrents of desire. An individual enters the moral universe of persons when the will, as a rational faculty, begins to tame desire (G 400; M 383). Conscious and consistent resistance to one's desires is what differentiates the I, or the moral self, from its desires. Whilst a holy will always wills the moral law, an imperfect will such as the will of a human being must be coerced to do so – this is why the moral law has the form of an imperative or command (G 413; CPractR 32). And since the moral law holds with necessity across the moral universe, or for all rational beings (G 412), it is like an objective law of nature (CPractR 19). Wood slips by describing the latter as a ‘regularity’ (p. 70),Footnote 2 given that Kant postulates a parallel between the moral realm and the realm of nature (which reflects metaphysical presuppositions): the moral law is like a law of nature or like a principle holding with universal necessity, causally, and not as a matter of regularity.

Wood laments the prevalent tendency of reading the Groundwork selectively by singling out section one, as it contains the crudest, universalizability, formulation of Kant's moral law (G 402, pp. xi, 41, 73, 81). This formulation is reiterated in the opening passages of section two (G 421) and Wood calls it the formula of universal law (FUL, G 402, 421, p. 66) or formula of a law of nature (FLN, G 421, 436, p. 66). Section one is especially problematic because there Kant links the moral law to the notion of abstract willing (Wille) and not to determined willing or choice (Willkür) (M 226, p. 121). Abstract willing is a principle of universally good motivation that must be object-independent but, as Wood intends to show, our moral experience is about concrete choice situations geared towards concrete ends. So we must read Kant's ethics as a theory of choice with a teleological bent – more on this in a minute. On to the second and third versions of CI: the formula of humanity (FH), and the formula of the realm of ends (FRE) (p. 67). FH requires persons to be always treated as ends in themselves and not as a means only (G 429), and FRE invokes Kant's ideal of a kingdom of ends constituted by autonomous persons who respect each other's ends without necessarily agreeing on their desirability (G 437–9). This picture of the moral life reflects scepticism about a single public standard of value, be that the common good, progress or happiness. Wood finally equates FRE to a formula of autonomy (FA; G 431–2) which gives us a superior rendition of the notion of universality compared to FUL/FLN (pp. 66, 75–6, esp. 84). So FUL and FLN are homologues the same way FRE and FA are.

Despite their homology, FUL is not identical to FLN. Unlike FLN which is a metaphysical principle underpinning the above-mentioned parallel between the natural and the moral realm, FUL is restricted to the moral domain of practical action. There, FUL operates as a procedure for testing the moral suitability of individual maxims of conduct. A rational individual who seeks to establish the rightness of an intended action must test it by first transforming it into a maxim of action, and then by transforming it into a general rule. Imagine this scenario – the agent pondering an action: ‘I want to do x’. Test via FUL: ‘Assuming that doing x is a maxim for me, what would happen if this maxim were a rule which holds universally, for every other rational creature including myself?’ FUL has a dual character – it is at once a testing procedure for individual maxims, but also an objective standard or ‘norm’ for deriving subjective maxims of conduct from objective principles of conduct. (Wood suggests that FLN is the testing element, and FUL the normative element, p. 70.) Central to this subjective–objective nexus is Kant's notion of duty. Duty is the objective moral law conceived subjectively or from the vantage point of the individual person labouring under its constraint. Duty expresses, without being coterminous with, the moral law – the subjective ‘necessitation’ of duty mirrors the law's objective ‘necessity’ (G 412–13; M 218).

But FUL is unfit to serve as the basis of Kantian moral philosophy, Wood argues, because moral conduct does not boil down to testing procedures (p. 71). This is a cogent argument. Yet it leaves open the question of the standard-setting aspect of FUL. And as Wood himself writes: ‘It is hard to see how one could justify in Kantian terms anyone's acting on a maxim (or on the presumptive universalized form of a maxim) unless it is possible for someone to will that maxim as a universal law’ (p. 217). This quote is from chapter 12 ‘Punishment’, with Wood contending that we should be wary of conflating the act of contemplating an action (in thought) with the act of willing it (as a maxim). None the less the bottom line is that universalizability, as a standard, has the standing of a first premise in Kant's moral science. This would suggest that FUL/FLN is basic to Kantian morality. Wood responds that, although the Kantian notion of universalizability is important (otherwise we would be left without a compass in the sea of uncertain moral choices), its correct formulation is FA (FRE), not FUL (or FLN). What are we to make of this proposition?

Wood's response alludes to hermeneutics as a tradition distinct from Kant's analytics. Abstract principles such as the supreme principle of morality (G 392) should be translated into determined rules or duties in a hermeneutic rather than analytic manner (pp. 60–1). Adopting this proposal is tantamount to a rejection of the equivalence thesis – regarding the three CI formulations as (moral) homologues – and endorsement of a hermeneutic model where the three are seen as ‘parts’ of a ‘whole’ or a moral system (pp. 69, 80–1). It is conventional wisdom that the first formulation of CI (FUL/FLN) stipulates what persons ought to do, granted that the moral law is a valid principle of rational conduct, whilst the second and third (FH, FRE/FA) reverse the puzzle and announce reasons moral agents ought to have for treating the moral law as valid (p. 56). (The topic of reasons will be elaborated below.) Unlike the first, Wood argues, the latter two formulations operate as justificatory standards rather than as criteria of logical derivation. To explain an act is to account for it in logical terms but to justify it is to adduce reasons for it, assuming that reasoning terminates in an ultimate ground of right and wrong. Justification in short constitutes morally grounded reasoning. Conceived hermeneutically, Wood continues, FRE is the key moral criterion because it condenses Kant's argument for duty-bound conduct into a totality. This totality or ‘complete determination’ in Kant's terminology is achieved by combining the form of FUL/FLN with the matter of FH (G 436; pp. 59, 80, 116). FH is the central logical criterion allowing the transition from the first to the third formulation so that the ‘parts’ of Kant's argument for duty can be conjoined into a ‘whole’. FH and FRE, as postulates of autonomy and freedom, emerge as focal in Wood's Kantian project.

The remainder of this review explores Wood's major innovation in the book, the proposal that Kant's moral metaphysics be overcome for the sake of a grounded moral analysis by:

1. a. replacing abstract moral principles with grounded rules and duties (p. 57);

2. b. replacing Kant's Reason (writ large) with Kantian reasons (in the plural).

Beyond Abstract Moral Principles

Wood portrays Kant's moral science as a self-standing system of practical reason which demands no complementary metaphysical presuppositions about the self, (theoretical) reason, causality and transcendental freedom such as those investigated in the first Critique (p. 138). But I wish to argue that stripping Kant's moral philosophy from its metaphysical attire is counterproductive. To begin with, Wood relies heavily on the concept of rationality and rational desire (pp. 35, 146, 150).Footnote 3 As he warns us, we must be careful not to attribute a simplistic reason vs. desire opposition to Kant since the supreme principle of respect for duty is itself a kind of desire: rational desire (to be dutiful). But whilst insightful, this argument enlists the premise of reason, and reason for Kant is a faculty operating across the moral and metaphysical domains. Practical reason is simply theoretical reason applied practically. It would be difficult to cut out theoretical reason, whilst retaining practical reason, because if we did that, the moral agent would be unable to imagine itself as a free agent acting within the moral domain. That is, the agent must first develop a conscious notion of its own self (CPR B422), as an integrated entity distinct from its environment which includes objects as well as other selves, and then as a causally efficacious entity which makes things happen (CPR A546–7). After these conditions have been satisfied, it can get into a position to think of itself as a person who in acting freely may ‘cause’ wrongs to other persons (G 450; CPractR 49, 55). Both selfhood (the ‘unity of apperception’, CPR B131–2) and causality (CPR A202) are metaphysical categories. Wood's excellent chapter on conscience (chapter 10) in fact gestures towards metaphysics, exhibiting the self as a bifurcated self, at once a ‘judge’ and an ‘accused’, in the court of reason.

The source of Wood's dissatisfaction with FUL is now obviated. FUL fares badly because it is a metaphysically grounded principle. Conversely, FH and FA/FRE fare much better as principles further removed from metaphysics: their abstract language can be translated into determinate rules and duties because it is a language of ends. FH as a formula of humanity demands rational beings to be respected as ends in themselves, and FA/FRE as a formula of autonomy requires free and autonomous persons to respect each other's ends within a pluralistic kingdom of ends. The notion of ends is determined rather than abstract; an action cannot have a determination – as this or that action – unless it aims at a particular wished-for end. Kantian moral action has a teleological structure – FH and FRE are felicitous formulations of CI because their logic is compatible with teleology. FUL, Wood concludes, should be overcome (or ‘sublated’ as we may say with Hegel) in favour of FH and FA/FRE.

The result is a Kantian ethics as a theory of action in modern parlance. Whenever the moral agent wonders, ‘What ought I to do?’, FUL cannot provide enough information. Wood's point is not that the supreme principle of morality announced by FUL is wrong but that it is incomplete. It cannot ever be complete because it generates commands about how persons ought to act, whereas all they need is indirect guidance to aid them in thinking for themselves, or autonomously, about a choice they have already made (p. 63). Completeness is another hermeneutic criterion: abstraction is incomplete without a determining ground. Notice that ground has a dual meaning – it refers to: (1) foundations (moral foundations in this case); but also to (2) the notion of differentiation or ‘particularising the general’.Footnote 4 If in seeking out the wise counsel of duty we are baffled by the abstract propositions in the Groundwork, we should not dig into the metaphysical depth of noumenal freedom in the Critique of Pure Reason but consult the Metaphysics to see what possible kinds of duties there are (pp. 73–4).

The taxonomy of duties in the Metaphysics illustrates an often neglected point: we are not subject to duty, full stop, but to various species of duties. Duties vary according to criteria such as the nature of a relation between self and others, inner motivation and coercive enforceability. Duties of ethics (which ought not to be enforced) are to be differentiated from duties of right (coercively enforceable) (M 219, 239), and similarly, imperfect duties from perfect duties (M 220–1). For example, the duty to aid the poor is imperfect whereas the duty not to steal is perfect. Perfectness marks the degree of objective necessitation residing in a given duty. Whereas duties such as these two are examples of external freedom, a duty to develop one's talents depends on inner motivation, hence it is an ethical duty. Because they regulate the external or public relations of persons, duties of right have a juridical character (M 219); but since they are duties, they rest on an irreducible ethical foundation. Wood's insight here is that moral conduct has context. If we intend a Kantian defence of the modern state as a coercive system of rules, our preferred idiom should be duties of right rather than abstract universalizability.

This plea for contextualized reasoning might evoke the notion of Hegel's ground as a complete, in the sense of ‘historical’, determination of an identity but Wood's scepticism about metaphysics is not provoked by a historicist critique. So his claim is not that metaphysical principles demarcate the contours of a moral system, whereas applying the system in assessing the moral conduct of an actual subject, say, the modern European state, hinges on an understanding of this subject as a historically situated being. The claim rather is that Kant's moral philosophy should be read as a theory of action and specifically as a theory of choice.

Reasons

Choice brings us to the topic of freedom. Because choice is a problem of free action, the focus in this part of the review is on FRE and FH. Wood's interpretation of the formula of humanity and autonomy emphasizes reasons rather than persons. Thus, I am not free because I am capable of freedom or autonomy (Kant) but because I am capable of acting for reasons, freely and autonomously (Wood). Other persons owe me respect, not in virtue of Kant's orthodox respect as non-interference (with my freely chosen ends), but because these reasons are themselves worthy of respect or are ‘good reasons’. ‘To act rationally’, as Wood remarks, ‘is to act for good reasons’ (p. 54). Wood in effect substitutes Kant's conventional thesis that moral agents act for reasons with an argument that agents act for moral reasons. Curiously, his definition of reasons is distinct from that of contemporary analytical philosophers who claim that to have an interest is to have a stake in one's own well-being.Footnote 5 So a reason for Wood is not a proxy of interest – reasons are ‘good’ on the condition that they refer back to a moral ground: Kant's moral law, articulated as FH and FRE.

But why ought X to respect Y's reasons (where ‘X’ and ‘Y’ are names of persons)? Wood argues that this respect materializes on the assumption that rational nature (pp. 40, 55, 58) is treated as a universal value (p. 94). To understand this proposition we must return to the notion of choice. A theory of choice must postulate a set of ends, and at least one basic value as a point of reference for measuring their desirability or appropriateness. For Wood, the basic value structuring moral choice is rational nature manifested as an ability of ‘setting ends according to reason’ (p. 94). The ends in question are certain ‘ends [which] it is our duty to have’ (p. 167) – individual perfection and the happiness of others (M 385–8, pp. 68, 219). Because ends which it is a duty to have belong to Kant's doctrine of virtue from the Metaphysics (M 239, 381, 383), the FH from the Groundwork must be revised to accommodate this doctrine (p. 143): what moral agents are urged to value as an end in itself is not personhood but rational nature (in a person).

Wood's reconstruction of Kant's theses of autonomy and freedom gravitating around the Metaphysics has three implications. First, rational in the expression ‘rational nature’ does not stand for Kant's faculty of subsuming the particular under general rules (CPR A195/B240). We can be agnostic about the metaphysical foundation of practical reason, as long as we accept rational nature as a basic value inside the moral realm. So for Wood practical reason is not predicated on the premise of theoretical reason but on rational nature which is not a faculty but a value (p. 94): it ought to be learnt. Second, practical reason does not announce principles of action: it announces rules of action linked to particular duties (and rights). Third, the differentia of duties is matched by a differentia of reasons deployed in the justification of moral choices.

Kant's reason – or Reason– is substituted by reasons. A discourse of reasons, Wood suggests, has its advantages because it permits us to incorporate considerations of welfare under the umbrella of respect. If Reason, as a faculty ordering objects in space/time via causal relations, were the true ground of morality, agents in whom this faculty is radically underdeveloped or lacking, such as animals (and perhaps senile individuals), would fall outside the moral universe and thus outside the remit of moral concern. If however reasons are the measuring rod of moral action – a rational agent may have good reasons to care about the welfare of animals, reasons which in turn ought to be respected by all other rational agents (pp. 95, 97). But the trouble is that ‘may’ is a permissive condition, not a stringent requirement. I may care for animals but I may choose otherwise: not to care. Wood's rejoinder to this objection runs like this. To ward off the possibility of autonomy sliding into rampant voluntarism – or the view that I am owed respect by others for whatever reasons I may have – a constraint must be imposed on all agents at the same time, as a universal criterion. This criterion, and this is an additional step, must delimit the scope of ethically permissible ends, leading us back to FH and FRE.

How are ethically permissible ends related to reasons? Kant evades this thorny question by excluding reasons and by elevating persons into ends in themselves: this is Kant's FH focusing on persons. Wood's modified FH emphasizes rational nature. Yet it is unclear whether this modification differs significantly from the original. Compare the conventional rendition of Kant: ‘I value your choice of ends because I value your person (as an end in itself) provided your act of free choosing does not violate FUL’, with Wood: ‘I value your choice of ends because I can see that you have good reasons and that you can in principle (granted that rational nature is a value) apply the same reasoning to a situation where I would be freely choosing ends and reach the same conclusion as I have just reached in relation to your reasons’. The domain of ethically permissible ends seems to overlap in both cases, questioning the distinctiveness of Wood's FH. When it comes to FRE, Wood's argument for the primacy of reasoning in terms of good or morally grounded reasons is more powerful. The kingdom of ends incorporates the notion of moral ground (freedom as autonomy) together with the notion of grounding as particularity: agents who are autonomous are free to pursue particular ends for particular reasons (provided these are good reasons).

Yet Wood's goodness-of-reasons argument is logically inconclusive. The culprit here is the notion of reasons itself. It is hard to defend any philosophical project centred on reasons because reasons are not end-states or ends; they are mid-points or means. Kant may have erred in thinking that moral action or action ‘from duty’ exhausts the entire sphere of prescriptive action – that duty is coextensive with oughtness – but philosophers like Wood who move backwards, from the realm of prescriptive action and reasons into that of morality, engage in an even more precarious enterprise. If we start with reasons, and then proceed to establish their ground, nothing guarantees that the said ground is moral. The grounding relation has two components, ground and reasons, and the direction of transition matters.

In Wood's scheme the move is from reasons to ground but its direction must be reversed to ensure that the grounding relation is moral in the sense Kant defines morality (G 390). Otherwise, the moral justification of reasons would collapse into an explanation by reference to reasons. If we can explain cruelty against animals, and if the ground-setting explanation apart from moral justification is context relative, we may end up justifying cruelty of this sort, as the boundary between the moral and the non-moral is contingent: the well-known danger of moral relativism. Here ‘contingent’ does not stand for Kant's ‘empirical’ (G 411) but for ‘changing’.

The cure for moral relativism is moral objectivity or a standard such as FUL/FLN. And if in our quest for moral objectivity we are prepared to discard Kant's Reason as an unwanted metaphysical baggage (dispensing with FUL/FLN), then we would be almost compelled to derive moral objectivity from history, be that a Hegelian system of historically contingent practices or Foucauldian genealogy (Foucault Reference Foucault1995). In-between Reason and history there seems to lurk a third way: Wood's ethics of reasons. But this via media, I think, is unlikely to complete the task of moral objectivity because reasons are supposed to import elements of subjectivity and contingency into moral conduct,Footnote 6 and cannot be invoked to objectify such conduct. The fundamental tension between contingency and objectivity has troubled moral philosophers before and after Kant. By affording us a glimpse into this aporia by means of a fresh and provocative analysis, Kantian Ethics discloses the virtues and limits of Kant's moral system.