1

I begin with Herman's response to the charge that Kant's ethics is formal with regard to its application. Again, the charge is that Kantian morality requires that in deriving particular duties from the supreme moral law, we abstract from the concrete particularity of cases. On this interpretation, the Kantian agent is taken to apply ‘fixed rules to particulars in the manner of a legal system’ (Herman Reference Herman2008: 1). Its applications, in Herman's words, are portrayed as ‘deductive rather than deliberative’ (ibid. viii).

Herman argues that it is a mistake to suggest that, for Kant, ‘the details of what we are like or the circumstances in which we act’ are of no relevance to our ‘fundamental moral questions’ (ibid. vii). In her view, Kant holds that our obligations cannot be understood without, ‘locating them within a social setting’ (ibid. 40). For the Kantian, in her words, moral reasoning starts with the ‘socially available knowledge we use to interpret our circumstances of action’ (ibid. ix). Our applications of the moral law thus require a grasp of our ‘social world’ and of local as well as contingent ‘norms of correctness for choice and action’ (ibid.). Our applications are in this sense ‘deliberative’ rather than ‘deductive’.

More precisely, the Kantian agent can only properly apply the supreme moral law if she possesses what Herman refers to as ‘moral literacy’. The morally literate person is sensitive to particularity. She recognizes that not all cases are alike; she knows that what is right to do here and now is often ‘a function of what is specific to a given situation’ (ibid. 1). She is thus sensitive to ‘local value’, and she takes seriously the fact of cultural pluralism (ibid. 44). She knows that in applying the Kantian rule not to harm, for example, we need to appreciate the fact that cultures define harm differently.Footnote 1

In addition to attending to particularity, the morally literate person possesses qualities of character that allow her to satisfy what Herman calls the ‘requirement of salience’ (ibid. 1). We can only adequately attend to particularity if we possess the ability to separate out facts that are morally relevant from those that are not. The morally literate person, then, exercises good judgement in identifying what is practically salient or significant (ibid. 6).Footnote 2

In these ways, Kantian moral agency requires moral literacy, according to Herman. The morally literate person does not abstract from the particularities and contingencies of cases; she takes them into account. She knows that the supreme moral law justifies a great variety of particular duties, because she is sensitive to the fact of cultural pluralism.Footnote 3

2

I now turn to Herman's response to the complaint that there is empty formalism in Kant's conception of moral agency. As Kant is standardly interpreted, the moral agent is motivated not by desire but by the categorical imperative, a law of pure practical reason. Herman notes that Kant's critics are mystified by the very idea of a purely formal motivational force, a force that, as she says, ‘seems to come from nowhere’ (ibid. 3). While most of us are ready to accept the idea that we can be moved by passion and desire, we are less comfortable with the suggestion that we can also be motivated by pure practical reason, by a force that is defined as wholly non-natural, and that is thus in no way tied to our natural desires.

Herman argues, however, that this standard reading makes the mistake of attributing to Kant a ‘rigid oppositional model’ of the relation of pure reason to desire (ibid. 13). On this model, Kant is supposed to hold that desire has no role to play in the motivation of the moral agent. At most, desire can be governed by and brought into conformity with reason. This interpretation, Herman tells us, represents the Kantian model of motivation as requiring the ‘oppression’ of the agent's affective life by morality (ibid. 20–1).

As Herman sees it, however, this account of the motivation of the Kantian moral agent is a mere caricature. It is a mistake in her view to assign to Kant the ‘rigid opposition’ model of the relation between practical reason and desire. Kantian moral motivation is neither merely formal nor wholly mysterious. On her proposal, we can accurately attribute to Kant the assumption that the moral agent acts from desire – at least, from desires of a certain sort (ibid. 11).

Now, in claiming that the Kantian moral agent acts from desire, Herman does not mean to deny that the source of the supreme moral law is formal. But the formal or a priori origin of the law, she says, accounts only for its ‘authority’, not for its ‘efficacy’ (ibid.). For Herman, the fact that the categorical imperative derives from pure practical reason does not imply that the motive of the morally worthy agent has to be purely formal (ibid.). Kant can in other words allow that desire is part of the motivation of the moral agent. Herman's point here is not simply that desire can be part of the motivation in that incentives of desire sometimes accompany the motive of duty.Footnote 4 Her point instead is that desires can be part of what makes the agent's motivation moral. In some cases, that is, it is the presence of desire that accounts for the fact that the agent has a (practical) reason to act (ibid. 8). Herman thus concludes that it is a mistake to attribute to Kant the view that there is a ‘rigid opposition’ between reason and desire. We need not assume, she writes, that the ‘purity of the motive of duty’ consists for Kant in its ‘complete separation from the empirical life of the agent’ (ibid. 10–11).

How does Herman make her case? In some of the richest passages in her collection, she argues that Kant can grant that in some instances desires can be ‘rationalized’. While it is true that some of our desires are simply given by nature or ‘brute’, most desires, in her words, are ‘responsive to reasons’ (ibid. 12). Most, that is, can be moulded or shaped – and not just constrained – by reason. Although what we desire in the satisfaction of self-interest, for example, may originally lead us to seek out nothing other than food and sex, our desires can evolve. Under the influence of reason and its practical law, self-interest can take on new meaning; we can learn to associate with it a new set of desires (including, possibly, the desire to attend to the welfare of others) (ibid. 23).

In rough outline, this is how Herman argues that Kant is able to ‘mitigate the hostility between reason and desire’ (ibid. 18).Footnote 5 Although desires on his account are not directly reasons for action, ‘valued desires’ are. Desires that have been ‘rationalized’ or ‘socialized’ by the supreme practical law can indeed be part of the motivation of the Kantian moral agent.

3

I think there can be no doubt that Herman successfully undercuts a certain set of common complaints about Kant, a certain version of the empty formalism critique. She convincingly demonstrates that, in applying the moral law, the good Kantian attends to rather than ignores the empirical particulars of individual cases. She is right to argue in addition that, in a certain sense, desires can have an essential and legitimate place in the motivational set of the Kantian agent. She effectively exposes the mistakes, then, of those who claim that the Kantian does not sufficiently attend to empirical content in either of these two ways.

But is this all there is to the empty formalism critique – to the critique that, as Herman notes in her Preface, goes all the way back to Hegel?Footnote 6 It is here that I have my doubts. For it seems to me that there is another version of the empty formalism critique that Herman does not address. This other (and, I believe, Hegelian) version aims its attack at the supreme practical law itself. It challenges the purported status of the law as a priori, as derived from pure practical reason rather than experience. It also calls into question the supreme law's universal and necessary validity.Footnote 7

Let us recall Kant's reasons for insisting that the supreme practical principle must be a priori or formal. We all agree, he writes in his Preface to the Groundwork, that ‘if a law is to hold morally … [it] must carry with it absolute necessity’. Kant argues in addition that the supreme principle must be universal in scope; it must bind all rational agents. Moral imperatives can enjoy this kind of validity, however, only if they are grounded in something other than experience. The ‘ground of obligation’, Kant thus concludes, has to be sought ‘a priori in concepts of pure reason’ (Kant Reference Kant1996a: 389).

The empty formalism critique I now want to explore defends two main points. First, it argues that the supreme practical law is not the formalism Kant claims it is. Were the law really formal, it would be empty; it would be unable to guide the derivation of particular duties in a non-arbitrary way. Kant is only able to derive particular duties from the supreme law because he presupposes content, because he relies on background assumptions about, for example, the nature and ends of rational agency. Second, when we subject these background assumptions to careful scrutiny, we discover their contingency. And this implies that the law Kant claims to be universal and necessary lacks this status after all.

As far as I have been able to determine, Herman has no doubts about the formal or a priori status of the supreme practical law. Instead (and as a true Kantian), she underscores the virtues of the law's formalism. Kant's supreme practical principle, she writes in ‘Rethinking Kant's Hedonism’, demands ‘our equality as persons with dignity’. It is to Kant's credit, she suggests, that he insists upon the ‘nondiscretionary’ nature of this demand (Herman Reference Herman2008: 197). Were the grounds of our actions material or empirical rather than formal, they would in her words be ‘fungible’ or ‘defeasible’ (ibid. 190–1); they would deprive us of any basis for ‘objective value’ (ibid. 177). This is why Herman argues that not even a sophisticated hedonism is able to supply an adequate foundation for practical value. Even a sophisticated hedonism rests on material rather than formal grounds.

As a true Kantian, then, Herman does not countenance the possibility that the supreme practical law is itself discretionary or contingent. Her commitment to the law's universality and necessity shows up, among other places, in her definition of moral literacy. The morally literate person understands, she tells us, that morality rests on a standard that is ‘presumptively the same across various moral communities’ (ibid. 80). While recognizing that our applications of that standard need to be sensitive to variations in context, the morally literate person affirms that the standard itself is invariable (ibid. 45).

The empty formalism critique I am now considering charges that this thesis of the invariability of the moral law collapses under pressure. As I noted a moment ago, this version of the critique calls into question the purported formalism of the law. It draws our attention to background assumptions that must be in place to guide the law's application. In exposing these assumptions, the critique aims to shake our confidence in the law's universal and necessary validity.

Moreover, the effort to shake our confidence in the universal and necessary validity of the supreme practical law serves a larger agenda. For ultimately, this version of the formalism critique challenges Kant's commitment to the assumption that human reason is capable of producing rules that are wholly a priori in origin, rules that therefore in no way reflect contingent historical reality. The net effect of this formalism objection, assuming it is successful, is thus to awaken in us scepticism about Kant's separation of the realms of the rational and of the actual. If successful, this objection is supposed to get us to embrace a new account of the nature of each of these realms and their respective faculties.Footnote 8

As we saw above, Herman seems convinced that the worry about a gap between the realms of reason and experience, or between the faculties of reason and desire, is misplaced. A proper or ‘enlarged’ reading of Kant, she urges, appreciates the extent to which his system ‘lets the phenomena in’ and is thus not a mere formalism. His system ‘lets the phenomena in’ in the two ways we reviewed earlier. The Kantian approach, in Herman's words, ‘mitigates the hostility between reason and desire’ (2008: 18). It can allow that we have desires that are rational, desires that are formed or moulded by reason. The Kantian approach ‘lets the phenomena in’, furthermore, by requiring that the moral agent be a good empiricist. The Kantian moral agent possesses a good character, and the person of good character is sensitive to the particularities and contingencies of experience.Footnote 9 She recognizes, in light of the differences of individual cases, that there will be variation in the particular duties we derive from the supreme practical law (ibid. 45).

But while Herman persuasively demonstrates that there is a sense in which, in ‘letting the phenomena in’ in each of these ways, the Kantian gap between reason and nature is not as wide as is commonly supposed, she follows Kant in keeping the gap wide open in another sense. The gap must remain open for reasons we noted earlier. If we are to ensure the ultimate objectivity of the supreme practical principle, if we are in other words to secure its necessary and universal validity, we need to insist that its basis is pure reason rather than experience. Kantian practical philosophy is compatible with the ‘great variety of human interests and traditions’, as Herman says; but there is, in her words, ‘room for only one supremely regulative value in the Kantian deliberative field’ (ibid. 45; emphasis added).

Herman therefore follows Kant in keeping the gap between reason and nature open in this respect. The question before us now is whether she should. Is there really good reason to endorse the Kantian idea that there is pluralism and variation on the side of our applications of the practical law, but not on the side of the law itself? Is there any validity to the empty formalism critique of this central Kantian assumption?

Obviously, these questions aim at the very core of Kant's philosophy. I am going to try to provide a few reasons, now, for taking them seriously. Later on, I will suggest that even Herman's Kantianism gives us reasons for taking them seriously. That is, I will eventually argue that we can discover in her essays evidence that not even she is able to remain faithful to the thesis that the supreme practical law is invariable.

To restate the question before us now: are there good grounds for holding, with Kant, that there is pluralism and variation on the side of our applications of the supreme practical law, but not on the side of the law itself? In some cases, this thesis seems obviously true. The categorical imperative obligates us to be beneficent. You send your money to OXFAM; I send mine to CARE. In this case, we each comply with the same rule; we merely do so in insignificantly different ways.

But cases get complicated fast. Indeed, two of Herman's own examples seem to me to suggest more than she intends them to; they suggest variation not just in our applications of a rule, but in our understanding of the rule itself. After reminding us in her essay ‘Pluralism and the Community of Moral Judgment’ of the Kantian position that ‘no maxim may be inconsistent with the principle of respect for persons’, she notes that the command that we respect persons may be legitimately instantiated in different ways (Herman Reference Herman2008: 45). She asks us to imagine two cultures committed to different accounts of the conditions of proper child rearing. A ‘culture that values individual autonomy’, she writes, may consider the ‘pain of separation from a parent’ to be a necessary stage of growth and thus not a ‘sign of injury’ (ibid. 42). In contrast, a culture that places great weight on ‘intergenerational bonds’ might consider early separation significantly detrimental to the development of a child. For Herman, the example illustrates how the general command to respect persons will sometimes yield different specific obligations; there will be variation in how we think the command should be carried out.

But is Herman right to conclude that the differences in this case have to do with nothing more than differences in how we apply the rule? The rule commands that we respect persons. The two cultures disagree over the conditions of child development. Underlying their respective views about child development, presumably, are different conceptions of human maturity. In the case of one culture, maturity requires separateness as an essential ingredient; in the case of the other, what matter most are tight familial bonds. Arguably, the two cultures embrace separate accounts, not just of how human maturity is to be achieved, but of its very nature. If so, then it would seem that what is ultimately at stake are competing metaphysical views of the self. But if there is disagreement at this level, in what sense are the two cultures really complying – or seeking to comply – with the same rule?

Discussing another example, Herman notes that, although it is ‘reasonable to suppose that every valid moral conception will have a standard of harm’, there is likely to be variation on the local level in what we identify as harmful (ibid. 40). There may be disagreement, for instance, over whether harm can be psychological as well as physical in nature. Once again, I wonder whether the disagreement Herman highlights in this case goes deeper than she admits. As she says, there is variation in what is considered harmful. The interesting question here, however, concerns the basis of the disagreement. In this case, the disagreement seems to have its source in different conceptions of the nature of human experience. As Herman tells the story, one culture denies the phenomenon of psychological pain. But if that denial is an expression of an underlying scepticism about the reality of the mental, then it once again looks as if the differences between the two cultures run deep. The case can be made, I think, that the cultures part ways not just in how they apply the rule not to harm, but in how they each understand the rule.

If these two examples do not thoroughly convince, hopefully this final one will. Consider the command, ‘Respect human freedom’, or the command, ‘Honour the dignity of persons’. Do we really want to say that the Aristotelian and the Utilitarian and the Kantian, for instance, differ only in what they take to be the implications of these commands for action? Is it not rather the case that the rules cannot have the same meaning for each of these three philosophical approaches? The rules cannot have the same meaning, because each of these approaches defends a unique account of what it means to be a person, a unique account of the nature of human dignity and freedom.

It seems to me that if we cling to the view that the two rules in this last example are the same for the Aristotelian, the Utilitarian, and the Kantian, it has to be because we have a high tolerance for vagueness. It is because we are content, for some reason, to leave the meaning of the key concepts of these rules largely unspecified. But why would we want to do that? One reason strikes me as obvious: only a highly vague and general rule is likely to pass as universally and necessarily valid. If we get too precise – if we add too much content to our rule and actually spell out what we mean by the conditions of personhood, human dignity and freedom – we are apt to encounter difficulty in defending our rule's invariability.

As we know, Herman is committed to the thesis of invariability. Perhaps this explains why her own expressions of Kant's supreme practical law tend to be rather vague. In reading these essays, I could not help noticing that the keen attention to detail that so enriches her discussions of applications of the law is missing from her accounts of the law itself. In ‘Pluralism and Community …’, for instance, she writes that there is ‘room for only one supremely regulative value in the Kantian deliberative field’ (Herman Reference Herman2008: 45). This value, she continues, is the ‘objective moral principle’ Kant calls the categorical imperative (ibid. 38). The categorical imperative, in her words, specifies that ‘local value has moral standing as it does or can express the value of rational agency’ (ibid. 44). The law commands that ‘no maxim of action may be inconsistent with the principle of respect for persons’ (ibid. 45). The law forbids us from disregarding ‘the moral status of another person’ (ibid. 41). The moral status of persons, Herman tells us in ‘Responsibility and Moral Competence’, refers to ‘a practical disposition that enables recognition of morally salient features of circumstance and action, as well as the regulating capacity to do what is seen to be right’ (ibid. 98). This status, in her words, is ‘sufficient to secure agents’ imputability for the effects of most of their actions’ (ibid. 98).

Herman is somewhat more specific in her essay ‘Bootstrapping’. She elaborates there what she believes Kant has in mind by the ‘minimum moral capacity’ responsible for moral agency. We have this capacity thanks to our possession of a pure will, of pure practical reason. This capacity gives us special causal powers (ibid. 169). As Herman says, the capacity may be understood as the freedom to act, ‘according to our conception of laws’ (ibid. 168).

But phrases like the ‘value of rational agency’ and the ‘moral status’ of persons are still highly abstract. Their key terms signify different things to different philosophers. If we survey the history of practical philosophy, we discover a variety of ways to fill in the conditions of human dignity, a variety of ways to explain what it means for us to respect the moral status of persons.

Now it may be that, in charging Herman with vagueness here, I am being uncharitable. Perhaps it is unfair to suggest, as I just have, that it is in the service of her commitment to the invariability of Kant's supreme practical law that her characterizations of it tend to be vague. In her defence, we should bear in mind that it is not the project of this collection of essays to assess the numerous interpretations of the categorical imperative, or to examine Kant's multiple formulations of it. Herman's express aim is instead to focus our attention on ways in which Kant's philosophy – formalist though it is – can nonetheless ‘let the phenomena in’. So it may be that what I am identifying as a sin of commission, on Herman's part, is really no more than a sin of omission. Her collection, Moral Literacy, is not after all an exploration of the categorical imperative.

Nonetheless, the essays in Herman's collection clearly rely on an interpretation of the categorical imperative. Accompanying her many discussions of the ways in which Kantian practical philosophy can accommodate contingency and pluralism, ways in which it not only can but must be sensitive to ‘local value’, are her repeated reminders of the supreme practical law's universal and necessary validity. The law must be universally and necessarily valid, and must therefore have its source in pure practical reason, she claims, if it is to adequately ground objective value. Herman thus repeatedly puts her cards on the table regarding what she believes is the special status of Kant's supreme law. She frequently asserts her commitment to this Kantian thesis.

In reading Herman's essays, however, I found myself hungry for a more contentful account of the supreme law. I wanted her to say more, not about the law's special status as universally and necessarily valid, but about its meaning. The problem with not giving us a more contentful account of the law is that Herman thereby makes it appear that a thesis that should be hard to defend is easy to defend. The thesis of the invariability of the practical law implies that the law is necessarily valid for all rational natures; it implies that the categorical imperative is the one and only supreme practical principle for all cultures, all traditions, at all times. This is a hefty claim. If we gloss the law simply as the command that we respect human freedom or honour the moral status of persons, it can seem obviously true, or at least uncontroversial; we can seem justified in simply taking the thesis of the law's invariability for granted. But we pay a high price for our vagueness here, I think. We leave out of our account features of the command that distinguish it as truly Kantian. We leave ourselves nothing to say about how Kant's command that we respect human freedom differs from that of other moral philosophers – differs, say, from that of Rawls.

4

Let me summarize the point of my critical remarks so far. I have been drawing attention to Herman's commitment to the Kantian thesis that the supreme practical law is invariable. I have done so because this is the thesis called into question by the version of the empty formalism critique she does not address, the version that raises doubts about the supreme law's universal and necessary validity. The formalism critique Herman does not address is sceptical about our prospects for drawing a sharp line between the invariability of our supreme practical principle and its variable applications. Although I have provided a few examples to motivate this scepticism, I do not take myself to have thereby established that Kant is ultimately unjustified in holding that the categorical imperative is the supreme and invariable arbiter of objective practical value. Perhaps the categorical imperative has exactly the special status Kant awards it.

But then again, perhaps not. It is worth noting, I think, that if we are looking for a defence of the supreme law's invariability, we are not going to find it in this collection of essays. I have already mentioned one reason for this. I noted that it is not Herman's intention in this collection to offer us either an interpretation or defence of Kant's categorical imperative. But there is a further explanation for why these essays provide no justification of the law's invariability. The further explanation is this: if we look closely, we discover evidence that, under Herman's treatment, the law that is supposed to be invariable in fact undergoes change. Although Herman is unwavering in her commitment to the thesis of the invariability of the law, she is not, I think, entirely true to that thesis. And this means that her treatment of the supreme practical law plays into the hands of the version of the empty formalism critique we have just been considering.

What I am proposing is that we can find in these essays evidence that Herman is not an orthodox Kantian after all. In common with a number of contemporary Kantians, she is reluctant to follow Kant every step of the way. We have seen that, with Kant, she insists that the supreme practical law has to be universally and necessarily valid in order to secure the objectivity of practical judgement. With Kant, she asserts in addition that the law can only be universally and necessarily valid if it is formal rather than material.Footnote 10 We have also seen that her collection is unified around the idea that, although Kant's practical philosophy is undeniably formalist in certain respects, it can also allow a role for empirical content; it can and does let phenomena in. The unorthodox nature of Herman's Kantianism, however, shows up in her essays in the following way: the formal features of Kant's theory (the features that are supposed to secure its objectivity) are barely in view. Not only does Herman give us little explicit discussion of Kant's formalism, the language she frequently enlists in describing key Kantian doctrines seems well suited to distract our attention away from it.

To cite just a few examples, Herman frequently tells us that Kant defends a ‘naturalistic account of reason’, and a ‘naturalized’ moral psychology (2008: 10, 24). She in addition stresses what she calls the ‘deep fact of sociality’ in Kant's work. Agency and judgement, she says, have for Kant a ‘social basis’ (ibid. 52–3). Moreover, she repeatedly reminds us of the sensitivity of Kantian moral theory to contingency and conceptual change (ibid. 23, 33, 45).

I am not suggesting that, in making these claims, Herman positively distorts Kantian doctrine. When she tells us that that agency and judgement have a ‘social basis’, her point is not that they are wholly grounded in actual social conditions. Agency and judgement are social, on her account, because they reflect our common rationality; they are grounded on ends shared by all rational natures, not on the interests of individual empirical wills (ibid. 59, 61). And in stressing the degree to which Kantian practical philosophy can accommodate conceptual change, Herman is not forgetting Kant's insistence that there can be no change in the supreme law itself. She merely means to emphasize that our applications of the law must respond to the variations of circumstance.

But even though Herman's references to the social and empirical elements in Kant are not actually distortions of Kantian doctrines, we have to look awfully hard to discover that they are not. Because Herman so foregrounds the role of empirical content in Kant's practical philosophy, the formal elements of his theory are often hardly detectable. Her frequent reminders of Kant's ‘naturalistic account of reason’ and of his commitment to the ‘deep fact of sociality’, tempt us to conclude that, on her account, Kant's approach is empiricist all the way down.

I think all of this is evidence of Herman's discomfort with the formal features of Kant's approach. The discomfort shows up even more obviously in the few passages in which she addresses the formalism directly. She tells us in her Preface that she aims to ‘liberate’ Kantian moral theory ‘from its excess noumenal baggage’ (ibid. 2). We learn in her essay ‘Bootstrapping’ that the ‘excess noumenal baggage’ from which she aims to liberate us is Kant's conception of the will as noumenon, as capable of a special, non-natural form of causation. On the one hand, Herman says she is convinced that we ‘have reason to accept the will as a cause because it provides the best explanation of what we do when we act morally’ (ibid. 174). We need not worry, she tells us, that the doctrine of the will as noumenon commits Kant to a ‘two-world’ ontology.Footnote 11 But then she confesses that she is ‘uneasy’ with the ‘two-standpoint’ interpretation as well. She confesses that she does not know ‘whether there can be a metaphysically acceptable account of the kind of cause Kant thinks the will is’ (ibid.).

There are hints, in these and other remarks, that Herman believes that the doctrine of the noumenal will is dispensable, and that we can get at the idea of a mental cause without it. This indicates a significant departure from Kant. Although it may be that Herman is right to assert that the idea of a mental cause ‘does not by itself require noumenal causality’, Kant's conception of practical agency requires noumenal causality. The freedom Kant attributes to the moral agent is not that of a mere mental form of causation. A mere mental or ‘internal’ freedom, as he calls it, is nothing more in his view than the ‘freedom of a turnspit’; it is the compatibilist version of freedom he so passionately rejects (Kant Reference Kant1996b: 218). On Kant's account, only one kind of freedom is capable of grounding a metaphysics of morals, and that is the freedom he refers to as ‘transcendental’. The agent who is free in the transcendental sense acts not just from a mental cause; she is capable of initiating a causal series from a standpoint outside time.Footnote 12

Herman may be wise to have reservations about this view, but it is Kant's view nonetheless. If we try to replace it with some non-noumenal alternative, we depart in a significant way from Kant's metaphysical grounding. For if there is no argument for freedom in the transcendental sense, then there is also no way of accounting for what Kant means by the unconditional ends the categorical imperative commands us to respect. There is no way of getting at his unique conception of human freedom and of the dignity of rational nature.

What I am suggesting, in conclusion, is that there are indications in Herman's essays of a significant departure from Kant, of a less than wholehearted embrace of the metaphysical basis of his theory. And this means that where there was supposed to be invariability, there is change.