Michael Cholbi. Ethical Theory and Moral Practice. Volume 16, Issue 3, June 2013.
One of the most resilient objections to Kantian ethics is that it is implausibly committed to a form of ‘rigorism,’ that is, that Kantianism displays a hopeless ‘insensitivity to moral complexity and righteous absurdity in requiring, as it is sometimes thought to, that we keep all promises, tell no lies,’ etc. (Herman 1993: 133. See also Hampshire 1978:28; Williams 1981:2-3; Hill 1992: 198-201) This objection interprets Kantianism as committed to a set of perfect duties, encapsulated in rules such as ‘don’ t lie,’ keep one’s promises,’ and so forth, and that these rules apply without exception.
Over the years, a number of Kantians have sought to blunt the objection that Kant’s perfect duties correspond to exceptionless rules. (Millgram 2003: 532ff.) For example, some Kantians claim that critics are overly fixated on his essay ‘On a Supposed Right to Lie’ and have therefore greatly exaggerated how committed Kant was to exceptionless perfect duties. The narrow, legalistic focus of the lying essay indicates that it should not be seen as an authoritative statement of Kant’s views about the morality of lying or about the stringency of perfect duties. (Paton 1953-54; Sedgwick 1991; Wood 2008: 244-251; Sussman 2009; Varden 2010) Another tactic is to suggest that at least with respect to lying, the prohibition in question is in fact narrow, since it applies to only one of the three senses in which (according to Kant) an utterance can be a lie. (Mahon 2009)
Others defenders of Kant point out that he does not explicitly endorse the claim that perfect duties admit of no exceptions—only that they do not admit of exceptions “in favor of inclination,” thus appearing to allow for exceptions to perfect duties in order to fulfill other duties. (Kant 1785: 73) In some passages, Kant actually endorses apparent exceptions to such duties, including the duty not to lie. (Cholbi 2009:28-31)
Still other Kantians propose that, textual evidence aside, none of the central elements of his ethical theory, including the Categorical Imperative, entail a commitment to rigorism and may even be inconsistent with it. (Singer 1954) Others argue that apparently reasonable exceptions to perfect duties can be made with minimal cost to Kant’s ethical theory. For example, Christine Korsgaard (1986) has argued that while duties such as the duty not to lie appear exceptionless if analyzed via one formulation of the Categorical Imperative, they are not exceptionless if analyzed via a different formulation.
Thanks to these efforts, crudely rigoristic interpretations of Kant’s ethics are no longer credible (even if they remain regrettably common in introductory treatments of Kant’s ethics: see Pojman and Fieser 2009: 138; Rachels 2012:128-32). Nevertheless, the rigorism objection retains some of its sting. For the fact that Kantianism is not committed to exceptionless perfect duties tells us little about the underlying motivation or logic of such exceptions. Under what conditions are exceptions to perfect duties permissible, and why? Unless Kantians provide at least a semblance of an answer to such questions, critics are likely to conclude that Kantianism provides insufficient guidance in just those situations where we most hope a theory can guide our choices, namely, when moral considerations speak in favor of violating a duty we otherwise take to be binding.
My aim here is to critique an approach, championed by Tamar Schapiro (2003, 2006) that appears to have the potential to provide a theoretical apparatus that identifies when, why, or how agents are morally permitted (or a fortiori, required) to act in opposition to an otherwise binding Kantian perfect duty. Schapiro’s approach, henceforth to be called the constitutive approach to Kantian rigorism, rests on the claim that moral requirements are subject to background conditions the presence of which lends these requirements their normative significance. Although there are probably many such conditions, including conditions concerning individual rationality or institutional justice (Rawls 1971:241-250; Cholbi 2002), Schapiro focuses on one such condition in particular, namely, others’ moral wrongdoing. The observance of moral requirements, she argues, can function as a background condition for other moral requirements to signify the forms of interaction they normally do. As a result, others’ failure to observe moral requirements can make it impossible to perform actions of otherwise impermissible kinds not in a literal sense, but as Schapiro argues, in the normative spirit they would possess if performed in Kant’s Kingdom of Ends. Thus, on the constitutive approach, others’ wrongdoing can excuse otherwise impermissible actions because these actions do not fully count as actions of the otherwise impermissible kinds. Hence, perfect duties whose fulfillment is otherwise mandatory are not mandatory in such ‘mitigating circumstances.’
Following Schapiro in using ‘defensive deception’—deceptive conduct that aims to thwart another’s wrongdoing—as a paradigm example of an action type that might be permissible in response to others’ wrongdoing, I argue that the constitutive approach fails to provide a credible picture of when and how we may violate otherwise binding perfect duties.
According to the constitutive approach, defensive deception is morally excusable because others’ intended wrongdoing makes it normatively incoherent to be honest to them in the spirit of Kant’s Kingdom of Ends. I argue that this approach lacks the resources to indicate where, if anywhere, the morally permissible responses (defensive deception aside) to thwart another person’s intended wrongdoing stop. Indeed, the constitutive approach seems to imply that any otherwise wrongful action would be permitted in such circumstances, an unacceptable practical consequence from a Kantian perspective. I conclude by briefly describing an alternative approach to the rigorism objection that avoids the objection I level at Schapiro’s approach while honoring the parameters she herself lays down for an adequate approach to the objection.
The Problem of Rigorism
As Schapiro observes, the strength of theories such as Kantianism resides in their plausibility as an ethics of the everyday, providing credible accounts of the common sense strictures against deception, failing to aid those in need, etc. Schapiro argues that Kantians confront the charge of rigorism because they, like many non-consequentialists, hold that ‘the actions prohibited by categorical moral rules are ‘wrong in themselves’—wrong in virtue of the features that make them count as the kinds of actions they are.’ (Schapiro 2006: 34) Hence, she seeks to outline ‘a strategy for holding onto the deontological thesis (that rightness and wrongness are intrinsic to actions) while avoiding rigorism.’ (Schapiro 2006:33)
Due to Kant’s controversial conclusion in his essay ‘On a Supposed Right to Lie,’ that it would be wrong to lie even in order to prevent a murder from occurring, such defensive deception has come to represent the ultimate challenge to Kantian rigorism. Presumably then, if Schapiro’s approach can answer the charge of rigorism and show that defensive deception is at least sometimes permissible, it can not only handle less troubling examples but also provide at least a sketch of a credible Kantian answer to the rigorism objection.
Again, doubts can be raised about how representative of Kant’s views his ‘On a Supposed Right to Lie’ is and about whether other Kantian texts support a rigoristic reading of Kant’s ethical theory. Nevertheless, Schapiro pinpoints why the charge of rigorism is a serious one for Kantians and other deontological theories—and would be even if Kant had not written the notorious lying essay. If an act’s deontic status hinges in some way on the kind of act it is, and acts that answer to certain general descriptions (lies, breakings of promises, etc.) are generally wrong, then acute problems arise when one and the same act falls under two relevant descriptions corresponding to such general kinds. If I have promised a crime boss that I will kill a local merchant, then my killing the merchant is both the keeping of a promise (which I am presumably morally required to do) and an unjust killing (which I am presumably morally prohibited from doing). So too in the infamous ‘murderer at the door’ scenario in Kant’s lying essay: Lying to the murderer violates an apparent moral requirement, but would simultaneously honor an apparent moral requirement: to protect innocents from aggression or violence. (Cholbi 2009; Sussman 2009) Resolving this problem seems to require Kantians to offer a principled way of identifying which of these descriptions is more fundamental or salient. As we shall see momentarily, a strength of Schapiro’s constitutive approach to the problem of rigorism is that it does not require identifying one act description as more fundamental or salient. For as Schapiro argues, wrongdoers can make it impossible to perform acts with the full ethical significance implied by certain descriptions those acts fall under. Wrongdoers in effect choose to make the honoring of an otherwise binding duty normatively incoherent.
Schapiro on Defensive Deception
Schapiro rightly observes that on a Kantian view, the wrongfulness of deception rests not on its presumed harmfulness, but on its ‘nature as a form of manipulation.’ (Schapiro 2006: 36) Deceptive conduct achieves its aim only when the deceived are kept in the dark about the fact of deception, thus making it impossible for one to consent to being deceived. Hence, deception amounts to an attempt to guide another’s will by inserting (assumedly) false beliefs into that person’s psychology, beliefs that the deceiver hopes will lead the deceived to behave in accordance with the deceiver’s wishes. In this regard, deception interferes with autonomy, or an agent’s right to govern her own behavior. If deception is intrinsically wrong in this way—that is, if deception attempts to interfere with another’s will in an effort to control that will—then (Schapiro asks) how ‘is it possible to hold on to the idea that the wrongfulness of deception stems from features that give it its deceptive character, while still allowing that deception can be made right, or at least not wrong, by the existence of mitigating circumstances?’ (Schapiro 2006: 38) How, for instance, could lying to a would be murderer be permissible (justified, excused, etc.) under this account of deception?
Schapiro’s answer is that moral rules do not simply prescribe certain sorts of action. Rather, compliance with such rules ‘constitutes’ certain forms of interaction. Moral rules require that particular sorts of interaction take place, but the sorts of interaction they require cannot be undertaken if particular ‘background conditions’ are not met: Actions falling under moral rules ‘must be both addressed and received in a certain spirit in order to count as having integrity as the kinds of actions they are.’ (Schapiro 2006: 33) I thus dub Schapiro’s approach to rigorism ‘constitutive’ because, on this view, the basis for making exceptions to perfect duties is that another individual’s wrongdoing undermines the conditions that make the fulfillment of such duties intelligible. Applied to deception, the constitutive approach entails that where these constitutive conditions that are not met, honesty is permissible not because dishonesty is justifiable, but rather because honesty, in the full-throated sense in which it operates absent others’ wrongful intentions, is impossible. In the case of defensive deception, the only sort of honesty possible is a ‘corrupted’ or ‘deficient’ form of honesty that allows for excusable deception.
What, then, are the background conditions which, when unmet, make honesty impossible, on Schapiro’s view? Honesty is characteristic of interactions governed by the reciprocity found in the Kingdom of Ends. That is, when individuals are honest with one another they evince the respect for others’ rational agency that would characterize interactions between rational co-equals. The would-be murderer, in contrast, has chosen an end that betokens a ‘basic lack of commitment’ to such a moral relationship. The murderer has thus adopted a normative stance that undermines ‘the integrity of the forms of interaction, like honesty, to which we ideally attach moral value.’ (Schapiro 2006: 54) The murderer cannot exit the moral community entirely, but by his actions he exemplifies more than a failure to comply with that community’s demands. Rather, he spurns the reciprocal or ‘colegislative’ standpoint that underlies such a community. As Schapiro expresses it:
The content of the murderer’s intention is relevant to the excuse [for your deceiving the murderer], because it reveals something about the orientation of the murderer’s will that is relevant to whether or not you can be honest with him in both letter and spirit. More specifically, his adoption of an end that is blatantly at odds with the ideal of reciprocity in a Kingdom of Ends shows that he is not in a position to take up your honesty in the spirit in which it would be offered, namely, a spirit of reciprocity. His choice of that end makes it appropriate for you to regard him as having refused to participate in the shared activity of which honesty is a part, the activity from which honesty derives its moral value. This is so despite the fact that on the surface, the murderer is treating you with respect. The fact that he intends to murder an innocent person shows that whatever game he is playing, it is not the colegislation game, even though in certain respects it may simulate the colegislation game. Now if honesty is conceived as a form of interaction rather than simply a form of action, then the murderer’s lack of reciprocity will have a bearing on your own capacity to be honest with him. If the murderer has refused to play the colegislation game, then whatever demand he is issuing, it is not the demand to share in thought and action, the demand to which honesty is properly addressed. Hence it is impossible for you to be honest with him in the spirit proper to honesty. (Schapiro 2006: 52)
Schapiro is therefore not defending the strange claim that it would be impossible to tell the would-be murderer the truth when asked for the whereabouts of his intended victim. Her point is instead that the murderer has, by his violent intentions, shown that he is not a participant in the ‘game’ of reciprocity, shared deliberation, etc. in which honesty can be legitimately demanded. As a result, one could honor the letter of honesty in such a situation, but to do so would not be to honor its spirit. In such a context, the only possible exercise of honesty would be ‘merely aspirational,’ a form of improvisation in which one seeks ‘a surrogate form of action that will come as close as possible to realizing the ideal of human relations that animates your normative conception of honesty.’ (Schapiro 2006: 55)
Schapiro says less about how other circumstances besides others’ wrongdoing might justify a violation of a perfect duty. Yet we can see how the constitutive approach would proceed with respect to other such circumstances. When the background conditions for an act fulfilling a given perfect duty are not met, then acting on such a duty would not amount to performing an act of that kind or in the intended spirit of the Kingdom of Ends. In other words, the persons affected by the fulfillment of the duty do not, for one reason or another, fully occupy the reciprocal or co-legislative standpoint. Hence, acting in opposition to such a duty can be excused in those circumstances. Thus, an otherwise binding perfect duty can admit exceptions without contradicting the thesis that violations of such a duty are wrong ‘in themselves.’
The ‘Anything Goes’ Objection
Schapiro takes her constitutive approach to answer the charge of rigorism while still honoring the conviction that otherwise impermissible actions, such as deception, are nevertheless wrong in themselves. Lying to a would-be murderer, for example, is permissible because the demands of honesty (or conversely, the prohibition on deception) cannot be fully honored in the circumstances created by the would-be murderer’s violent intentions.
However, Schapiro’s constitutive approach must walk a fine line concerning the moral implications of the murderer’s rejection of the reciprocal standpoint. On one hand, Schapiro describes the murderer’s intentions not simply as failing to comply with the demands of that standpoint, but as a refusal or betrayal of that standpoint. (Schapiro offers no criterion with which to distinguish failures to comply with the moral demands of a given standpoint from betrayals of that standpoint, but I will not pursue that issue here.) The murderer’s actions are more than an offense against the norms governing that standpoint; he is not imperfectly complying with those norms, but is in fact rejecting them or denying their normative authority. We may be tempted to infer that because of this betrayal, all bets are off—that we no longer have any obligations to the murderer at all (in which case, there are no moral limits to what we may do in order to thwart his murderous aims). But Schapiro blocks this inference, claiming that even if we may manipulate the murderer in ways that would be otherwise wrong, our efforts at manipulation are still ‘constrained,’ since the murderer has neither ‘put himself’ ‘beyond the pale’ of the moral community,’ nor has he ‘exited’ that community. (Schapiro 2006: 54) Inasmuch as Schapiro intends to defend a Kantian view, she is right to reject this inference, since Kant, and most contemporary Kantians, do not suppose that even the most serious wrongdoing or evil establishes a permission to do whatever we please to wrongdoers.
On the other hand, Schapiro emphasizes that the proper basis for deceiving the murderer is an excuse rather than a justification, because (in her estimation) it is only when conceived as an excuse that such an act of deception can honor the Kantian thesis that deception is wrong in itself. For to excuse an act of deception is nevertheless to categorize the act as deception and to implicitly refer back to a moral standard disallowing deception. In other words, to be excused for one’s deception is still to have engaged in deception. (Schapiro 2006: 41) In contrast, if defensive deception is justified rather than excused, then what makes such deception permissible is some feature of the act’s context that intervenes to modify the act itself, so that the act’s morally relevant description—the description that makes its performance morally justified—is not ‘X is an act of deception’ but, for example, ‘X is an act of rescue’ or ‘X is an act to counter aggression.’
Yet it is not at all clear what, on her approach, serves to block the inference that anything whatsoever may be done to thwart a would-be murderer. Put differently, it is unclear what entitles her to land on precisely the normative prescription that lying is permissible when it would thwart wrongdoing such as the violent intentions of the murderer. Schapiro contends that not all bets are off with respect to how we respond to the murderer’s aggression. Yet the murderer’s plans indicate that he is not abiding by the constitutive rules of interaction within the Kingdom of Ends, rules that give honesty its very point, thus making defensive deception excusable. But in not playing the colegislation game, the murderer is also not abiding by rules of interaction that give many other Kantian restrictions their point as well. If the murderer has made the only possible form of honesty a corrupted form of honesty, then so too has he made possible only corrupted forms of other behaviors ordinarily required by Kantian morality. And there are after all other possible defensive actions, many of which are at least as likely to thwart the murderer’s intentions as deception would. We could kill the murderer, threaten him, or buy him off with promises we have no intention of honoring. But the constitutive approach then proves too much, for we could equally well excuse ‘corrupted’ killing, ‘corrupted’ threatening, or ‘corrupted’ false promising on the grounds that the murderer has made genuine non-violence, non-interference, or fidelity—of the kinds that are honored in the Kingdom of Ends—impossible, at least in their genuine spirit. As Schapiro writes, ‘if honesty is conceived as a form of interaction rather than simply a form of action, then the murderer’s lack of reciprocity will have a bearing on your own capacity to be honest with him.’ But similarly, if we accept Schapiro’s claim that adhering to moral rules is intelligible only when a background condition of reciprocal interaction is met, then the murderer’s lack of reciprocity also has a bearing on our capacity to respect the murderer’s autonomy in the various other ways that Kantianism ordinarily requires. If I kill the murderer in order to stop the murderer from murdering, may I reasonably claim that I did not wrongfully kill him—since his actions have made it impossible for me to refrain from murdering him in the spirit proper to not murdering—, but only excusably killed him in order to prevent him from murdering?
This, then, is the ‘anything goes’ objection: The constitutive approach does not appear to have the resources necessary to explain why lying, as opposed to any more dubious defensive action, is uniquely permissible. Insofar as the constitutive approach is supposed to fill out the details of a Kantian response to the rigorism objection, this is worrisome, for it allows that any perfect duty may be violated in order to thwart others’ wrongdoing. Despite Schapiro’s claim that the murderer has not lost his status as a moral subject, the constitutive approach appears to eviscerate that status by nullifying all our obligations to him.
Schapiro’s constitutive approach to defensive deception thus has the unfortunate consequence that others’ moral wrongdoing appears to justify a form of moral carte blanche on our part. However, Schapiro might have the resources to avoid this ‘anything goes’ objection: Recall her remark that although the would-be murderer has made it impossible for others to be honest in the spirit exemplified in the Kingdom of Ends, it is nevertheless possible to act in a way that aspires to realize ‘the ideal of human relations that animates [our] normative conception of honesty.’ Admittedly, Schapiro says little about how we are to be guided by this ideal. Nevertheless, perhaps the ideal of human relations that ‘animates our normative conception of honesty’ could serve as a constraint on the permissible responses to the murderer’s violent intentions. Moreover, this constraint might function to rule out morally dubious ‘defensive’ responses to the murderer (murdering the murderer, say) while honing in on deception as the uniquely justifiable defensive response. If so, then my charge that the constitutive approach results in too many permissible responses to the murderer could be deflected on the grounds that the ideal of the Kingdom of Ends constrains which responses to others’ wrongdoing are morally permissible.
Again, Schapiro says little about how responses to the murderer honor, or fail to honor, the ideal of the Kingdom of Ends. However, the general notion that our responses to others’ wrongdoing are subject to moral constraints of their own has a long deontological heritage. We see, for example, in discussions of self-defense and of the doctrine of double effect appeals to norms of proportionality intended to ensure that agents who act in the pursuit of a morally defensible end do so with maximal respect for moral rules such as not violating rights, not causing harm, etc. Hence, Schapiro’s constitutive approach appears to be on solid ground in holding that deviations from the ideal of the Kingdom of Ends might nevertheless be constrained by that ideal.
Yet even here, the constitutive approach cannot duck the ‘anything goes’ objection. Suppose we grant that some second-order moral norm—a norm that might imply that though defensive deception is morally permissible in response to the murder’s intentions, defensive killing would not be—would bind us in the Kingdom of Ends. In order for such a second-order norm to answer this objection, it would have to be true that the murderer, for example, does not betray the Kingdom of Ends in toto. It does after all seem possible to betray some component of the Kingdom of Ends without betraying this ideal in its entirety. If so, then perhaps killing the murderer is ruled out (a) because the morally permissible responses to the murderer’s intentions are proportional to the extent of his betrayal, and (b) killing the murderer is disproportionate to his less-than-total betrayal of the Kingdom of Ends. But granting this still does not solve the ‘anything goes’ problem. For recall that, on Schapiro’s view, the murderer is ‘blatantly at odds with the ideal of reciprocity in a Kingdom of Ends’; his conduct amounts to a ‘basic lack of commitment’ to this ideal; he is refusing to participate in the shared activity of ‘colegislation’ in which honesty is indispensable; etc. Schapiro thus claims that the murderer has betrayed the ideal of the Kingdom of Ends with his violent intentions. Since this is presumably a betrayal of all the norms associated with the Kingdom of Ends, his wrongdoing, according to the constitutive approach, makes otherwise binding moral precepts, such as the precept not to lie, inapplicable. But if so, then there is no reason to think that second-order norms that might rule out, e.g., killing the murderer, are applicable either. Hence, if lying to the murderer is excusable on the grounds that honesty, in the sense demanded by the Kingdom of Ends, cannot be provided to the murderer, then so too can killing the murderer be excused on the grounds that, though there is a second-order norm ruling out such killing, the murderer’s intentions have made it impossible for that second-order norm to be honored in the sense demanded by the Kingdom of Ends. Of course, in Kant’s Kingdom of Ends, such second-order norms would never need be invoked, since in such a kingdom, individuals not murdering (and in general, agents treating one another as ends in themselves and not as mere means) is itself the norm. Thus, it becomes murky exactly how norms governing our responses to wrongdoing could be based on an ideal of interaction where such wrongdoing does not occur. Yet even if this could be made sense of, the constitutive approach lacks the internal resources to show why we must abide by these second-order norms when responding to others’ wrongdoing. Abiding by such norms is probably morally permissible or admirable, i.e., we might think an agent who merely lies to the murderer shows a morally admirable restraint or evinces respect for the ideal of the Kingdom of Ends. But in order to answer the ‘anything goes’ objection, the constitutive approach needs more than this. It needs to show, in a way that is not simply ad hoc, that we are obligated to follow second-order norms through which particular responses (e.g., defensive deception) turn out to be uniquely justified or obligatory because other responses (e.g., defensive killing) are wrong. But the constitutive approach cannot do this.
A final note on this objection: Some may be untroubled by this objection because they do not find it morally worrisome that the murderer’s violent intentions might permit us not only to lie, but also to kill the murderer, say. Consequentialists would not likely be troubled by this, but for them, the ‘anything goes’ objection would not be forceful in the first place. Yet the worry the constitutive approach presents for those sympathetic to the Kantian project extends beyond the specific responses to the murderer that it licenses. Perhaps killing the murderer is morally permissible on a Kantian view. Yet a conscientious Kantian must recognize some limit to how the murderer may be treated. For instance, in his defense of capital punishment, Kant argues that the dignity and autonomy of the criminal morally demands his death, but that in putting the criminal to death, we must nevertheless treat him in a humane and respectful fashion consistent with his dignity. There is no appeal here to the Lockean notion that criminal behavior results in the criminal’s forfeiture of his moral status such that our duties toward him are vacated. (Murphy 1970; Scheid 1983; Byrd 1989; Fleishacker 1992; Hill 1997; Holtman 1997) By analogy, then, perhaps torturing the would be murderer is a morally unjustified response. But again, the ‘anything goes’ objection has less to do with precisely which responses to the murderer are morally licit as it does with the constitutive approach’s inability to demarcate—and to rationally ground—the boundary between the morally licit and the morally illicit responses are, and when combined with Schapiro’s claim that the murderer has betrayed the ideal of the Kingdom of Ends, this seems to imply that anything goes. Thus, the constitutive approach answers the rigorism objection but only by providing agents far more latitude in overlooking otherwise binding perfect duties than seems plausible.
Conclusion: Maxims and the Challenge of Rigorism
For Kantians, answering the charge of rigorism requires showing how it is possible for there to be ‘intuitively reasonable exceptions’ to perfect duties ‘without abandoning the basic Kantian point of view.’ (Hill 1992: 201) I have argued that the constitutive approach is counterintuitive and un-Kantian in allowing too many exceptions to perfect duties, in effect abandoning the Kantian point of view. Still, the shortcomings of the constitutive approach are instructive. Merely acknowledging that Kantianism permits exceptions to perfect duties is not enough to provide a satisfying reply to the charge of rigorism. What is needed is guidance as to when, why, or how we are permitted not to fulfill otherwise authoritative perfect duties. The challenge here is to show how the Categorical Imperative, the supremely authoritative principle of morality and the source of perfect duties, can permit exceptions to these duties in response to circumstances including but not limited to the wrongdoing of others. I hope to have shown that the constitutive approach neither honors the authority of the Categorical Imperative, in effect denying its relevance to such circumstances, nor yields plausible conclusions as to what we morally ought to do in such circumstances.
However, the flaws in the constitutive approach do not require Kantians to revert to strategies wherein one or another act description is identified as more fundamental or salient in a given context. Indeed, such strategies wrongly assume that Kantianism is a theory that tests acts under some generic description. Rather, Kantianism tests particular acts by way of maxims, principles that express an agent’s reasons for action and incorporate not only an act description but also the agent’s understanding of the relevant context of her action as well as the end she seeks to bring about through acting on that maxim. (Glasgow 2011:153) In other words, maxims themselves supply the morally relevant description of what an agent does because they embody the agent’s reasons for acting as she does.
This observation suggests that a more plausible avenue for Kantians to answer the rigorism objection is to argue that once we fully comprehend how that Categorical Imperative is supposed to operate in terms of how maxims are to be selected, tested, etc., it not only permits exceptions to perfect duties but is in fact sufficiently flexible to provide guidance concerning when and how exceptions to such duties ought to be acknowledged, including how to resolve apparent conflicts of perfect duties. For those Kantians Schapiro refers to advocates of the ‘fine print’ approach to Kantian ethics, perfect duties correlate with prima facie moral rules applicable in ordinary circumstances but too coarse to apply universally. So whereas a maxim M allowing deception ordinarily fails Kant’s universalization test (or acting upon such a maxim would treat others as mere means, etc.), a maxim M embedded within a conditional antecedent such as ‘if deception would prevent the death of an innocent person’ may not fail such a test. The salient task is therefore to identify those maxims that, when subjected to the Categorical Imperative’s demands, generate defensible exceptions to perfect duties in response to such phenomena as others’ wrongdoing. Note that such an approach would not be vulnerable to an equivalent ‘anything goes’ objection: For some maxims that, for instance, enjoin lying would not pass muster under the Categorical Imperative precisely because of the maxim’s other features. For instance, a maxim to lie to a putative murderer in order to keep a person alive because you plan to use the person as an unwitting subject of a medical experiment would presumably fail the Categorical Imperative.
Obviously, I cannot provide a full defense of this ‘fine print’ approach here. Yet this approach has the obvious advantage of respecting the basic Kantian point of view, without embarking on a potentially fruitless attempt to prioritize among act types. Moreover, the constraint Schapiro proposes, that an adequate Kantian answer to the rigorism objection must honor the notion that the wrong making features of acts stem from the nature of the acts themselves, is also honored by the ‘fine print’ approach. A wrongful lie, for instance, will be wrong because a maxim incorporating it cannot be consistently universalized. In this respect, the fact that an act is a lie plays an ineliminable role in explaining its wrongfulness when it is in fact wrong. My proposed approach thus appears to satisfy the desiderata, articulated by Schapiro, for an adequate response to the rigorism without succumbing to the nihilistic conclusion suggested by the ‘anything goes’ objection.