2. Traditional philosophical approaches to moral dilemmas

A moral dilemma, as philosophers have traditionally understood it, is a situation where an agent has moral reasons to perform two different actions but it is impossible for her to perform both. The philosophical literature abounds with tragic examples: Sophie is forced by a diabolical Nazi to save one but not both of her children; Gauguin must choose between fulfilling his familial responsibilities and expressing his artistic genius; Sartre’s student must decide whether to join the resistance or care for his ailing elderly mother; Agamemnon is torn between saving his daughter and leading the Greek troops to Troy.

Until relatively recently, however, genuine moral dilemmas were thought not to actually exist. Because it was assumed that a fully comprehensive moral theory should be able to solve all moral conflicts, dilemmas were seen as problems to be explained away on pain of threatening the legitimacy of a moral theory as a whole. Philosophical interest in moral dilemmas was primarily insofar as they allowed theorists to use their pet theory’s ability (or competitors’ inability) to solve them to confirm or defend their theory. The assumption that a satisfactory moral theory should not permit dilemmas remained more or less unchallenged until the middle of the last century, until the question of whether genuine moral dilemmas exist rose to prominence as one of the central metaethical debates in the 1980s and 1990s.

One of the central motivations in this debate for denying the possibility of moral dilemmas was a concern to avoid logical inconsistency. Were genuine moral dilemmas to exist, agents could find themselves in a situation where they are simultaneously obligated to ϕ and not to ϕ: Sophie would be simultaneously obligated to save, and not to save, each of her children. Another key motivation for rejecting dilemmas was the conviction that a moral theory must be uniquely action-guiding. If the fundamental purpose of morality is supposed to be to tell us what to do, genuine dilemmas threaten this project because they present us with situations where morality fails to generate practical advice. Morality’s task is not merely to make sense of the easy parts of our moral lives, the thought goes; we need a theory that does not abandon us to coin-flipping when the going gets tough. Many opponents of dilemmas were also motivated on realist grounds – the thought that there could be a situation where there is no one right course of action was seen to threaten anti-realism or even nihilism.

Proponents of the existence of genuine moral dilemmas, on the other hand, tended to focus on the phenomenology of certain moral experiences. When agents face moral dilemmas, complex emotions such as guilt, conflictedness, remorse or anguish almost inevitably ensue, and these emotions are an important and necessary aspect of the moral experience. It is also commonly believed that dilemmas tend to generate compensatory duties in their aftermath: no matter which course of action an agent chooses in a dilemmatic situation, she will be left with moral obligations to do what she can to make things better. Proponents of moral dilemmas capture these related intuitions – that in the aftermath of a dilemmatic situation an agent will be left with certain morally appropriate (or possibly even morally required) emotions, as well as certain compensatory duties – with the term moral residue. Sophie cannot skip off into the sunset no matter which of her children she chooses to save, the intuition goes; this is precisely what is so soul-crushingly tragic about her circumstances. To claim that there is no genuine dilemma here, that Sophie has open to her an action that is ultimately morally justifiable, is to profoundly misrepresent the moral contours of the situation. There are times in life where an agent can feel like she is faced with the impossible, where morality is presenting her with competing incompatible and maybe even incommensurable demands. It is only if we permit the existence of genuine moral dilemmas that we can explain the interrelated phenomena of moral residue.

It is important that we define moral dilemmas carefully. A genuine moral dilemma is not merely a situation where an agent must choose between a number of morally permissible actions – that is, properly speaking, just a description of everyday life. An agent faced with a moral dilemma is not someone in the position of Buridan’s ass, faced with reasons to do incompatible things all of which are themselves morally permissible. Rather, the thought goes, an agent in a moral dilemma is faced with reasons to do incompatible things, and no matter which course of action is taken the agent is left having done something immoral. Choosing one horn of a moral dilemma rules out the possibility of choosing the other, this latter failure to act is itself morally culpable, and the rules of morality are silent on which course of action should be prioritized. We must be careful to capture the sense of moral failure, or of having been abandoned by morality, that is characteristic of the phenomena in question here.

Since its intellectual heyday in the 1980s and 1990s, the question of whether genuine moral dilemmas exist has more or less gone out of vogue. I am here advocating a revival of this debate, one that takes on board the intervening philosophical work in the decades since. Returning to this debate, whilst rationally reconstructing Kant for radical purposes, I will argue that Kantianism both can and should permit the existence of moral dilemmas.

3. Kant on moral dilemmas

In a Kantian framework, the question of whether there can be genuine moral dilemmas is understood in terms of the question of whether moral duties can come into conflict with one another. Those approaching Kant with this question tend to start and finish with his familiar ought implies can, concluding rather too quickly that moral dilemmas are impossible in a Kantian system. After all, the phenomena that dilemmas are meant to capture include feelings of morality facing us with the impossible, of being torn between competing incompatible and maybe even incommensurable moral commitments. When Kant’s dictum is read as a blatant insistence that the moral law cannot demand the impossible from its subject – that it cannot be that we ought to do something that is impossible for us to do – this is seen as definitive proof of the inability of the Kantian system to accommodate the phenomena in question.

But this turns out to be one of those rare instances where almost everyone in philosophy has gotten things wrong. Robert Stern challenges the conventional attribution of the ought implies can principle to Kant, pointing out that Kant actually uses this principle in surprisingly few passages, and explicitly discusses or defends it in literally none. While Kant did clearly accept some version of ought implies can, Stern argues that Kant’s version of this principle is considerably weaker than many have supposed. A strong reading of ought implies can would hold that ‘nothing can be right that we are incapable of achieving’, or that ‘no act can be right (rather than just obligatory for us) unless we are able to perform it’ (Stern Reference Stern2004: 59, 57). But all Kant argues is that ‘we cannot be obliged to do what is right unless we are capable of acting in that way’, and that ‘the moral law only has its status of being obligatory for us because we are able to act upon it, and that we can thus only explain this obligatoriness by accepting certain claims about our capacities and their conditions’ (Stern Reference Stern2004: 59, 57). ‘What Kant is focused on here is not the moral law as such (so to speak), but how the moral law relates to us, as something that commands us … that tells us what to do unconditionally … that has a certain authority over us’ (Stern Reference Stern2004: 56) None of this, then, rules out the possibility of moral dilemmas.

Recognizing that Kant’s understanding of ought implies can does not necessarily rule out the possibility of moral dilemmas yields a Kantian picture on conflicting duties that is considerably more complex than many have supposed. Turning to the primary texts confirms that, contra prevailing assumptions, Kant can permit the existence of moral dilemmas. The primary passage in which Kant addresses the possibility of conflicting duties is found in the Metaphysics of Morals:

It is, at first glance, perhaps natural to read in Kant’s claim that ‘a collision of duties … is inconceivable’ a denial of the possibility that moral duties can conflict. But, as emphasized by Allen Wood, reading further in this passage makes clear that what Kant really means by this claim of inconceivability is that conflicting duties cannot cancel each other out (Wood Reference Wood2008: 163).Footnote ¹

Proponents of moral dilemmas, note, do not generally claim that conflicting moral duties cancel each other out. Their claim, rather, tends to be that conflicting duties give rise to moral residue such as obligatory feelings of guilt, or to compensatory duties to attempt to rectify what has gone wrong. Though Kant himself admittedly does not explicitly work through these possibilities, it actually follows from the claim that conflicting duties cannot cancel each other out that situations of this kind might result in some sort of moral residue. The thought that duties retain the full force of their pull on us even when they conflict is precisely the phenomenon that proponents of moral dilemmas are trying to capture. It is actually the converse claim – that conflicting duties do cancel each other out – that would rule out the possibility of moral residue, and, again, this converse claim is not one Kant makes. Instead, while he does not think duties can conflict (in the sense of cancelling each other out), what Kant does believe can conflict are obligating reasons – the grounds or reasons why we have a duty. But, again, when an agent’s obligating reasons conflict, this will, in at least some circumstances, give rise to moral dilemmas understood in the sense of a situation that generates moral residue. With Wood, I am inclined to believe that Kant is not the stubborn refuser of the existence of moral dilemmas that many have painted him to be – or, at the very least, that Kantians need not be.

Rather than failing to admit the possibility of moral dilemmas, a larger cause for concern here, according to Wood, is Kant’s failure to address a variety of other issues in the moral dilemmas debate. There is, for one, the point of deontic logic: if I am obligated to ϕ then it follows that it is false that I am not obligated to ϕ, but it does not follow that I am not obligated not to ϕ. Or, at the very least, not being obligated to not ϕ would require a separate deontic principle unlikely to be accepted by defenders of moral dilemmas. Kant is silent about what to make of this issue, but, again, does not deny in the above passage that such conflicts could arise. When Kant allows that obligating reasons can conflict, he is admitting that the reason one has a duty to ϕ might conflict with a different reason one has a duty not to ϕ. The only assistance he gives us to arbitrate such conflicts is that the stronger obligating reason is to prevail (fortior obligandi ratio vincit). Kant does provide some additional practical guidance in how to think through which obligating reasons are stronger than others – chiefly that wide or imperfect duties are always trumped by narrow or perfect duties – but this meagre guidance leaves many questions unanswered. Wood admits that Kant neither asks nor answers a host of important questions here:

Kant’s silence on these and other related issues has led many to infer that he denies the possibility of moral dilemmas, but it is important to note that this impossibility does not actually follow from any of Kant’s ethical principles. The reason for Kant’s silence here, Wood insists, is that the above sorts of questions would have been considered to have been ‘casuistical’ and thus not appropriate for discussion in his more abstract metaphysics of morals.

What is important for present purposes is that what Kant actually means when he says that ‘a collision of duties and obligations is inconceivable’ does not rule out the moral residue that proponents of moral dilemmas have in mind. Kant is merely ruling out the much narrower possibility that conflicting duties cancel each other out. Defenders of moral dilemmas are thus wrong to strawman Kant as stubbornly refusing to admit the possibility of their existence. And while we might have liked Kant himself to have given us considerably more guidance on what to do when moral dilemmas arise, there is nothing in the Kantian system to forbid those in his wake from picking such questions up.

Unfortunately, most prominent Kantians’ contributions to the old moral dilemmas debate have failed to properly appreciate this point about the possibility of accommodating moral residue in a Kantian framework. ‘A moral system is inconsistent only if it allows the possibility that, without any wrongdoing on his part, a man may find himself in a situation in which he can only escape doing one wrong by doing other, that is, only if it allows the possibility of perplexity simpliciter’, insists Alan Donagan, and this perplexity or inconsistency is precisely what is ruled out by ‘rationalist or “natural law” systems of morality, like those of … Kant’ (Donagan Reference Donagan1977: 145, 148). Donagan’s mistake here is representative of that of Kantians in general, both uncritically attributing to Kant a too-strong reading of ought implies can, and taking the first part of the passage where Kant discusses the ‘inconceivability’ of a collision of duties out of context and ignoring his immediate clarifications about what this inconceivability actually consists in.

Donagan does go on to admit that Kantian morality can put agents in impossible situations, but he insists that this will occur only in an extremely constrained set of circumstances. Moral dilemmas can genuinely arise, Donagan argues, but only when an agent has done something to bring the situation upon herself – by making two promises that she knows to be incompatible, for example. The presence of a moral dilemma is thus an indication that an agent has committed a wrongdoing; for Donagan, Kantian morality’s demands would never place someone in such a situation on their own. The role of morality is to tell us what we ought to do, but if we insist on violating morality’s commandments then we should perhaps not be surprised when we find ourselves in a mess.

4. Dilemmas, gaps and moral failure

The logical entailment of Donagan’s contention that moral dilemmas only ever arise in Kantian morality as a result of an agent’s wrongdoing, however, is that the agent who finds herself in dilemmatic conditions must have done something to deserve it. An agent with clean hands would never find herself morally compromised in such a way, the thought goes, and that she has brought this upon herself thus threatens only our assessment of her individual character, not the legitimacy of Kantian morality as a whole.

Suggesting at least some awareness of the egregiously privileged victim-blaming lurking here, Tom Hill takes issue with Donagan’s claim that moral dilemmas only ever arise from an agent’s wrongdoing (Hill Reference Hill2002: 362ff.). Even when an agent has done something immoral, Hill reminds us, the dignity of their humanity requires that we respect them as autonomous agents capable of choosing to do otherwise in the future. Were we to allow that an agent could do something that would put them in a situation where the moral law no longer had any purchase on them, he argues, we would have given up on Kant’s conception of the value of the person.

Genuine moral dilemmas, according to Hill, should be understood as situations where there is no morally permissible course of action, where every option is morally impermissible and yet one or another must be taken. Hill insists that such dilemmas, understood as situations where ‘practical reason, after due reflection, unequivocally directs you … to refuse to take any of your available options, including doing nothing’, make no sense in a Kantian system (Hill Reference Hill2002: 375). What could arise, Hill says, are situations of what he calls apparent moral dilemmas – situations where an agent is torn between incompatible courses of action by competing moral duties. In such situations, it is important to remember that each course of action is itself not contrary to duty because each horn of the apparent dilemma is, itself, morally permissible. And therefore, while we might commonly characterize whatever an agent ultimately decides to do in an apparent moral dilemma as the ‘lesser of two evils’, Hill insists that this characterization is inaccurate. Instead, if an agent has acted from duty then their action is definitionally permitted by duty, thus so long as the action is done from good motives, with a good will, then such an action cannot properly be called ‘evil’ at all.

Hill characterizes this refusal to admit the possibility of genuine moral dilemmas as an expression of Kantian respect for persons:

The reason Kantians should deny the possibility of genuine moral dilemmas, according to Hill, is that permitting them would be tantamount to giving up on people who have done immoral things. And this is something the moral law would never permit us to do.

What Hill thinks the Kantian system does allow, however, are what he calls moral gaps. If a genuine moral dilemma is a situation where there is no morally permissible course of action, a moral gap is a situation where the moral law provides no (or not enough) practical guidance. Strikingly, it is the very same feature of Kantian morality that is supposed to rule out the possibility of dilemmas that raises the possibility of gaps. This is because the Formula of Humanity’s requirement to unfailingly respect people in virtue of their dignity – their unconditional and incomparable value – will inevitably generate situations where we are drawn powerfully in different directions.

In such cases, the moral law cannot tell us definitively what to do. Hill appreciates the tragedy of moral gaps, but insists that it is not a defect of Kantian theory that it permits them. A theory that failed to permit moral gaps, he argues, would ignore or distort certain important features of our moral experience – chiefly, the deeply felt value conflicts that are often what lead us to feel the need for moral theorizing in the first place.

Gaps also allow Kantianism to explain why we should do everything in our power to avoid getting into tragic situations where morality can no longer function properly as our guide, for they accommodate the intuition that ‘we have strong reason in advance to do everything that we permissibly can to avoid later being forced to do what we have reason not to do’ (Hill Reference Hill2002: 384). This is an intuition first put into words in the moral dilemmas literature by Ruth Barcan Marcus, who argued that the desire to avoid moral residue can have a ‘dynamic force’, and that properly appreciating the tragedy of situations where morality cannot offer unique action-guidance should commit us to ‘a second-order regulative principle: that as rational agents with some control of our lives and institutions we ought to conduct our lives and arrange our institutions so as to minimize [such] predicaments’ (Marcus Reference Marcus1980: 121). It is worth noting that Marcus’ interest in this intuition was primarily metaethical, raising the point only to bolster her arguments in favour of the existence of genuine moral dilemmas. The broader implications of thinking that there can be compelling reasons to address the social conditions that give rise to conflicts like these would go on to be largely unexplored in the literature for too long. We will return in a moment to a broader analysis of the social and political implications of moral dilemmas – one that recognizes that these conflicts are often better understood as wires in Frye’s birdcage than as random instances of bad luck.

But first there is the question of whether Hill’s gaps are a suitable replacement for genuine moral dilemmas – that is, whether they can capture everything of phenomenological significance to those concerned about moral residue. Gaps do capture one important aspect of moral residue: the recognition that there will be tragic instances where the moral law cannot be our guide, where no amount of virtue nor steadfast commitment to being a good person who does the right thing can save us from the possibility of finding ourselves in a situation where morality cannot tell us the uniquely best thing to do. This captures the sense of abandonment an agent faced with a dilemma might feel, the sense that having spent your entire life committed to a set of principles is sometimes not enough to insulate you from the possibility of being left on your own by these principles precisely when you need their guidance the most. Of course, agents in dilemmatic situations have not been abandoned entirely. The moral law still presents them with a set of permissible options; it just cannot help decide between them. But while it might be acceptable in everyday life to have decisions about which of a permissible set of options to choose from left to judgement or discernment, it is in precisely the difficult cases like the ones presented by dilemmatic situations that we might reasonably have wanted morality to have given us more. Hill’s gaps explain this aspect of moral residue, at the very least.

To be faced with a genuine moral dilemma is to be faced with the uncanny recognition that you have had forced upon you an action that is simultaneously permissible and impermissible, of knowing in your bones that you must do a thing and yet knowing with equal certainty that doing this thing will make it impossible for you to do another thing that you also must do. Too many in the moral dilemmas literature have wanted to resist this uncanniness, to tidy it all up or explain it away, but properly capturing the phenomenology insists on it. ‘Theories that satisfy ideals of neatness, completeness, and elegance may serve some purposes less well than theories that unabashedly highlight the moral conflicts we experience’, Hill recognizes, insisting that ‘[g]aps may reflect important features of our moral experience that closure would distort’ (Hill Reference Hill2002: 382). But while Hill sees the importance of letting go of the need to contain all of human experience within confines of too-narrow logical formalism, he does not properly appreciate the absurdity of dilemmatic situations. There are other crucial aspects of moral residue that his gaps fail to capture.

Hill means gaps to comfort the agent caught in a dilemma, to reassure her that while her situation is tragic, whatever she ultimately chooses to do is not impermissible so long as it is done from a good will. But it should be cold comfort to Sophie that choosing her son was permissible; the whole point here is that not choosing her daughter was impermissible! Gaps fail to capture the impossibility of what is being asked of an agent in a dilemmatic situation. It is not enough to reassure an agent that each horn of their dilemma is permissible; the tragedy here is that each horn is morally required and yet performing one rules out the possibility of performing the other. Sophie is not in the position of Buridan’s ass, caught between two equally attractive piles of straw that will be perfectly fine if she chooses the other. She is torn between abandoning one of her children or abandoning the other, and no amount of tying oneself in philosophical knots can or should finesse the situation. Hill wants to console Sophie that choosing her son over her daughter is not the lesser of two evils, but this is simply false – abandoning one of your children to their death is evil, even if fate has made it impossible for you to do otherwise. Thus, while Hill’s gaps might be able to capture some important phenomenological aspects of moral residue, there are others captured only by admitting the possibility of genuine moral dilemmas, situations where what the moral law is requiring of someone is outside the bounds of what can be made sense of within the confines of logical possibility. Whether or not Kant himself saw this, Kantians must.

5. Oppression theorists on moral dilemmas

This is something that oppression theorists have seen all too well. It is striking but ultimately unsurprising that Marcus’ insight about the importance of doing what we can to ‘arrange our lives and institutions’ to avoid being caught in moral dilemmas remained almost entirely unexplored in the old moral dilemmas literature. When participants in these debates did bother to attend to this concern, it was usually only ever to fret about whether if this second-order principle were too stringent it might prevent people from entering important leadership roles in political life, because of the inevitability of such actors finding themselves with ‘dirty hands’. Also striking in this literature was which cases were given central focus: virtually every moral dilemma taken to be representative involved an agent facing a dilemmatic situation because of geopolitical forces that would have been considered to be well outside the scope of everyday life. Both Sophie and Sartre’s student were thrown into their impossible situations by the Nazi regime, and when the stuff of Nazis was not fantastical enough for these debates, they turned to the tragedies of Greek myth.

Indeed, it seems to have been a central assumption in these debates that genuine moral dilemmas will be exceptionally rare. Hill confirms this assumption explicitly, assuring us that ‘we are not forced every day to face choices’ like the unfortunates caught in moral dilemmas, and justifying the acceptability of gaps by explaining that, ‘[i]f our best theory abandons us only in these extreme cases, then perhaps we can live with this’ (Hill Reference Hill2002: 382). (The only real exception to this tendency to treat dilemmatic situations as occurring only in the most extreme of circumstances was Bernard Williams’ discussion of Gauguin (Williams Reference Williams1981: 38ff.). The hypothetical conundrum over whether Gauguin ought to have abandoned his family for the sake of his art is admittedly considerably more banal, but noticing how live and reasonable an option Williams took this abandonment to be tells us everything we need to know about the blithe indifference of this debate’s participants to the struggles of anyone other than those occupying extremely privileged social positions.)

It matters a great deal to the urgency of the discussion of moral dilemmas which cases we take to be representative. Ultimately, it is not a coincidence that theorists have tended to focus on the cases they have, for it is a mark of privilege to not be faced with dilemmatic situations except in cases of extraordinarily bad luck – this is, to be honest, probably why anyone could have thought that dilemmas were impossible to begin with. The lessons of oppression theory remind us that when social institutions are constructed and maintained by and for people in positions of social privilege these institutions will insulate those within them from being faced with moral dilemmas, but those outside their bounds, at the margins of society, enjoy no such protection.

It is unsurprising to find a similarly blinkered focus in the adjacent canonical literature on moral luck, where the questions introduced by Williams (Reference Williams1981) and Thomas Nagel (Reference Nagel1979) about the impact of luck on ethical life were structured by an underlying assumption that such questions arise primarily in the context of extraordinary events such as accidents, risky choices and dramatic reversals of fortune. What was considered to be philosophically interesting was how unpredictable incidents can affect moral assessments of goodness, blameworthiness and punishment; it does not seem to have occurred to most engaged in this literature that the odds of facing such incidents are in fact entirely predictable. Claudia Card (Reference Card1983) coined the influential phrase ‘the unnatural lottery’ in response to this literature, calling attention to the fact that while everyone can be vulnerable to bad moral luck, we should not pretend that all people are equally vulnerable. Card saw that an agent’s position in a racist heteropatriarchy would give rise to vast differences in moral luck, and that these differences would result in vastly different experiences of moral life.

The influential work of María Lugones can also be read as a contribution to this philosophical discussion. Referring to herself as a ‘multiplicitous’ human being with a ‘fragmentary’ or ‘curdled’ identity, Lugones rejected both the tenability and desirability of unifying her identity into a subjectivity without conflicting elements. This kind of unification might be possible for a ‘Christian, white, bourgeois man’, but for someone like her – an activist against the racist oppression of Latinx people by the dominant American culture who simultaneously participates in the privileges of that culture as an academic; a lesbian who faces different manifestations of homophobia in both cultures; a woman who faces different manifestations of sexism in both cultures – any attempt at unification would be an act of ‘self-mutilation’, a misrepresentation of who she actually is (Lugones Reference Lugones2003: 131, 41). Because she cannot simultaneously endorse the conflicting values of the different oppressed cultures she is a member of, she argues that there is no unified perspective from which to navigate her moral life. Resisting oppression, for Lugones, thus takes the form of refusing to play the game of what she calls ‘rationality’, and instead embracing ambiguity and plurality. (This ambiguity and plurality is represented in part by allowing her narrative voice to shift back and forth between English and Spanish.) Her point here is that certain cultural aspects of one’s identity can be not just incompatible but also incommensurable with aspects of one’s identity that stem from a different culture. Lugones argued that this incommensurability means that any attempt to resolve the conflicts one experiences as the result of being multiply oppressed risks being conceptually impossible, practically misguided and morally undesirable.

Recognizing the ubiquity of dilemmatic conditions under oppression, Lugones saw all too well that morality can abandon us precisely when we need it the most. Recall that metaethical theorists who want to deny the existence of genuine moral dilemmas are usually motivated by the fear that dilemmas threaten the legitimacy of morality as a whole. In a clever demonstration of the aphorism that one person’s modus ponens is another’s modus tollens, Lugones takes the recognition that morality can abandon us to suggest that we should be the ones abandoning morality (at least as it has been canonically understood, as a closed logical system).

But there are philosophical implications here beyond the metaethical. A central motivation for those concerned with understanding the moral residue that is attendant upon moral dilemmas, remember, was the ethical concern that certain emotions such as guilt or remorse are morally appropriate reactions to dilemmatic situations. That is, what was seen to be chiefly at issue in these debates was the question of how a particular moral agent – an agent understood in the most abstract and idealized terms – should be morally required to feel when faced with a moral dilemma. But there is another frame that is just as interesting and perhaps even more important to consider here, one that takes the central issue to be less one of judging individual agents for how they feel, or even what they do, in dilemmatic situations than it is to have us attend to the patterned and systemic nature of how and when such situations arise in the first place. The fact is that certain people are far more likely to find themselves trapped in a birdcage or torn apart on a rack than others, and it is not random fate upon whom such misfortunes befall. The insights of oppression theory suggest that the social and political significance of moral dilemmas is just as important as the metaethical and ethical.

The oppression theorist who has worked most explicitly, and fruitfully, with this insight is Lisa Tessman. Tessman’s central contention is that philosophers have tended to deny the possibility of moral dilemmas because they have mistakenly thought that the only point of morality is action-guidance. While generating practical advice about what to do is, of course, one of the central tasks of moral theory, she argues that another equally important but heretofore less appreciated consideration is a theory’s ability to make sense of our actual lives. This is particularly important when it comes to the lives of those who are oppressed, for it is precisely these vulnerable people who are more likely to find themselves in situations where morality’s ability to guide action has run out. ‘Because oppression is a significant source of dilemmatic moral conditions, and as a result moral failure tends to be ubiquitous under oppression’, Tessman argues, ‘at least some normative theorizing about oppression and resistance [should] direct attention to the failures that shape moral life’ (Tessman Reference Tessman2015: 176).

Letting go of the mistaken assumption that morality can always guide us out of dilemmatic situations without moral residue is an important element in the shift from ideal to non-ideal theory. It is only when a theorist starts with the false, idealizing assumption that the world is not unjust that they could reasonably conclude that morality could never demand unfair or impossible things of an agent. The chief virtue of non-ideal theory is that it has us look at the world as it actually is, not as it should be under idealized conditions that abstract away from the concrete material conditions people actually find themselves navigating. In his ground-breaking skewering of ideal theory, Charles Mills (Reference Mills, DesAutels and Urban Walker2004) shows how philosophy’s historic idealizations have functioned to exclude (or make appear ‘defective’) members of oppressed groups, and how its counterfactual characterizations too often wipe the theory clean of real-world problems. Non-ideal theorists in Mills’ wake recognize that ideal theory is not equipped, nor even intended, to address real-world injustices, much less those that result from the systematic forces of oppression. In this respect, non-ideal theory represents a clear improvement over ideal theory’s tendency to be overly optimistic about morality’s prospects for making sense of the lives of those who are oppressed. An important way to de-idealize moral agents is to understand them as people who will very likely face complex moral conflicts and emerge from such conflicts bearing identity-constituting moral remainders.

But Tessman cautions that non-ideal theorists can also be guilty of a too-narrow focus on action-guidance and that they, too, can be ‘falsely cheerful about the possibility of moral salvation’ (Tessman Reference Tessman2015: 175). ‘Not all wrongs can be rectified, not all losses can be compensated, not everything can be repaired or replaced, and – especially given the limits of psychological resilience – not everyone can recover’ (Tessman Reference Tessman2015: 178). Mills’ insight, that aiming at an ideal society with no acknowledged history of injustice can conceal the existing wrongs that need to be righted, can be extended here: if our aim is only ever to rectify the wrongs of oppression, we can forget that it is virtually inevitable that we will fail at this task, and we can obscure the wreckage of, or make untheorizable, past failures. Properly appreciating the harms and injustices of oppression requires recognizing that to insist that agents could never find themselves in a situation without moral residue is to function as an apologist for precisely those unjust systems one is committed to dismantling.Footnote ²

6. A radical Kantian analysis of moral dilemmas

I have argued here that, despite the many shortcomings of the mainstream philosophical canon, it can remain a productive strategy for oppression theorists to use the resources of this canon for more radical purposes than those intended by its creators themselves. A radical project to rationally reconstruct the Kantian corpus – one that both rejects an overly strong reading of ought implies can and that takes on board the radical political insights of oppression theory – must be properly cognizant of Kant’s failings. But such a project, one that retains elements such as Kant’s commitment to the inalienable dignity of humanity, his proscriptions against servility, and his recognition of the importance of self-respect, remains fruitful for anti-oppression purposes. What, concretely, might this approach look like when applied to concrete dilemmas?

Let us briefly consider how a radical Kantian might approach one of Tessman’s representative examples of how dilemmatic conditions can arise under oppression (an example first discussed by Patricia Hill Collins):

The work of oppression theorists allows us to see how the intersections of racial, gender and classed-based oppressions systematically and institutionally constrain the option sets for women like these in ways that are anything but random. Instead, these Black mothers find themselves facing moral dilemmas that are patterned and predictable. Kant’s (and too many Kantians’) callous indifference to this political reality is indefensible. But this is an oversight that can be rectified.

Properly reconstructed to accommodate the material realities of oppression, Kant’s moral framework can provide a robust means of articulating the distinctive moral harms and injustices brought about by dilemmatic material conditions such as the ones these women face. The fraught sexual politics Black women must navigate forces mothers to teach their daughters to protect themselves from harassment and abuse by keeping their heads down and knowing their ‘place’, but doing so ensures their continued exploitation. A Kantian will condemn this exploitation, for example, as incompatible with the respect these mothers and daughters are owed as persons. A Kantian will also point out that these mothers’ protections come at the expense of their daughters’ self-respect, and that they encourage their daughters to violate Kant’s proscriptions against servility.

Because neither the duty to ensure their daughters’ physical survival, nor the duty to foster their daughters’ emotional resilience and self-respect, nor the duty to avoid servility, cancel each other out, these mothers are placed in a position where it is impossible to do everything the moral law requires of them. They cannot protect their daughters physically without endangering them emotionally and morally; they cannot protect them emotionally and morally without endangering them physically. The moral residue in such situations – being complicit in the ‘emotional destruction’ of one’s own child – is extreme. Kant’s Formula of Humanity captures the unconditional and incomparable moral worth these daughters possess. His account of the importance of self-respect captures the distinctive moral harm of these daughters’ emotional destruction. His proscriptions against servility capture the distinctive moral failure these daughters face. And his requirement that people are to be respected in virtue of their dignity explains why these mothers’ failure to protect their daughters is so tragic.