Is a ‘dirty hands’ (DH) scenario conceptually possible or, as critics argue, an incoherent notion that merely adds further confusion to those already perplexing cases of apparent moral dilemmas?Footnote 1 If DH scenarios are indeed conceptually impossible, then all extant discussions of their sui generis properties and their place in our moral vocabulary are at best redundant. Asking whether we ought to punish those with DH would be as pointless as searching for the properties of square circles or seeking to list the habits of married bachelors. Moreover, mistakenly advocating for the possibility of DH would not simply be an inconsequential conceptual error. To wrongly maintain that it is possible to engage in a form of justified wrongdoing, to advocate the possibility of doing wrong in order to do right, can and will be used as moral cover to excuse a plethora of immoral and evil actions while claiming that they were necessary, unavoidable, and honorable. So, for those who argue for the existence of DH scenarios it is necessary to successfully respond to charges of incoherence.

1 The Project and Strategy

I have argued elsewhere that DH scenarios are an inescapable and essential part of our moral reality (de Wijze 1994, 1996, 2003, 2005, 2012, 2013, 2018). By ‘moral reality’, I mean the complex and entire experience of our moral lives—the choices made, the moral emotions felt, the sense of one’s own goodness and integrity, the special normative demands of our social roles, the pull of different moral imperatives from personal duties to universal consequentialist and deontological obligations. This moral reality is reflected in our moral phenomenology—our moral intuitions and emotions—and this phenomenology must play a role in properly understanding it. I reject the charge that talk of DH is conceptually incoherent. This paper seeks to outline two arguments against the conceptual possibility of DH and offer three responses. My aim here is to consolidate and extend a number of arguments scattered throughout the recent literature on DH, which have not been systematically brought together as a comprehensive defence of this unique feature of our moral reality.

This is not a straightforward task since the charges of incoherence by critics rely on wider views about ethics in general. So, the argumentative battleground, so to speak, is less about the details of DH scenarios themselves and more about what any moral theory can allow and must achieve. In short, those who dismiss DH as incoherent do so because they reject any ethical view that claims it is possible to experience unavoidable and morally justified wrongdoing. An efficient and coherent moral theory precludes this possibility. The widespread phenomenology that contradicts this is based, so they argue, on a false and primitive understanding of morality, which fails to recognize the deeper conceptual and logical constraints on any efficacious ethical theory.

The strategy of this paper is to set out the two main arguments explicitly used by critics of DH (or who may simply assume them in the background) and offer three responses. I will not be able to give definitive knock-down arguments against my opponents, but I do seek to at least to raise reasonable doubts. I hope to persuade critics that there is a need for modesty in claiming that unavoidable justified moral wrongdoing is incoherent and ipso facto so is their rejection of DH on these grounds. These debates have long occupied moral theorists especially when arguing about the possibility of moral dilemmas, whether there is a pluralism of values, how to best to conceptualize ‘values’, ‘reasons’, ‘agents’, and much more besides. I will be helping myself to what I take to be the best arguments, both for and against dilemmas, in the literature. My own contribution is to argue that if we acknowledge that the disagreement about moral dilemmas and all-things-considered moral judgments are inconclusive and essentially contested at the theoretical level, we should be open to a tie-breaker argument, namely that rich literary and lived experiences strongly support the view that we ought to accept the possibility of justified wrongdoing (or unavoidable moral wrongdoing or moral failure) which results in a moral remainder of some sort. DH scenarios are just these kinds of situations where there are specific conditions for a particular kind of moral staining, that is a dirty experience, even though the action which has been performed is morally justified. In short, our intuitions and moral emotions which arise in such situations provide a strong reason to reject the argument that DH scenarios are incoherent, despite there being no agreement among moral theorists concerning two core conceptual issues: the existence of real or genuine moral conflicts and the preeminence of all-things-considered judgments.

2 Paper Structure and Argument

I begin the paper (Sect. 1) by disambiguating two common ways of understanding the origins of DH scenarios. There is confusion among supporters and critics of DH scenarios who, since the publication of Michael Walzer’s seminal article “Political Action: The Problem of Dirty Hands’ (Walzer 1973: 160–180), disagree about what exactly constitutes cases of DH and how they arise as part of our moral experiences. This results from a reductionist error that leads to an incorrect understanding of the concept itself and causes unnecessary confusion when arguing against the critics of DH. So, I want to ensure in this defence of DH scenarios that there is a better understanding of the two different ways of setting out this concept. I call them the ‘Walzerian’ and ‘Machiavellian’ interpretations, which while different nevertheless share in common what is essential to all DH scenarios. Each view, with its own philosophical background and specific focus, educes different reasons for why DH situations arise, why they are morally significant, and why they are unavoidable, but share the fundamental assumption that such a concept is neither a contradiction nor incoherent.

In Sect. 2, I set out the three jointly necessary and sufficient conditions that are entailed by the concept of DH. Here I stipulate that, firstly, such scenarios are cases of genuine moral dilemmas and that, secondly, in such circumstances some decisions on how to act, given the ‘all-things-considered’ judgment, will result in the violation of some moral value or principle. The third stipulation is that the dirt in DH scenarios arises from the obligation to carry out an ‘impossible ought’ which results in the morally justified action which seriously mistreats persons. Victims of DH actions have a legitimate complaint that they have been treated immorally, their trust, integrity, and status as ends in themselves violated, dishonoured, and betrayed.

Section 3 examines sets out the two most common and influential arguments against the coherence of DH. I refer to them as the ‘Arguments from Deontic Logic’ (ADL) and the ‘All-Things-Considered Argument’ (ATC), each of which has much initial plausibility and are the primary reasons for why DH faces widespread opposition from contemporary moral theorists. These arguments are not directed specifically at DH but rather argue that to claim there are genuine moral dilemmas is incoherent. However, since DH scenarios necessarily presuppose the existence of genuine moral dilemmas, the ADL and ATC arguments consequently indirectly reject the conceptual possibility of DH scenarios.

Section 4 offers three responses to the charges of incoherence. The first responds to ADL by reiterating some extant arguments in the moral dilemma debate. My aim here to raise reasonable doubt among those who think that moral dilemmas are impossible by questioning the assumed correctness of deeper assumptions underlying these arguments. The strategy is to persuade my opponents that there is a reasonable case to bracket the claim that there are no genuine moral dilemmas, and then attend to other arguments in favour of DH. It may be the case that DH ought to be accepted as a coherent moral phenomenon given a wider understanding of how DH scenarios might explain our moral phenomenology combined with theoretical arguments supporting their coherence. What I am seeking is a broader coherence between certain theoretical arguments for DH scenarios and our practical moral experiences on the ground incorporating other notions such as moral emotions, punishment theory and so on.

The second response focuses on showing why ATC is also vulnerable given its starting assumptions about the nature of values, persons, and actions. Again, the focus here is to show why any confidence in the ATC argument needs to be tentative since, as I will argue, moral theory is not merely, or even primarily, focused on action guiding information, and this is especially so in situations of genuine moral dilemmas. The third, and final, response to those who claim that talk of DH is incoherent offers the evidence of our moral emotions, specifically ‘tragic remorse’Footnote 2 as evidence that such scenarios exist. The dismissal of moral emotions as a legitimate argument for the existence of DH is based on a particular and much contested view of how we best obtain knowledge of our moral reality. DH provoke a special kind of moral emotion that fits with our strong intuitions about our moral reality, one that acknowledges their authenticity and strongly questions the viability of dismissing such moral phenomena. While DH scenarios are widely represented, discussed, and taken seriously in literature, popular culture, political tracts, historical writings, films and TV dramas, they have been given short shrift by most contemporary moral theorists. This suggests that, at least prima facie, contemporary moral philosophers may be failing to notice something of great importance to our moral lives of which other disciplines are cognizant. If my responses to critics of DH are credible, then we need to continue refining our understanding of the DH phenomenon and explore its wider implications for a range of political, psychological, and moral issues.Without a clear account of DH scenarios and their moral implications, we lose our ability to describe this unique aspect of our moral experience and, instead, respond in inappropriate and damaging ways to these difficult and unavoidable situations of moral conflict.

3 DH—Two Models

The recent plethora of papers concerning the notion of DH can be traced back to Michael Walzer’s seminal article ‘Political Action: The Problem of Dirty Hands’ (Walzer 1973). His paper posed a direct challenge to the orthodoxy in contemporary moral theory on how to respond to moral dilemmas in public life. Walzer’s central claim, that we can do wrong to do right, highlights the paradoxical natureFootnote 3 of DH and argues that it is an important part of our moral reality, one that strongly and properly resonates with our moral intuitions. While Walzer’s discussion of DH uses contemporary hypothetical political problems in a democratic state under the rule of law, this moral problem can be traced back to antiquity.Footnote 4 While Walzer’s account of DH has become the focus of much recent commentary (it is seen as the paradigm formulation of this problemFootnote 5), it represents a particular strand of theorizing on this concept that fits with recent concerns within contemporary moral theory. However, another important way of thinking about the problem focuses on the tension between delivering public goods or benefits while maintaining individual moral virtue.Footnote 6 This latter formulation of the DH problem arises primarily as a political problem, one that suggests that political action needs to be assessed by a different ethic or moral code and for political agents to adhere to a different set of moral virtues.Footnote 7

So, we find in the contemporary literature different explanations for DH scenarios. The first version, the Walzerian Model, asserts that DH scenarios arise when there is an unavoidable clash between incompossible moral obligations arising from competing approaches to moral theorizing. While deontological reasoning focuses on an agent’s adherence to fundamental moral principles, which are based on reason or revelation, consequentialists look to the effects of possible actions in maximizing a particular non-moral good such as utility or happiness.Footnote 8 Deontologists evaluate the motivations and intentions of agents, while consequentialists seek to justify actions by determining what is necessary to achieve a desired end-state. Since all complex moral judgments typically combine both deontological and consequentialist considerations, there is potential for a serious conflict to arise between them.Footnote 9 Walzer’s argument that a politician may face a situation where ‘exactly the right thing to do in utilitarian terms’ leaves ‘the man who does it guilty of a moral wrong’ (Walzer 1973: 161) sums up this inherent instability when consequentialist and deontological considerations’ part ways. Nevertheless, those who reject the possibility of DH scenarios frequently do so from either an absolutist form of their preferred approach, or argue for a threshold deontology of some kind, one that establishes the right action without a moral remainder.Footnote 10 Given this, Walzer’s claim that it is possible and sometimes necessary to do wrong to do right, is deemed incoherent and based on a serious error in moral reasoning.

The second version of DH, the Machiavellian Model, focuses on the problem of achieving laudable and noble ends when this requires using immoral means and, in the process, the loss of moral virtue.Footnote 11 It has been long recognized that achieving a desired end-state often requires engaging in practices and actions that are morally dubious or even reprehensible. Nowhere is this more so than in the domain of politics when seeking to bring about justice, peace, and other social benefits for the society. Human vice abounds and the actions and projects of evil persons give rise to unavoidable situations that compel the use of morally problematic actions to counter them. Hence, the problem of DH has been characterized as the ‘ends/means’ problem in politics. This view differs from the Walzerian approach in that it is not primarily about means and ends if this is interpreted as merely a clash between deontological and consequential moral reasoning. As Niebuhr and Hampshire point out,Footnote 12 this political dilemma is unavoidable since human social relations in general, but particularly in large societies, are more or less in a permanent state of conflict. Consequently, effective and successful politics that strives for social justice, peace and security, will sometimes leave a politician facing a tragic clash between her personal and political moralities.Footnote 13 Those who take on such political obligations, to use a religious metaphor, endanger the salvation of their souls in order to benefit the people they serve (Weber 1958: 126). ‘To be a good man according to God is one thing; to be a good man according to men is quite another (Huxley 1994: 148).’Footnote 14 Successful politicians face the burden of being ‘suffering servants’ or ‘tragic heroes’ (Walzer 1973: 176–177).

The focus of the Machiavellian Model, then, is concerned specifically with the clash of public duties with private moral convictions. This model is committed to some form of role morality and value pluralism, at least in the distinction between, and the recognition of, the valid demands of both a public and private morality. The pluralist view of value endorses the claim that people hold and cherish many different values, which are neither reducible to each other, nor placeable in a hierarchy, nor made subservient to a single supreme or overriding value.Footnote 15 Consequently, there will always be situations when the clash between public and private values ensures that the realization of one cherished value undermines the other. The deep philosophical error made by those who reject the Machiavellian model results from their commitment to a reductionist abstract ethical position, which purports to provide clear and unambiguous action-guiding advice even when public roles clash with private values. However, this claim of ethical neatness based on a singular concern with action guidance starkly contradicts the messiness of our moral reality reflected in the widely reported phenomenology of our moral and political experiences. A fully engaged life embraces many different and sometimes incompatible values and projects. For example, a society that places the value of political liberty above all else undermines the importance of striving for political equality, and vice versa. If this is correct, then as Berlin points out, there is no social life without loss. ‘We are doomed to choose, and every choice may entail an irreparable cost (Berlin 2013: 14).’ The Machiavellian model, then, is defended by pluralists who point to the ethical and practical implications of facing incompatible or incompossible political and personal moral values and duties. They argue that failing to accept this reality is both theoretically naïve and has practical implications which are dangerous.

However, whichever model of DH is adopted, critics of this concept raise the charge of conceptual incoherence. So, for the purposes of this paper there is no need to differentiate between different models of DH. Since both trigger the charge of incoherence, in the following section I stipulate the minimal necessary and sufficient conditions to be met for any action or project to be plausibly described as a genuine case of DH. Or to put it another way I set out the de re conditions for the concept of DH which can be widely supported by the many different conceptions in the literature to date.Footnote 16

4 DH—3 Jointly Necessary and Sufficient Conditions

The claims I defend in this paper are that DH scenarios are a sui generis form of justified moral wrongdoing, and they are a coherent and necessary part of our moral vocabulary. The concept of DH is a member of a subset of concepts within the class of genuine moral conflicts. In the DH scenario, all possible choices necessarily entail a serious moral wrongdoing where agents become morally stained for so acting. For an action to be properly labelled as DH it needs to meet three jointly necessary and sufficient conditions. I simply stipulate these conditions here since the charge of conceptual incoherence is not based on a dispute about what DH theorists and others deploying this family of similar concepts claim about themselves per se. Critics do not dispute that if there were a tenable coherent concept of DH, then these three conditions would be a reasonable account of the concept itself. Their concern is with the very possibility of these criteria applying to moral actions. So, DH scenarios, as a unique type of justified wrongdoing, must satisfy the following conditions:

  1. i.

    All cases involve genuine moral dilemmas. Genuine dilemmas involve conflicting pro tanto reasons to act (or not act).Footnote 17

  2. ii.

    All action-guiding decisions result in an unavoidable moral violation. (This condition endorses the ‘Remainders Thesis’ (RT) and rejects the Eliminativist Thesis (ET)).Footnote 18

  3. iii.

    DH include ‘impossible oughts’Footnote 19 such that the action involves committing a determinate moral crime involving the violation of persons, principles, or values.Footnote 20

Conditions (i) and (ii) situate DH as members of the set of moral scenarios involving genuine moral dilemmas which are both unavoidable and leave a moral remainder of some kind. This remainder may be a sense of moral failure, or the clear emotional acknowledgement that moral wrong has occurred, or that whatever choice is made ensures that the agent becomes a bad person, or the recognition of evil done to others with the acknowledgement of the need for some form of reparations.Footnote 21 However, DH scenarios also need to meet condition (iii) to set them apart from these other forms of unavoidable moral wrongdoing. The ‘dirt’ in DH arises from the sui generis and pernicious nature of particular moral dilemmas, where the need to violate a stringent moral duty arises due to a ‘complex of immorality’ (Stocker 1990: 25).Footnote 22 There is considerable difference among theorists concerning whether a certain kind of betrayal, or the circumstances under which this betrayal becomes needed, best defines the concept of DH.Footnote 23 However, this debate is not relevant to responding to the charge of conceptual incoherence that is levelled at DH scenarios.Footnote 24

So, for the purposes of this paper, DH scenarios are dilemmatic situations, where no matter how the agent (or, collectively, a committee, government, or electorate) acts there is no avoiding the serious violation of persons, values or principles when seeking to bring about the lesser evil. Agents who commit DH actions compromise their moral goodness and unavoidably become morally polluted. This unique type of inescapable moral wrongdoing usually involves very serious cases of moral violation. Such cases are especially pernicious as they involve good persons committing determinate moral crimes when doing, all things considered, the morally justified and laudable thing to do in these circumstances.

5 Rejecting DH

Defenders of DH scenarios face many criticisms. In this paper I respond to only one, the charge of conceptual incoherence. If I succeed in demonstrating that DH scenarios are possible this would still leave those criticisms which argue that it is unwise to endorse this notion as a legitimate part of our ethical toolbox. The foci of these other criticisms fall under two broad approaches. The first rejects DH as it would face considerable practical difficulties in both its implementation and integration within a cluster of other important normative concepts, such as guilt, shame, regret, pride, legitimacy, and accountability (to mention a few). For example, justifying punishment for wrongdoing becomes problematic when applied to scenarios of justified wrongdoing that involve getting DH.Footnote 25 Secondly, there is much concern that invoking the concept of DH begins a slippery slope toward general moral decline by creating a theoretical space which can easily be exploited to vindicate impermissible actions or projects. DH rationalizations far too easily substitute for the rare genuine cases of DH dilemmas. While these criticisms require a response, space limitations prevent me from addressing them here. At any rate, if the arguments for why DH scenarios are conceptually possible are unpersuasive, then the need to respond to the above criticisms become redundant. So, I turn to the incoherence arguments against DH which are the concern of this paper.

5.1 Two Incoherence Arguments (IA)

There are two versions of the IA which seek to show that the concept of DH is incoherent. The first appeals to deontic logic to demonstrate that genuine moral dilemmas are impossible. Given that all cases of DH are special cases of moral dilemmas, ipso facto such scenarios are conceptually impossible. The second approach focuses on the action guiding priority of moral theories and what it means to say that an all-things-considered judgment has been established. The concern here is that in cases of unavoidable or justified moral wrongdoing, the all-things-considered reasons for a particular action does not negate the pro tanto reasons to also refrain from such an action. To assert this is to claim that an action can be simultaneously both right and wrong. Or put another way, when deciding whether to do A or B (actions which preclude each other) there are overriding reasons to do both A and B.Footnote 26 This, critics argue, is simply incoherent. I set out the details of each of these approaches below.

5.2 The Impossibility of Genuine Moral Dilemmas—Arguments from Deontic Logic

The arguments from deontic logic seek to show that if we accept certain axiomatic claims about morality which logically preclude the possibility of moral dilemmas, then any further claims (such as the existence of DH) that contradict these axioms are incoherent. There are four such axioms which underlie the two arguments from deontic logic, and they cannot be held consistently with the claim that genuine moral dilemmas are possible. These are:

  1. i.

    Principle of Deontic Logic (PDL)—if doing A brings about B, and A is obligatory, then B is also obligatory.Footnote 27

  2. ii.

    Principle of Deontic Consistency (PDC)—if A is obligatory, it cannot also be forbidden.Footnote 28

  3. iii.

    ‘Ought implies Can’ Principle (OiC)—all moral obligations presuppose actions which are also physically possible.Footnote 29

  4. iv.

    Agglomeration Principle (AP)—If an agent ought to do each of two actions, A and B, then she is required to do both.Footnote 30

These axioms are very widely taken to be self-evident, true, and fundamental to the logical foundations of our thinking about the possibility and necessity of moral action. Given this, when claiming that we face a genuine moral dilemma, we incoherently assert by implication that we are morally required to do both A and B but that it is also impossible to do so.Footnote 31 This claim cannot be consistently held in conjunction with either PDL and PDC or with OiC and AP. Consider the following argument to demonstrate that asserting the axioms OiC and AP prevents us also claiming that moral dilemmas are possible.Footnote 32 Asserting the five premises below results in a contradiction.

  1. 1.

    OA

  2. 2.

    OB

  3. 3.

    ¬C(A&B) [where ‘¬C’ means ‘cannot’] (Conjunction of 1-3 represents a moral dilemma)

  4. 4.

    OA→CA (for all A) [OiC]

  5. 5.

    (OA&OB)→O(A&B) [AP]

    --------

  6. 6.

    O(A&B)→C(A&B) (an instance of 4)

  7. 7.

    OA&OB (from 1 and 2 - Conjunction)

  8. 8.

    O(A&B) (from 5 and 7 - Modus Ponens)

  9. 9.

    ¬O(A&B) (from 3 and 6 - Modus Tollens)

Lines 8 and 9 contradict each other, asserting that one ought to do A and B and one also ought not to do A and B.Footnote 33 Consequently, either the claim that real moral dilemmas exist is true or the axioms of deontic logic are true, but they preclude each other. And, given that the axioms of deontic logic are self-evident in a way that the existence of moral dilemmas are not, the former claims about the existence of real moral dilemmas need to be dismissed as incoherent assertions about our moral possibilities, and ultimately our moral reality. In the literature, as McConnell (2018) notes, much of the debate on whether moral dilemmas exist has concentrated on whether it is possible to avoid the inconsistency shown above.Footnote 34 For many, the above argument definitively settles the issue concerning the existence of real moral dilemmas. When we think we face such situations, these moral dilemmas are merely prima facie and disappear when moral theories are properly applied to the ethical conundrum.

I return to this issue in the next section but here I simply want to endorse Weber’s plea that there be modesty ‘concerning what theoretical reason can accomplish in the moral dilemma debate’ (Weber 2002: 470).Footnote 35 While most ethical theorists agree that the axiomatic claims underlying the deontic arguments appear to be self-evident, I will argue that they are problematic and contestable. In Sect. 4, I call attention to the fact that the deontic arguments are not definitive since they are based on a set of unchallenged assumptions about the role of moral theory that problematically endorse what Stocker refers to as ‘strange theories of value’ (Stocker 1990: 26).Footnote 36 For example, if it is thought that the primary purpose of moral theory is to be action-guiding, then there is a strong motive to use axiomatic assumptions which already preclude the possibility of genuine moral conflicts. I return to Stocker’s arguments later but here my concern is to raise reasonable doubt that real moral dilemmas are impossible and argue that, at the very least, this judgment needs to be bracketed and later re-considered in the light of a number of positive arguments for the existence of DH scenarios.

5.3 DH and ‘All Things Considered’ Judgments

The second argument for the incoherence of DH scenarios focuses on how disparate moral reasons are managed, compared, and then acted upon when we face what appears to be cases of intractable conflict. Such conflict, for example, arises when moral obligations or duties, generated by the generic demands of morality, clash with the responsibilities and duties arising from different social roles in public life. In these situations, an agent may face an obligation to do X and simultaneously have a duty not do X. Or an agent may be required to do X yet also be required to do Y even though actions X and Y are morally incompossible. However, this is an intolerable situation for moral theorists who maintain that action-guidingness is a core and primary virtue of moral theories. They insist that the raison d’être of a properly functioning and efficacious moral theory is to provide, in all circumstances, an ‘all-things-considered’ judgment that offers clear action-guiding information.

If this is indeed the case, then it is not difficult to see why we need to reject DH scenarios. Given the jointly necessary and sufficient conditions for DH scenarios set out in Sect. 2, they cannot be part of any coherent and effective moral theory. To claim that an action is the morally right thing to do, yet that it is also somehow wrong is intolerable. Indeed, any moral theory that endorses the Remainders Thesis (RT) (See note 18) is simply not credible given that it deeply misunderstands its core purpose. Consequently, apart from the concern that there can be no genuine moral dilemmas, there is also the deep problem that DH scenarios entail that there can be an all-things-considered moral reason for action, yet also maintain that such an action is somehow morally wrong. As Nielsen insists, to endorse such a claim is to heap confusion on already difficult and confusing situations.

However this (do what is right by doing what is wrong)... is a mistaken way to conceptualise things. Where whatever we do, or fail to do, leads to the occurrence of evil or sustains it, we do not do wrong, everything considered by doing the lesser evil. Indeed we do what, everything considered, is the right thing to do; the thing we ought, through and through ought, in the circumstances to do. In doing what we ought to do we cannot (pace Walzer et al.) do wrong. We do things that in normal circumstances would be horribly wrong, but in the circumstances of dirty hands, they are not everything considered wrong. It is difficult enough in such situations to ascertain what the lesser evil is and to steel ourselves, when we are the agents who must so act, to do it, without, by adding insult to injury, making artificially and confusedly, a conceptual and moral dilemma out of it as well. (My emphases) (Nielsen 1996: 2).

Nielsen’s concern is widespread among moral philosophers who endorse an argument which takes the form of the one below:

  1. 1.

    There are two kinds of reasons for actions. Those reasons which contribute for and against how we ought to act, and those reasons which underlie the action-guiding decision itself. Let us call the former ‘contributory’ reasons and the latter ‘all-things-considered’ reasons.

  2. 2.

    Contributory reasons support (or reject) an action, but can be outweighed (or, more controversially, silenced) by other contributory reasons.

  3. 3.

    The all-things-considered reasons mark the act as right (or wrong). To put it another way, to have an all-things-considered reason supporting (or rejecting) an action is precisely how we decide if the action is morally right (or wrong). The all-things-considered reason is one which has examined and evaluated all the contributory reasons, and its core feature is that it cannot be outweighed. Claiming that there is an all-things-considered’ reason to act is to say that this act is morally ‘right’ or ‘permissible’ in these particular circumstances.Footnote 37

  4. 4.

    Moral dilemmas arise when moral principles deliver different judgements on the same action, or we are unable to decide between two incompossible moral obligations. If the conflicting moral principles are taken to be absolute and inviolable under all circumstances, then this would prevent the possibility of establishing an ‘all-things-considered’ judgment on how to act.

  5. 5.

    However, the claim that moral principles are absolute under all circumstances is not a reasonable construal of moral principles. Both consequentialist and deontological approaches take cognizance of possible clashes but resolve them by either a) rejecting one of the conflicting principles as having only prima facie force, or by b) claiming that the clashing principles are pro tanto so that each stand as a contributory reason, which may or may not be outweighed by other reasons that give rise to an all-things-considered judgment.

  6. 6.

    DH scenarios, in contrast, claim that some actions are simultaneously right and wrong so that an agent is engaged in morally justified wrongdoing which leaves her guilty of a moral crime.

    Consequently,

  7. 7.

    Given 1–5 above, DH scenarios amount to saying that there can be an all-things-considered reason for performing X, and an all-things-considered reason to not do X. Or that there is an all-things-considered reason to do X and an all-things-considered reason to do Y even when X precludes Y.

  8. 8.

    However, conceptually, the notion of ‘all-things-considered’ reasons must be a singularity and thus there can only be one set of reasons to justifiably act in a certain way in each situation.

  9. 9.

    Since a foundational claim of DH scenarios is that there are cases where agents do wrong to do right, all such talk is incoherent since such scenarios are conceptually impossible.Footnote 38

For critics of DH scenarios, the rejection of genuine moral conflicts combined with the incoherence of claiming that there can be differing all-things-considered judgments when moral duties conflict, unequivocally settles the issue. DH scenarios are simply not possible.Footnote 39 Moreover, both arguments against the possibility of DH cohere with, and further support, the strongly held view that any efficacious moral theory must be able to give clear and unambiguous action guiding advice in all circumstances. Hence, the rejection of DH prevents confusion and moral error that would result if an agent believed that in the face of an apparent moral conflict, both options were obligatory or, alternatively, prohibited.

6 Responding to the Incoherence Arguments—3 Pathways

I now turn to responses to the arguments above and argue in support of DH as a coherent concept. Each of the three pathways below if taken alone is unlikely to shift the views of critics. But the hope is that their cumulative weight, first by raising significant concerns about the strength of the incoherence arguments and then offering a positive argument for DH, might bring about a shift in the views of even the most stringent opponents. Or, perhaps more realistically, those who reject the possibility of DH will be less sure that their dismissal is reasonable and will be more receptive to the arguments by DH theorists concerning the application of DH, be it to theorising about warfare, targeted killings, torture, punishment, moral emotions, and more besides. It may turn out that the plausibility and efficacy of such analyses for exploring these difficult issues will strongly suggest that the concept of DH is not incoherent and perhaps a better way of characterising part of our moral reality.

6.1 Deontic Arguments and Circular Reasoning

The use of deontic arguments to dismiss the possibility of genuine moral dilemmas, and hence DH, appear to be decisive since it is assumed that they are valid arguments based on self-evident true premises. The OiC and AP axioms appear to formalise the basic foundations of all legitimate moral reasoning and they cannot be consistently asserted with the claim that moral dilemmas are possible.Footnote 40 Yet, many have questioned the truth of these premises. Consider Weber’s concern with the deontic argument using OiC as a basic premise:

What I am saying is that one of the premises of the argument – ‘ought implies can’ – is not plausible unless one already accepts the argument’s conclusion – that inescapable moral wrongdoing and moral dilemmas are impossible. Thus the argument, while valid, is question-begging in textbook fashion (Weber 2002: 466).

Weber is claiming that what appear to be conceptually true and self-evident premises, the OiC and AP, are in fact assertions about the nature and purpose of moral theory that has arisen from prior and, at the very least, disputable intuitions.Footnote 41 These intuitions include a commitment to an ethical system that does not permit an agent to face situations of inescapable moral wrongdoing. If prior intuitions, or commitments to other principles of beliefs, such as ideas of moral responsibility and blameworthiness, are sufficiently dominant, then moral theory that conforms with these commitments will not allow for the possibility of genuine moral conflicts. For example, those who hold certain religious views might reject the possibility of genuine moral dilemmas because of a prior belief that a good God would not permit them. To allow genuine, as opposed to prima facie, moral conflicts would unacceptably force good persons to choose between sinful acts.Footnote 42 So while OiC and AP are used as self-evident conceptual truths about our moral obligations, they are axioms which reflect a controversial and contested prior conviction about the purpose of moral theory. Their use in the rejection of the possibility of moral dilemmas appeals to fallacious circular reasoning.

Another deep concern is with what Holbo (2002: 264–272) refers to as the ‘poverty of deontic logic.’ It is not clear what role the axioms of deontic logic actually play in commonplace moral reasoning, especially for those facing dilemmatic situations. Our complex moral reality takes place under non-ideal conditions and given this it is not clear why AP is a legitimate constraint on my moral obligations or duties. Recall that the AP states that if we ought to do each of two actions, A and B, then we are required to do both. Yet it is entirely plausible that in actual situations that arise in an imperfect world, I ought to do both A and B, yet they preclude each other. For example, I may be confronted with a choice that entails a duty to protect from harms yet doing this requires the use of violence that would take lives. The usual response to this scenario is to either reject one of these obligations as overridden or as simply prima facie using consequentialist moral reasoning or a hierarchy of deontological obligations. This manoeuvre retrospectively protects the AP and maintains the exclusion of incompossible moral demands. Yet this move simply instantiates the AP as an unassailable foundational assumption about moral systems and ignores our more complex experiences of moral dilemmas on the ground.

Similarly, invoking the OiC principle in the face of a moral dilemma does not remove the fact that in some situations all possible actions result in a serious moral violation. Yet, this principle implausibly rejects the very possibility of such a scenario. Holbo (2002: 265–266) uses the analogy of rules in a chess game to show why this stipulation is not credible. The King is constrained by two inviolable rules: it may not move into check and must always move out of check. When the King is checkmated one of these rules is unavoidably violated and the game ends. It is not that the King cannot find itself in a situation of checkmate, since clearly this occurs at the end of every chess game if there is no draw, but rather that when this happens there is no possible move without the violation of a fundamental rule of chess. This analogy highlights the situation faced by agents in cases of genuine moral conflicts. When situations arise where two moral obligations are incompossible, we face a practical problem where it is physically impossible to act without bringing about a moral violation. What bearing the OiC axiom has on this reality is not clear, but ruling out the possibility of such scenarios violates the constraints of space and time which circumscribe human limitations.Footnote 43 This suggests that these deontic axioms may not be relevant for certain moral conflict situations and hence they do violence to the phenomenology of our moral experiences. The denial of moral dilemmas by appealing to the OiC principle fails to resonate with our strong sense of having committed a determinate moral crime, albeit for justifiable reasons.

The poverty of deontic arguments, then, lies in their denial of lived experience in favour of prior theoretical commitments concerning the purpose of moral theories for an abstract ideal world. The deontic axioms make it impossible in dilemmatic situations for a person to act badly. Yet a morally tragic dilemma is just those situations where agents are forced to act immorally. It is not impossible to be in a situation where our options are limited to only two morally bad actions. Whether being forced to act in such situations makes me a bad person is a complex and different issue, but that I can be forced by circumstances beyond my control into such situations is undeniable. This is the basis of the claim by Nussbaum (1986: 25) that good people can be morally ruined or damaged by having to act in morally objectionable ways in response to events that are out of their control.Footnote 44 When this happens, it seems appropriate for good persons to feel some kind of moral residue for having so acted.

Finally, even if we agree that OiC is an important principle underling moral discourse, there is considerable disagreement on whether this standing is due to the strong feeling that it is self-evidently correct. Despite it being a longstanding principleFootnote 45 there is recent empirical research in cognitive science arguing that OiC is not an intuitive principle of moral psychology. While our moral duties are strongly related to whether we are able to carry out the act, the truth of this claim cannot be based solely on strong intuitions that this is so. Recent studies by Buckwalter and Turri (2014) have found that when presented with scenarios which violated the OiC principle, subjects consistently ascribed moral obligations to those situations.

Furthermore, there are many theoretical concerns with holding OiC as an axiomatic principle for assessing moral discourse. Buckwalter (2020) offers seven areas where OiC is in deep tension with foundation areas of ethical theorising; namely, ‘apologies, excuses, promises, moral dilemmas, moral language, disability, and moral agency’. While this is not the place to review this body of work, it is worth noting that at the very least it ought to raise considerable doubt about using OiC as a fundamental axiom in arguing against moral dilemmas. Weber (2002) is right to point out that arguments both for and against moral dilemmas are largely inconclusive. My task here is to persuade those who dismiss DH, due to their rejection of genuine moral dilemmas, that they ought to have reasonable doubts about the soundness of deontic arguments on which this is based. Given this, it would be productive to explore arguments for the possibility of DH, which in turn may offer a plausible alternative argument for why moral dilemmas are indeed possible.

6.2 The Problematic Assumptions Underlying the ‘All-Things-Considered’ Argument

The all-things-considered argument, like the deontic arguments, at first seems to be a decisive knock-down argument. It reveals that those who support the concept of DH fatally misunderstand the notion of an ‘all-things-considered’ judgment; namely, that it is the way in which we establish what is the moral action, and one which rightly discards overridden reasons as no longer morally relevant. Furthermore, if there is a prior commitment to the view that efficacious moral theories must, as their first priority, focus on providing action-guiding information in all situations, then claiming that an action is both wrong and right is counterproductive as well as incoherent. However, on closer inspection, this view is far from persuasive if we reveal and critically examine the prior theoretical commitments upon which it is based. Theorists who defend the Remainder Thesis (RT),Footnote 46 maintain that even when deciding on the correct action to take in moral dilemmas situations, these conflicts nevertheless sometimes involve moral wrongdoing that leaves a moral residue. As pointed out in Sect. 2, this view stands in stark opposition to the Eliminativist Thesis (ET) that maintains it is not possible to do what is morally right and yet by so acting also commit a moral wrong. The ‘All-things-considered’ argument against the possibility of DH endorses ET and rejects RT as incoherent.

What then can be said in defence of RT, and why do some overridden values leave a moral residue while others do not? There are at least three different but not mutually exclusive approaches to defending RT. The first arises from an interrogation of key assumptions about the nature of values, agents, and choices which underlie the validity of the ‘all-things-considered’ argument. It also involves an examination of which moral obligations can be overridden or rejected by the all-things-considered judgment yet still exert a moral force and leave a moral stain. The second, an approach that is used by those who favour the Machiavellian Model of DH, posits that we face a pluralism of values, and this leads to an inevitable clash between the demands of certain institutional roles and the generic demands of morality. The third approach arises from a methodological disagreement about how to best think about ethical issues. Those who reject RT generally evoke rationalistic and abstract notions to argue their position typified by the two incoherence arguments in Sect. 3. By contrast, those who support RT rely on actual experiences and the strong moral emotions they generate. (Holbo 2002: 271 refers to this difference as the rationalist/experientialist divide). The scholarship on these three differing approaches is vast and adequate coverage of it is far beyond the scope of this essay. However, for my purposes it will suffice to outline the core claims of each to show that there are plausible responses to the incoherence arguments thereby making conceptual space for the possibility of DH scenarios.

6.2.1 Disputing Assumptions About Values, Choices and Costs: ‘Impossible Oughts’

Perhaps the best and most extensive response to the incoherence arguments has been offered by Michael Stocker. While he is not the only theorist to defend the RT, he does so while specifically defending the possibility of DH. His strategy is to critically examine the assumptions made about values, actions and choices by those who defend the incoherence arguments. To this end, Stocker argues that while DH scenarios are difficult and troubling moral phenomena, their conceptual possibility is not in doubt. In all situations where choices are made, and this includes cases involving moral choices, there are ‘double-counted unavoidable costs’. Depending on the situation, some justified costs (costs for foregoing one option for the other) are regrettable. For example, losing a friend for telling him the truth is regrettable, but the cost of foregoing a pasta dinner for fish and chips is not. What distinguishes DH scenarios from the usual type of unavoidable costs is that they face ‘double-counted impossible oughts’ , those oughts with ‘dirty features’ that we are normatively unable to obey as they are deeply repugnant to our moral integrity and sense of self. In such situations this ‘dirty feature’ is double counted so it is ‘taken into account once in determining the overall value of the act (the lesser evil and all-things-considered judgment) and again on its own (Stocker 1990: 13).’ ‘What is morally unavoidable is said to tell against act and agent (Stocker 1990, p. 12).’ It is important to note here that in DH scenarios the action guiding decision does not negate or eliminate the moral force of those defeated reasons to refrain from so acting. What is overridden is the action-guiding force of such reasons, not their moral wrongness. The action remains a disvalue even though it is justified and this ought to be noted and regretted. I have argued elsewhere that DH scenarios leave an agent morally polluted, and given the unique circumstances that bring this about, the appropriate moral emotion when dirtying your hands is ‘tragic remorse’ (de Wijze 2005).

Why then do those who offer the ‘all-things considered’ argument reject the points made above? Stocker argues that their position is based on ‘radically implausible views of value and action’ (Stocker 1990: 13).Footnote 47 Those who reject the conceptual space for DH scenarios do so because, firstly, they assume values are always independent of each other. That is, if an action is impermissible and incompatible with what ought to be done then it cannot be good or justified in any way. Similarly, whatever is deemed necessary and morally justified cannot be bad or immoral in any way. The result of this view is that genuine moral conflict is deemed impossible. Secondly, moral values are seen as having similar properties to beliefs in that those values which are overridden by competing values disappear qua relevant values.Footnote 48 They no longer exert any moral force. But this is to misunderstand the nature of moral values which, as Williams (1973) points out, have in some respects qualities more like desires than beliefs.Footnote 49 Overridden desires (unlike overridden beliefs) remain to exert some force on how the person feels and acts, and what they have become. So it is with moral values, and this resonates with our moral phenomenology and moral emotions experienced in DH scenarios. Thirdly, those who reject DH are concerned solely with finding an action-guiding option given that we can only face prima facie moral dilemmas. All such situations are resolvable through a meaningful comparison between values so that a clearly better or top ranked option can be established. There is no interest in examining any other moral features such as whether the action would impugn our integrity or violate an absolute side-constraint in these special circumstances.

The key point to note here is that theorists who hold the above views cannot simply assume that they hold the correct and undisputed a priori presuppositions underlying moral discourse. Stocker’s questions above have been echoed by others who have also sought to develop alternative ways of thinking about the clash of values where a ‘responsibility overridden in deliberation about what to do remains a responsibility’ (Gowans 1994: 18). Contrary to the ET, our moral reality does sometimes entail situations where our moral wrongdoing is inescapable. Gowans, for example, defends this position with a phenomenological argument using a method of ‘reflective intuitionism.’Footnote 50 He argues that the moral distress we feel in situations of inescapable moral wrongdoing provides information about our moral reality and the kind of social relationships we are obligated to maintain. Tessman similarly argues that in some situations we need to accept that there will be ‘moral failure’ since the task of moral theory is not solely about seeking all-things-considered action guiding information. Our intuitions and emotional responses in some cases of moral dilemmas reveal to us that fundamental moral values underlying our most precious social and personal relationships cannot be overridden without a moral remainder.

It is important to stress here that supporters of the RT do not claim that all cases of moral dilemmas result in a moral residue of some type. To claim this would, at the very least, be contrary to the phenomenology of our moral experiences and misunderstand why moral residues arise in certain situations. The mere fact that we have competing moral obligations does not preclude the fact that in many, if not most, situations the overridden value is not double counted to exert a moral remainder. For example, if I promise to take my child to the park but get called away to care for someone who is very ill, breaking this promise will rightly provoke a sense of regret, but there would not be a moral remainder, let alone the moral pollution from a DH situation. However, if as Chief of Police, I threaten a suspect with torture to save the life of a child he has abducted, the act itself, while justified, is a case of DH and leaves a serious moral remainder.Footnote 51 What then explains this difference given both involved a clash of incompossible values? The answer lies in understanding the distinction between negotiable and non-negotiable moral obligations. Tessman explains the distinction as follows:

Non-negotiable moral requirements – those that cannot be absorbed into an all-things-considered ‘ought’ through either substitution or compensation – remain requirements, contravening the principle that ‘ought implies can’ (Tessman 2015: 44).

A negotiable moral requirement, then, is one where the overridden value does not result in the loss of a unique value. Breaking my promise to my child can be either substituted for another value that fully counterbalances the loss, or compensated for at a later date. I can bring my child a present in lieu of the visit to the park or go for a walk another day with the compensation of an ice-cream. Non-negotiable moral requirements, in contrast, involve what Tessman calls ‘sacred values’; those values which we are unable to even consider compromising for another value, even if the alternative is another sacred value (Tessman 2015: 4–5). When a sacred value is violated, it almost always involves a special tragic kind of cost and leaves a moral residue, one that needs to be noted and appropriately felt by good persons.Footnote 52 The threat of torture by the Chief of Police violates a sacred value and even if it results in the saving of the child, it has violated a value that is unique and not compensable. It is such dilemmatic cases that the OiC principle becomes inoperable since the moral prohibition on torture remains even though the all-things-considered judgment is to so act. Tessman sees this as unavoidable ‘moral failure’, Gowans as a case of inescapable moral wrongdoing, Nussbaum as a tragic conflict and Stocker as a DH scenario. What is common to all their accounts is the endorsement of the RT where there will be in such cases a moral remainder to be noted and regretted (see table below).

 

Moral failure (Tessman)

Inescapable moral wrongdoing (Gowans)

Tragic conflicts (Nussbaum)

Dirty hands (Stocker)

Different accounts for non—negotiable values which leave a moral residue

Violation of sacred values—values we cannot think about compromising even for another sacred value. These values remain oughts even if they cannot be carried out

Violation of the unique value of persons so we lose something of irreplaceable value. This occurs when universal duties to all persons clash with specific roles and obligations

Those value conflicts which result in tragic choices—those choices where an individual’s capability to live a dignified life is unavoidably violated

The agent faces ‘impossible oughts’. Those actions which make the agent complicit with the evil projects of others and requires acting in ways that undermine/ destroy the agent’s sense of moral integrity

6.2.2 Hampshire’s Pluralism and Ineliminable Conflict

An alternative response to the ‘all-things considered’ argument is found among those who adhere to the Machiavellian Model of DH. Stuart Hampshire argues that DH arise because we face value pluralism in our complex societies, evidenced by different yet valid conflicting demands of a public and a private morality. Hampshire puts it this way:

Our everyday and raw experience is of a conflict between contrary moral requirements at every stage of almost anyone’s life: why then should moral theorists – Kantians, utilitarians, deontologists, contractarians – look for an underlying harmony and unity behind the facts of moral experience (Hampshire 1983: 151)?

He contends that human nature (conceived in terms of common human needs and capacities) always underdetermines a specific way of life with an order of priorities among virtues, which in turn underdetermines those moral prohibitions and obligations that support a particular way of life. If this is correct, then there are three powerful claims against those who hold the ET in favour of the RT. Firstly, there cannot be such a thing as a complete human good. Secondly, there can never be harmony among all the essential virtues in a complete life. And, thirdly, we can never infer what is universally the best way of life from propositions about human nature (Hampshire 1983: 155).

For Hampshire moral theories commit a serious error when they insist that there is, or ought to be, fundamental agreement on, or a convergence over, moral ideals. The ineliminable conflict of values exist both within a single human life and between people and groups within societies. As a result, there are different virtues, obligations, duties, and character dispositions needed for a public or political morality (within a successful political way of life) from those required of a private morality. This insight forces itself on us with the ‘perennial problem of the necessity of ‘dirty hands’ in great political and social enterprises (Hampshire 1978: 49)’. DH scenarios in politics reflect this reality because, firstly, they acknowledge that a politician’s public role entails a special responsibility for the consequences of their actions, one that is absent from private life. Secondly, they acknowledge that when politicians implement public policy, they may be forced to employ coercion and even violence to obtain worthwhile goals, such as protecting their citizens and ensuring fairness and justice. This will unavoidably produce a deep tension with a private morality that eschews such actions. In short, the special obligations and duties attached to the role of a politician requires different virtues (such as ruthlessness) and the willingness to use means that are not morally acceptable in private life. As a result, in some situations where there is a serious clash between public and private values, politicians are forced to dirty their hands. Contrary to the ‘all-things-considered’ argument, an agent will do right in her role of politician but also will do wrong in terms of her personal morality. So, for Hampshire, DH are conceptually possible and inevitable because there will always be competing ways of life and conflicting social roles due to a pluralism of values. These conditions leave a moral residue, a part of that rich and pluralistic moral tapestry of human life.

6.3 Tragic Remorse and the Case for DH

The final argument supporting the coherence of DH essentially turns on a deep disagreement about methodological issues. As noted earlier, those who argue for the coherence of DH scenarios cite relevant examples, the moral emotions they generate, and privilege (or at least take seriously) the experiential when it clashes with theoretical claims. In contrast, the arguments which reject DH are rationalistic and rely on abstract notions such as logical and conceptual principles of deontic logic or insist on a particular understanding of the nature of values and what this implies for ‘all-things-considered’ judgments. The latter view excludes or deeply distrusts the use of moral intuitions or emotions in deciding moral obligations. They insist that moral emotions can be misleading and, no matter how strongly we feel about a particular action, such feelings are not sufficient to show that an action was moral or otherwise. Feeling guilty does not entail being guilty and the converse is also true.

Yet, as those who support a notion of DH have argued, moral emotions need not be seen in this way and the DH revival restores moral emotions back to their rightful place in moral discourse as prior to, and constitutive of, moral thought. Emotions are at the centre of our moral sensibilities and an essential aspect of how we understand the world around us. As Stocker points out, ‘a life without emotions would be at best a pathologically deficient life—perhaps the life of a severe schizophrenic, psychopath, or sociopath (Stocker 1990: 31–32).’ In the case of DH, feeling ‘tragic remorse’ is a specific and appropriate reaction to the situation an agent faces given the morally dirty incompossible choices available. Tragic remorse is a special way of feeling guilt and shame. It differs from ordinary remorse in that it is not a feeling of being morally defective (as should be the case with those agents who commit ordinary moral wrongdoing), nor does it equate with shame and guilt that hates and condemns the sinner and the sin. Rather, it is based on a feeling of being violated, compromised, and morally dirtied. It also recognizes the unavoidable harm done in the attempt to bring about a lesser evil. This occurs because in certain circumstances we are forced to carry out an impossible ought and choose between evil options caused by the evil actions or projects of others. Put another way, we face non-negotiable moral requirements that involve the violation of persons, values and principles.

Tragic remorse, seen in this way, provides important reasons to accept DH scenarios as conceptually coherent and an important part of our moral reality. It acknowledges and responds to how we experience a distressing aspect of our moral reality. This moral emotion is not a vestige of a primitive moral sensibility or social manipulation (although in a corrupted form it clearly could be) but serves to focus on crucial evaluative information which is fundamental for all grounded moral theories. This information sets out one way in which we experience unavoidable moral conflict that results in moral violations. Acknowledging the reality of tragic remorse provides a proper starting point for developing a rational ethical system that accurately reflects this moral reality. Agents who dirty their hands will suffer an unavoidable loss of moral goodness, bear witness to the evil done, and feel the moral pollution for so acting. Tragic remorse acknowledges the pernicious nature of DH, where good persons, acting in good faith, are forced to engage in a particularly nasty form of justified wrongdoing thereby undermining their own integrity and moral virtue.Footnote 53

7 Concluding Remarks

If the responses in Sect. 6 are credible, then the charge that DH scenarios are conceptually incoherent is not persuasive. Or at least, I hope that I have provided sufficient reasons to harbour significant doubts. DH scenarios are frequently examined in literature, popular culture, political tracts, historical writings, films and TV dramas, yet contemporary moral philosophers insist that this very notion is incoherent. Perhaps, at least prima facie, moral philosophers may be failing to recognize something of great importance to our moral lives, and this ought to give them pause and the motivation to engage with the phenomenological arguments for DH.Footnote 54 Rationalists against DH need to explain why our social practices suggest a much ‘thicker’ or nuanced moral reality than they are prepared to sanction. By dismissing moral emotions as primitive or incoherent if they do not fit with a set of prior beliefs about the purpose, limits and nature of morality is itself a prejudicial way to engage in moral theorizing. There are alternative plausible views on these issues which can be supported by good arguments (as I have tried to show above) and they offer a better fit with the phenomenology of those experiencing DH scenarios. As Holbo rightly points out, accepting the validity and fundamental importance of the phenomenology of moral experiences is a

‘consequence of an antecedent, deeply-held conviction that, in life, matters of deontology and axiology (duties and goods) are rather promiscuously met and mated, not just with each other but with questions of character, virtue, moral psychology, judgment – perhaps as well as culture, politics, society and history’.

He adds that

…it takes a keen eye of the literary artist to capture life’s shifting ethical lights and shadows – to take in its full sweep and minute detail. This is why novelists and playwrights often seem ethically deeper and more insightful than the systems of the deepest thinkers (Holbo 2002: 271).

Finally, if DH are possible then it is important that we continue to refine our understanding of this moral phenomenon and explore its wider implications for a range of political, social, psychological, and meta-ethical issues.Without a clear account of DH scenarios and their moral implications, we lose the ability to accurately describe a fraught part of our moral reality and, instead, respond in inappropriate and damaging ways to these specific kinds of unavoidable situations of justified moral wrongdoing.