PUNISHMENT, VALUES, AND JUSTIFICATIONS

[☛ read the rest of the Symposium on Leo Zaibert, Rethinking Punishment (2018)]

Leo Zaibert*

I would like to thank the Centre for Ethics at the University of Toronto, and especially its Director, Markus Dubber, for organizing this Book Symposium around Rethinking Punishment. And I would, of course, also like to thank Alejandro Chehtman, Leora Dahan-Katz, Stephen De Wijze, Margaret Martin, and Sascha Ziemann for participating in it. It is a great honor – and also a tall order – to engage with the participants’ reactions to and criticisms of the book. Advancing the debate over punishment – which I described in Rethinking Punishment as having reached a stalemate – was one of my main motivations for writing the book. The thoroughness with which the participants have read the book and the nature and substance of their suggestions and objections bodes well, I think, for the potential that this conversation has to advance the debate.

Below I respond to each contributor separately, trying to focus on unique angles. Still, common themes and concerns run through some of the contributions. Thus some cross-referencing is inevitable, and aspects of my response to one contributor are to be found in my responses to other contributors.

1. Chehtman and the Distinction between the Axiological and the Deontic

Many, if not most, of Chehtman’s criticisms can be traced back to a general misgiving regarding the axiological turn that I recommend in Rethinking Punishment, which is accurately captured in the title of his contribution. This has been the sort of misgiving that I have encountered most often when presenting parts of the book, and a misgiving that is in fact discernible in some of the other contributions to this Book Forum as well. I would then like to devote a few lines to clarifying my main goals in the book before turning to Chehtman’s more specific criticisms.

One of the central themes in Rethinking Punishment is that we should pay more attention to the axiological dimension of punishment, particularly in light of the overwhelming emphasis that punishment theorists have traditionally placed upon the deontic dimension. The axiological dimension relates to questions of value (which are not necessarily, and at any rate not directly, action-guiding). The deontic dimension relates to what we ought to do, to our rights and duties, to our prerogatives and our obligations. Grosso modo, Chehtman’s general misgiving is that, independently of whatever merits the book may have in terms of “pushing the contemporary debate forward,” or of “broaden[ing] our understanding of punishment,” it fails to provide sufficient guidance in terms of assessing “our current practices of legal punishment.” This is a version of the “so what?” general sort of misgiving: what is the upshot of what I am saying – where “upshot” is understood in fundamentally practical, deontic ways.

There is a sense in which this misgiving has to be well taken: our current practices of legal punishment do call for serious, probing assessment. And doubtless one could assess them in more direct ways than I do in this book. Admittedly, Rethinking Punishment does not offer a list of policy recommendations, but it is important to underscore that the book was simply not meant to do that. A book focusing on those practical matters would have been an altogether different book. But I want to explain why the “so what?” misgiving does not worry me as much as it worries Chehtman (and, as we shall see, other commentators).

Chehtman’s view that “the central move Zaibert advocates in Rethinking Punishment is to challenge the type of question that a philosophical account of punishment must focus on” slightly – albeit importantly – misrepresents the main goal of the book. Although in Rethinking Punishment I indeed pay (what I take to be long overdue) attention to axiological questions, I do not think that only axiological questions matter, or even that they matter more than other types of questions. Nowhere do I deny that the sorts of practical problems with which Chehtman (and others) would have liked me to engage deserve attention. My view, rather, is that the theoretical problem which the axiological turn helps foreground is important too. Early in the book, and immediately after canvassing the importance of practical problems I put it like this:

By and large, it has been these sorts of practical problems that have attracted the attention of contemporary punishment theorists. This focus may partly explain the fact that “the problem of punishment” is often taken to be either exclusively or predominantly “the problem of state punishment.” There is no denying that these practical problems are both pressing and depressing. But this does not, I think, justify the comparatively little energy that has been expended in addressing the theoretical problem of punishment. In this book, thus, I attempt to shed much needed light on this theoretical problem (2).

And, as I put it towards the end of the book: “independently of any consideration of policy or strategy, there is a purely theoretical problem relating to the axiological complexity of punishment” which I describe as “unabashedly philosophical” (242).

My approach in Rethinking Punishment does seek to both diagnose and begin remedying the exaggerated fixation with the deontic dimension (to the detrimental of the axiological dimension) that I see in contemporary punishment theory. Independently of the importance of practical problems, to also pay attention to axiological, theoretical problems strikes me as salutary. Of course, disrupting received wisdom about punishment does not by itself prove that the axiological turn I recommend is indeed necessary, or that my pluralistic justification is, in the final analysis, better than its alternatives. But it is a way of enriching and reorienting the debate: the axiological “turn” does not mean abandoning non-axiological concerns.

As the title of the book suggest, I do mean to offer new ways of thinking about punishment. Some of this rethinking may involve “purely” (or at least predominantly) axiological matters. Consider, for example, the view I advance in the book that a much more theoretically sound starting point for classifying different justifications of punishment is to look at the purely axiological question as to whether or not there could be value in deserved suffering, as such. Not only does this move allow me to offer an alternative to the deadlocked debate between retributivism and consequentialism, but it further allows me to suggest that justifications of punishment that have traditionally been seen as diametrically opposed to each other are in fact rather similar. Thus, in my proposed new map of the (theoretical) territory both deontic retributivism and consequentialist justifications based on Benthamite, classical utilitarianism[1] – the two opponents in the debate, as typically construed – belong together under the heading of axiologically monistic justifications. Against these monistic justifications, I defend a pluralistic justification of punishment.

I would be satisfied about these (and other) theoretical contributions in Rethinking Punishment, even if they were “mere” exercises in conceptual analysis, for this is part of what I mean when I refer to my own approach as “unabashedly” philosophical. But I do not think that the book’s contributions are just that, for they often have obvious (albeit admittedly indirect) practical implications. After all, it is precisely this “theoretical” pluralism that I defend that allows me to expose the unsettling fact that punishment, even when justified, is dilemmatic and generates moral emotions in the punisher, such as guilt or remorse. And it allows me, too, to highlight the “simplemindedness” with which punishment theorists and policy makers tend to approach the subject.[2] Consider some other views for which I argue in the book: that punishers (and punishment practices) can be criticized even when they are justified; that punishment (even when justified) is best conceived as necessarily dilemmatic (and indeed as a case of dirty hands); that it almost always taints those who inflict it; that the sense of “justification” typically employed by punishment theorist is unduly narrow; or that the central axiological tension in punishment theory is between the value of punishment (in the sense of deserved suffering) and the value of forgiveness (in the sense of the deliberate refusal to inflict that deserved suffering). The axiological turn foregrounds the considerable disvalue that punishment, even when justified, generates. And it does this without committing the widespread twin mistakes of demonizing retributivism and of denying the value of deserved suffering. Surely if these views are correct they must have practical implications, even if downstream and indirect, and they would recommend that we approach the practice of punishment with much more humility and caution than we tend to do now.

Let us now turn to Chehtman’s two specific reasons for thinking that whatever practical implications my views may have, they strike him as “rather thin.” These reasons flow from his engagement with two of my theses in the book: first, my thesis that “inflicting deserved suffering is not (necessarily) bad,” and, second, my thesis that “we cannot fruitfully think about the moral value of punishment without also referring to the value of forgiveness.” I shall deal with these two in reverse order.

Chehtman’s dissatisfaction with my treatment of forgiveness itself breaks down into three closely interrelated parts. He disagrees with my “narrow understanding of forgiveness” because he thinks it neglects “other values which are arguably relevant for a plausible axiology of punishment.” He takes issue with the fact that I conceive of “forgiveness independently of other closely associated phenomena, such as repentance, reconciliation, and apologies.” And he finds it “somewhat perplexing that a book that continually complains specifically about ‘simpleminded’, ‘flat’, ‘narrow’, accounts, in favor of a ‘complex’ ‘properly’ or ‘richly pluralistic’ axiology contends that the underlying or competing sources of value in the practice of punishment are but two.”

Nowhere do I contend that only two sources of value matter for punishment. In fact, I explicitly recognize that there are “many” (3), and indeed “potentially infinite” sources of value (34). For a variety of reasons in Rethinking Punishment I do indeed choose to focus on the value of inflicting deserved suffering and the value of its merciful remission, but, again, not because I think that these are the only two values that matter. Similarly, nowhere do I deny that there could be value in reconciliation, or in repentance, or in apologies. While I do suggest that these things need not always be valuable, I explicitly admit that, in general, they are “worthy goals” (124).

What Chehtman calls my “narrow” understanding of forgiveness is the result of my taking a side on one of the central debates in the specialized literature on forgiveness: the debate between conditional and unconditional accounts of forgiveness. Roughly, defenders of conditional forgiveness see value in it only on condition that the wrongdoer has repented, or apologized, or that the very act of forgiveness is part of a process of reconciliation (etc.). In contrast, defenders of unconditional forgiveness – myself included – see value in it even absent repentance, apologies, or indeed any other condition. This is not to say that those other things are value-less or unimportant – it is just to say that, even if they are absent, forgiveness could be valuable. While I cannot here dwell on why defenders of unconditional forgiveness believe that it is the conditional account of forgiveness that is “narrow,” the above clarification should suffice to stave off the charge of narrowness. After all, embracing unconditional forgiveness is not to exclude anything. But I would like to briefly explain the relevance of my mobilization of unconditional forgiveness within the context of punishment theory.

There is, I submit, a parallel structure at work in forgiveness theory and in punishment theory: just as (some of us believe that) there could be value in deserved suffering in itself (independently of other considerations), (some of us believe that) there could also be value in its merciful remission (independently of other considerations), in itself. Again, this is not to deny that there could be value – and very often there is – in these other considerations. Nor is it to say that these things that could be valuable in themselves perforce defeat other values. But it is to reveal interesting and hitherto insufficiently explored connections between the debate about conditional and unconditional forgiveness, on the one hand, and the debate about retributivism and consequentialism, on the other. The reasons for endorsing conditional forgiveness are very similar to the reasons for endorsing consequentialism in punishment theory, and the reason for endorsing unconditional forgiveness are very similar to the reasons for endorsing (axiological) retributivism.

While Chehtman finds aspects of my views “illuminating,” he suggests that I am “rather ungenerous” to my opponents, and that I misrepresent them in that they are “precisely concerned about what is permissible to do even in situations in which it would be morally costly, and even deeply hurtful” to act. But recall that a very large number of my opponents are the classical utilitarians and the deontic retributivists: all of them monists. As I (and others) have argued, these opponents cannot really – on pain of incoherence – countenance the axiological conflicts I discuss in the book. Strictly speaking, their theories force them to insist that if we did the best we could (or what desert unproblematically commanded), then there really is no theoretical room for remorse or other moral emotions. These accounts could only cogently admit non-moral, generic sadness. They could of course instrumentalize these essentially misguided feelings and views for the betterment of society, but they cannot recognize or support them in earnest. This is what Williams’s talk of “simplemindedness,” P.F. Strawson’s talk of “one-eyed utilitarianism” and Isaiah Berlin’s talk of “technological problems” “simplemindedness,” seek to highlight. And I have sought to show the relevance of this discussion for punishment theory.

I wish Chehtman were right in that “no one” working in the ethics of self-defense “would suggest that there is no moral residue” in, say, “a mother killing her son if he was unjustly threatening her life.” The lesson of Williams’s, Strawson’s, and Berlin’s critiques of utilitarianism is that classical utilitarianism does not really have theoretical space for these residues. I am sure that not everyone working in the ethics of self-defense is (or thinks of herself as) a classical utilitarian. But, as I try to show in Rethinking Punishment, classical utilitarian views are extraordinarily widespread, and even critics of utilitarianism I have shown to be themselves utilitarians when it comes to punishment. (In the book, I show how these criticisms apply as well to deontic retributivism: they apply generally to monistic justifications.) Underscoring these theoretical implications may prompt those people working on self-defense who Chehtman has in mind – and who feel the weight of moral residues – to question whether their feelings are compatible with their underlying axiological views.

Chehtman’s third worry is that I say “very little, or at least not enough” about the deontic implications of the axiological tension between the value of deserved suffering and the value of forgiveness. Since this is both an instance of the same general dissatisfaction Chehtman has vis-à-vis the allegedly thin practical implications of my view, and in fact the same sort of objection he has against my other thesis – that, as he puts it, “deserved suffering is not (necessarily) bad” – I turn to that discussion at once.

My central view is that it is, in general, good when people get what they deserve, even when they deserve to suffer. But since I readily grant that the default value of suffering is bad, the case of deserved suffering is especially problematic. (As we shall see, Dahan-Katz suggests that perhaps my view really is that suffering is invariantly disvaluable.) In the specific case of punishment, my view can hardly be understood unmoored from the context of the discussion of organic wholes and of moral particularism. Strictly speaking, then, my view is neither that deserved suffering is intrinsically good nor that it is not intrinsically bad. (Incidentally, it appears to me that Chehtman equivocates somewhat between these two different positions.) Rather, my view is that deserved suffering can by itself (absent any consequences) add value to organic wholes. (I will return to this topic in my discussion of Dahan-Katz.)

Chehtman’s worries here are legitimate, for we are not clear as to when exactly deserved suffering would add value, how much value it would add, or even why exactly it would add it. These are, I am afraid, consequences of taking seriously the sort of flexibility afforded by the approach based on organic wholes and moral particularism. Of course, I understand that to offer this explanation is not to assuage Chehtman’s worries. But it is to put them in their proper place. While Chehtman’s worries are pressing, they do to an extent underestimate the problems with the alternative (extant) approaches to the justification of punishment. The classical utilitarian, for example, has little – if any – difficulty assuaging these worries – even though, for other reasons (which I discuss in the book) her approach is rather untenable. And something similar goes for the deontic retributivist. My view does face the difficulties that Chehtman mentions – but these are difficulties worth having, and which contemporary punishment theorists have ignored for too long.

I am not sure that Chehtman disagrees with my view regarding the fact that deserved suffering may add value to an organic whole. But he is not sure “what is behind this very strong intuition.” He again worries that even granting that my view in this regard is correct, the view simply fails to provide enough practical guidance. It also fails to offer ways to determine how “weighty the intrinsic goodness of deserved suffering is and how this would translate for a deontic account of legal punishment.” Thus, Chehtman offers examples of concrete situations in which my view evinces “very little moral traction in terms of allocating the right (power) to punish between different authorities or people.” And he insists that it is precisely these sorts of situations that “a deontic account of legal punishment needs to illuminate.”

My account, however, is explicitly not a deontic account of legal punishment. Not because I think that deontic accounts of legal punishment are bad, or inferior to the sort of axiological approach I undertake in Rethinking Punishment. I agree with Chehtman in that one’s axiological position regarding punishment better have some practical consequences. And I do not believe that in Rethinking Punishment I have spelled out all those (or even many) of those concrete practical consequences. Much needs to be done, although determining what exactly should result in terms of policy, criminal legislation, or indeed of the matters having to do with the extraterritorial validity of the criminal law that interest Chehtman, as a result of the rethinking I recommend are not amongst my goals in the book.

My goals in the book are at once humbler and more ambitious. They are humbler in that they do not seek to provide the concrete guidance Chehtman (and others) would like to see. They are more ambitious in that they invite a real rethinking of the very way in which we do punishment theory. If these latter goals are even partially successful, they should generate a general change of attitude vis-à-vis punishment, and there is every reason to suspect that this new attitude will not be deontically inert.

2. Dahan-Katz, Monism, and Mereological Axiology

Dahan-Katz raises a number of criticisms against Rethinking Punishment. I can quickly address the criticism that she appears to consider most serious, since it is a ratcheted version of the “so what?” criticism that we have just discussed. So, regarding her general misgivings concerning the deontic implications of my views in Rethinking Punishment, I will simply emphasize once more that in urging punishment theorists to pay more attention to axiological matters I am not dismissing deontic matters. To the contrary, I believe that such consideration is per force bound up with a reassessment of practical consequences punishment, even if my book is not directly focused on the latter.

Admittedly, there is one important difference between Chehtman and Dahan-Katz. Whereas Chehtman harbors doubts as to the potentially very “thin” deontic implications of my views, Dahan-Katz finds my view “incapable of bearing fruit with respect to the question of how one ought to act” – and she believes that this incapability is “structurally guaranteed.” I must confess that I simply cannot find Dahan-Katz’s arguments for this position. So I will discuss some of her other criticisms.

Dahan-Katz believes that I unfairly criticize “nearly all existing theories of punishment” when I claim that, in different ways, they are “simpleminded, impoverished, and flatfooted.” She further believes that this view of mine is based on the “grandly stated charge of monism,” and this is why my view “appears exaggerated.” She suggests I lump together

a broad swath of views, including all deontic retributive theories, as well as communicative theories of punishment, and views as diverse as those held by Scanlon, Nussbaum and Tadros, marshaling the same charge of monism against these theories as well.”

Chapter three of Rethinking Punishment is devoted to problems with utilitarianism, chapter four to problems with (deontic) retributivism, and chapter five to problems with communicative theories. An obvious initial counter to Dahan-Katz’s criticism is to point out that I have explicitly presented the theories I criticize in order of increasing plausibility, and in doing so I have explicitly distinguished among them. But there is much more I would like to say in response to this general criticism.

It is easy to show that I have not marshaled the charge of monism of which Dahan-Katz speaks against communicative theories. For I expressly admit that the communicative approach appears to be both the “most promising approach to the justification of punishment” (28) and “particularly well equipped to underwrite a pluralistic approach to punishment” like mine (119). My main worry vis-à-vis the communicative approach relates to its refusal to engage with moral conflicts. This refusal leads Duff (arguably the most influential contemporary communicative theorist) to dismiss important aspects of moral inquiry by claiming that mercy/forgiveness is an “intrusion” into the logic of punishment.[3] It is for this reason that, in my view, his communicative approach fails to be properly pluralistic. As I explain in the book (23 ff.), a properly pluralistic justification of punishment focuses on these two values (punishment and forgiveness/mercy); Duff’s theory is pluralistic, but not properly so. I find Tasioulas’s communicative approach more promising because it explicitly seeks to be properly pluralistic.

So, I do not suggest that the communicative approach is monistic. Dahan-Katz, moreover, admits that my charge of monism

appears well suited to fit the theories of punishment he criticizes, most straightforwardly classic utilitarianism, which entertains only one intrinsic good in terms of which all other putative goods are to be expressed – pleasure, as well as Kantian retributivism.

It is not easy to see, then, what is wrong with my monism charge.

Perhaps, however, I have unfairly lumped punishment theories together not quite because their monism but in some other way. Dahan-Katz mentions three specific authors that I may have lumped together unfairly: Thomas Scanlon, Martha Nussbaum, and Victor Tadros. In the early chapters of the book I argue that many superficially disparate approaches to punishment share the quintessential axiological tenet of Benthamite utilitarianism: that suffering is invariantly bad, that punishment is in itself evil, as Bentham famously put it (8, ff.). I argue that these three authors endorse this axiological tenet. Very briefly: Scanlon believes that the idea that there may be value in inflicting suffering on those who deserve to suffer is “morally repugnant” (64 ff.); Nussbaum (now) believes that Bentham’s monistic suffering-reduction agenda constitutes a “deep insight” (184 ff.); and Tadros admits that he shares Bentham’s “idea that the suffering that punishment causes is not intrinsically good” (62 ff.).

As I recognize in the book, these authors have advanced influential and indeed admirable criticisms of many aspects of Benthamite utilitarianism. Still, the similarity I underscore in Rethinking Punishment is perfectly real. It is only when it comes to punishment that these generally non-utilitarian authors can be seen as Benthamite utilitarians, at least in the sense of endorsing its main axiological tenet. This typically overlooked or underestimated commonality is quite revealing in terms of the pervasiveness of utilitarian modes of thinking. Doubtless there is room for disagreement: perhaps endorsing what I take to be the main axiological tenet of classical utilitarianism has fewer (or less momentous) consequences than I think it does, etc. But those potential disagreements would not render unfair my grouping together justifications of punishment that do share an endorsement of the view that deserved suffering is invariably bad (barbaric, repugnant, etc.).

Despite admitting that I am right about Kantian deontic retributivism being monistic, Dahan-Katz’s finds my view that deontic retributivism is monistic “striking.” While some of Dahan-Katz worries here stem from a genuine disagreement between us, I think others stem from a misunderstanding of my views. The genuine disagreement concerns my view that “retributivism, first and foremost, should be seen as an axiological doctrine” (14).

Retributivists can disagree as to the complicated questions concerning the deontic implications of their axiological commitments. But no retributivist can deny the axiological point that deserved punishment is intrinsically valuable. In this sense, then, this axiological commitment is revealed as absolutely essential to retributivism in ways that no other commitment is. Without adhering to this axiological commitment, you are not a retributivist (15).

Dahan-Katz disagrees: “the deontic retributivist need not, despite Zaibert’s protestations, grant the intrinsic goodness of suffering.”

The general discussion of intrinsic value is famously fraught with difficulties. The specific discussion of the intrinsic value of deserved punishment is no different. I use the expression in order to capture the sense in which deserved punishment can be good independently of whatever downstream consequences it may bring about. If you agree with my view, you necessarily disagree with Bentham’s dictum that punishment (deserved or otherwise) is in itself evil, and vice-versa. There are only two options: either you agree that punishment is in itself evil, or you do not.

I am all for unsettling received wisdom. For example, in the book I make much of the fact that G.E. Moore – a utilitarian (of a special, ideal sort) – can be seen as endorsing retributivism. But a Benthamite retributivist is a contradiction in terms. And this means that if you are a retributivist, you must agree with the view that deserved suffering can in itself be of value – otherwise you would be a Benthamite utilitarian. Contra Dahan-Katz, then, a retributivist – deontic or otherwise – cannot deny the intrinsic goodness of deserved suffering (in the sense sketched), because denying entails embracing Benthamite utilitarianism, and these two positions are mutually exclusive. I stand by my view, and am genuinely curious as to how Dahan-Katz may conceptualize retributivism so as to disagree with this point.

This disagreement brings the misunderstanding I mentioned earlier to the fore. By way of a counter-example to my view, Dahan-Katz claims:

Deontology further admits the possibility that the duty to punish is not immediately grounded in the intrinsic goodness of deserved suffering. This might be, for example, because it bottoms out at further deontic (rather than axiological) claims or because it is not the value of deserved suffering that underlies the duty, but some other value.

There is quite a bit of contraband in this passage. Dahan-Katz smuggles the word “immediately,” and with it the idea of temporal connections between values and duties – a topic unrelated to my goals in the book. But Dahan-Katz smuggles another term: “deontology,” and this is the crucial misunderstanding I want to discuss. Throughout her contribution, Dahan-Katz, it seems to me, uses “deontic” and “deontological” interchangeably.

While it is committed to a deontic position vis-à-vis deserved punishment – that there is a duty to impose it – a more complex deontology than countenanced by Zaibert allows that this pro tanto duty (or obligation, if one prefers) can meaningfully conflict with others within the context of punishment. […] Deontology can further countenance insoluble conflicts of a more robust nature, wherein neither duty overrides, allowing the kind of conflict between desert and forgiveness that Zaibert seeks in axiological retributivism.

Crucially, however, I do not use “deontic” and “deontological” interchangeably. I am interested in the deontic/axiological distinction (as explained early in the book’s opening pages, and as I repeated in my reply to Chehtman, above), not in the very different deontological/teleological distinction (about which I have very little to say in the book). In any case, paying more attention to axiological questions – as I entreat punishment theorists to do – is compatible with endorsing a number of moral theories, including (complex) deontological normative ethical theories.

While I, for a variety of reasons, harbor doubts about traditional deontological and teleological approaches, I do not deny that there could be complex deontological and teleological positions. I can hardly be guilty of countenancing an insufficiently complex deontological framework – for I have neither countenanced nor rejected any deontological framework whatsoever. Again, this discussion is orthogonal to my concerns in Rethinking Punishment. The complexity of which Dahan-Katz speaks in the passage just quoted is a property of normative moral doctrines, not of the general axiological or deontic aspects or dimensions I discuss in the book.

I must mention one more problem stemming from the confusion between “deontological” and “deontic.” A paragraph devoted to showing how “deontology” can accommodate moral dilemmas ends with Dahan-Katz claiming that “nowhere is his dismissal of deontic views more absurd than in his insinuation that only the axiological has the resources and motivation to deal with” the fact that the US constitution permits slavery as a punishment. I guess I would agree with Dahan-Katz that the bald view that only non-deontological moral doctrines have the tools (and motivations, etc.) to deal with these sorts of cases would be absurd. That is evidently not my view. Rather, my view is that overly deontic approaches, which operate with a narrow and mechanistic understanding of “justification” (etc.), are likelier to bracket out the axiological horror of the sheer fact that slavery is an allowed punishment in the United States.^[4]

I would like to conclude my discussion of Dahan-Katz’s reaction to Rethinking Punishment by paying attention to what she considers “only a minor concern,” but which I consider a particularly interesting line of criticism, since it forces us to engage in precisely the sort of axiological discussion so frequently absent in contemporary punishment theory.

Probingly, Dahan-Katz suggests that since my central thesis is that suffering can add value to organic wholes, I am not actually denying that suffering, qua part, remains always bad. Thus, although I claim that suffering has a default disvalue, I am unwittingly committed to the view that its disvalue is invariant. If this is so, then it appears that I am after all in the same camp as my (utilitarian) opponent: for we all find suffering invariably bad. This is why she suggests that, rather than “invariantism,” I reject “exclusivism.”

What Dahan-Katz seems to me to underestimate is that, even if I were an unwitting invariantist along the lines she suggests, my position would still be quite different from that of my utilitarian opponent. Unlike Bentham (or Scanlon, or Nussbaum, or Tadros, etc.), I can claim that an organic whole may gain in value by the presence of deserved suffering alone. My truly invariantist opponents cannot do this. This seems enough to render myself a “variantist” about the value of suffering. And I think that my position is consistent with both the part qua part changing its value and with not changing it.^[5]

Dahan-Katz, however, offers a second reason why I may perhaps be an unwitting invariantist: she suggests that “denying the invariant negative value of suffering would seem to be at odds with Zaibert’s defense of the necessarily dilemmatic nature of punishment.” My response here is to underscore that there can be serious moral conflicts between “merely” default values – conflicts that leave remainders and that taint us in significant moral ways. To refer to a value as “default” is not to say that it is somehow not real or somehow unimportant. It is simply to say that this value can change depending on the context. My central point is that the badness of suffering does not vanish into thin air if punishment is justified – and, again, this is consistent with this badness being default or invariant.

I think that this helps clarify a related point. Dahan-Katz claims: “per Zaibert, forgiveness is intrinsically valuable and is in terminal conflict with the intrinsic value of deserved punishment.” Furthermore, she claims that “in what is perhaps the greatest lacuna of the book, Zaibert offers no account of the value of forgiveness – no reason to accept that it is necessarily valuable.” I do not think I ever used the expressions “terminal conflict” or “necessarily valuable,” and I am not sure I know exactly what they mean. But I wish to underscore once more that the sense in which I say that forgiveness (or deserved punishment) is “intrinsically valuable” is not that it is “always” valuable.

A leitmotif in Rethinking Punishment is that both forgiveness and punishment stand in need of a justification. I reject both the demonization of punishment and the sanctification of forgiveness:

I do not believe that punishment is necessarily barbaric or that forgiveness is necessarily admirable. Deserved suffering sometimes has value (in the sense that it can add value to the organic wholes in which it appears) independently of any further consequence, and the same is true of forgiveness (183).

I do maintain that deserved suffering can be valuable in itself (in the sense already explained whereby by itself it adds value to an organic whole), and that forgiveness can be valuable in itself (in this same sense). Perhaps these two positions are too stipulative, or too “intuitionistic,” as Dahan-Katz (and others) suggests. But the denial of these possibilities continues to strike me as extraordinarily implausible (uninspiring, perverse, etc.), and I can only here refer to my remarks on intuitions in the book (76 ff.).

Contentious and difficult as these axiological matters unquestionably are, I think that they are terribly important, and that punishment theorists have neglected them for far too long. Disagreements about these matters are worth having. Thus, I hope that Dahan-Katz turns out to be correct when she predicts that “the debate over whether this [the axiological turn] is a turn we ought to take will no doubt be a source of fruitful critical reflection and discussion, driving punishment theory precisely in the direction that Zaibert intends to steer.”

3. De Wijze, Infinite Regresses, and the Cheapening of Innocence

De Wijze is not centrally concerned with the deontic implications of my account, however incipiently developed thus far, or however “thin” they may in the final analysis turn out to be. De Wijze and I agree that, as he puts it, “there is more, much more, to morality than how we ought to act in certain situations.” So, he recognizes the value in the framework that I advance in the book, independently of practical considerations. He in fact sees my project as rather broad, and as “necessarily theoretical.”

In earlier work, De Wijze explored the thorny question of how to punish dirty agents – if they are to be punished at all – and he confesses to having “struggled to understand why the punishment theories on offer were so problematic.” Generously, De Wijze claims that thanks to Rethinking Punishment, he now better sees that there simply were “no existing theories of punishment that were suitable for the problem of DH.” This is, of course, not to say that De Wijze does not have objections to my account – even though they are quite different form the ones I have discussed thus far. Before addressing De Wijze’s objections to my views in Rethinking Punishment, it is important to briefly go over the reasons for his dissatisfaction with the ways in which the standard debate concerning the justification of punishment plays out in the context of dirty hands.

Justifications of punishment based on Benthamite utilitarianism (which, as I argue in the book, constitute the vast majority of what is on offer, despite their defenders’ protestations to the contrary) are evidently inadequate, since we want the dirty agent to act as she did – we want her to do wrong (in the context of doing right) – and then there seems to be no forward-looking rationale, say, prevention, that would be realized by punishing someone for doing what we wanted her to do. And “we want her” to act as she does, because we deem it, all-things-considered, better that she acts in this way than in any other. Deontic retributivism does not fare better: what exactly is the desert basis in the case in which someone’s action was justified (and we would like the person to do it again, if the occasion arose, etc.)? Drawing on views developed more fully in Punishment and Retribution, in Rethinking Punishment I also argue against the so-called “mixed justifications” of punishment. I argue that either they do not really “mix” much, or the “mixing” is theoretically unstable and at any rate incapable of transcending monistic and mechanistic strictures.[6] De Wijze’s work on punishing dirty hands reveals yet another dimension of the inadequacy of the mixed justifications of punishment: they are as profoundly incapable of illuminating the punishment of the dirty agent as are the “unmixed” justifications upon which they sought to improve.

At this point, the idea that perhaps the solution to this problem is to deny that dirty agents should ever be punished may look appealing. But this strikes me as self-servingly stipulating the problem away. This strategy should be anathema to me, since I after all claim that all (or virtually all) punishers are dirty. Were I to accept this strategy, I would be unwittingly forced, regarding agents with dirty hands, into a sort of abolitionism antithetical to some of my central views in Rethinking Punishment and elsewhere. Plus, I would be engaging on precisely the sort of facile refusal to recognize the world in its real complexity that I have criticized. Wisely, De Wijze also avoids this stipulative, self-serving path. But then what is to be done?

De Wijze has explored three different ways of perhaps beginning to justify the punishment of dirty agents: by attending to cathartic (or cleansing) reasons, by considering the suffering that they may have themselves (or the circumstances) brought about, and by noting the unavoidable causal connection between the agent and the deed in these cases (reminiscent of the sorts of considerations at play in Williams’s famous example of the lorry driver who faultlessly kills a child which I discuss in chapter four). It should be obvious that I am quite sympathetic to De Wijze’s emphasis on the importance of these factors. But I also agree with him in that these considerations do not seem to provide the full-blown justification that we are trying to get. But, given how much our stances share, how can he object to my account?

Let me take De Wijze’s two inter-related objections in order. The first objection takes as its point of departure the almost exasperatingly provocative way in which Michael Walzer ended his seminal article “Political Action: The Problem of Dirty Hands.”[7] Although Walzer does not say it explicitly, or discuss it systematically, his view appears to be, in embryo, precisely the view that I explicitly and systematically defend in Rethinking Punishment: that to punish is to get one’s hands dirty. Roughly, if we are justified in punishing dirty agents, and we then punish them, then we ourselves become dirty, and then presumably liable to be punished ourselves and on and on: an infinite regress would obtain. Now, it seems to me that the objection is as pertinent against my views as it is against De Wijze’s own views, or as it would be against anyone who may (sometimes) justify the punishment of the agent with dirty hands. But this is no consolation: De Wijze’s objection is worth our attention beyond tu quoque invocations.

I think that in Rethinking Punishment I offer some tools to at least begin dealing with this sort of objection (even if perhaps not to completely answer it). I have just presented the gist of Walzer’s passage that motivates De Wijze’s objection. Interestingly, however, the way in which Walzer formulated the problem is slightly different from the way in which (for expository reasons) I have just presented it. After describing cases in which we are supposed to punish dirty agents, Walzer, as De Wijze reminds us, concludes his famous article with the provocation: “We won’t be able to do that, however, without getting our own hands dirty, and then we must find some way of paying the price ourselves.” The first thing that I would like to note is that in the concluding paragraphs (immediately preceding the final sentence of the article just quoted), Walzer had been talking quite explicitly about the punishment (or non punishment) of dirty agents in preference of the odder “paying the price” locution that he used by the end of his famous article. In the last sentence of the article the word “punishment” has altogether disappeared. I do not believe that this slight point about word choices changes the essence of the problem. But I do not believe that this is merely cosmetic either. The word change, in conjunction with Walzer’s ostensibly evasive “we must find a way” locution, signals in the direction of a solution.

That solution is, again, to take the distinction between axiological and deontic considerations seriously, as in Rethinking Punishment I suggest we do. Strictly speaking, Walzer has identified a potential infinite regress only if we assume that we must punish the dirty agent. But whence that assumption? My view is that there can be value in inflicting deserved suffering, despite suffering simpliciter’s default disvalue, a disvalue that can be affected by desert, etc. To the extent that some dirty agents may deserve to be punished, then there may be value in punishing them. Importantly, this potential value, however, could coexist or conflict with a variety of other things that could also be valuable – including, centrally, the merciful remission of this deserved suffering. It is at least prima facie plausible that in normal circumstances the dirt in a person who punishes a dirty agent may get increasingly less significant by dint of each recursive iteration down the path to the infinite regress, somehow asymptotically tending towards infinitesimal significance. At any given point down this path, the value of inflicting deserved suffering on a dirty agent may be lesser than the value of, say, forgiving her. Nothing in Rethinking Punishment commits me to the view that punishing a dirty agent is always more valuable than forgiving her, or even than refusing to punish her in ways that are not instances of forgiveness.

It would be a serious misunderstanding to parse my view as if it held that the fact that a certain suffering is deserved provided us with a duty to punish. But it would also be a serious misunderstanding to assume that thereby all I am saying is that desert gives us a mere right to punish. Desert is not merely a necessary condition for its infliction being justified – but it is not a sufficient condition either. It is an important condition – and this is a fact that gets lost in the maze of rights and duties that, via consideration of Williams’s criticism of the morality system, I explore in Rethinking Punishment. That a certain punishment is deserved is important, and we are much better off not straightjacketing this notion to the workings of rights and duties. This point holds identically whether we are punishing a garden-variety wrongdoer, or someone with dirty hands.

Of course, my solution to the infinite regress may be resisted, in that it may appear too abstract, or too formal, and it would immediately invite the sort of demand regarding the deontic implications of the axiological turn, as we have just seen Chehtman and Dahan-Katz do. For my solution does not specifically tell us which dirty agents should be punished and which should not, or to what extent, or by whom, etc. Once again, my answer is to emphasize the theoretical nature of my approach. As it turns out, then, it is not only that much needs to be done in order to spell out the exact, concrete practical consequences of the approach I recommend: much needs to be done, too, to spell out some of the purely theoretical implications of my approach. Having said that, however, I do believe that my approach does contain the resources necessary to formally stop this specific infinite regress that worries De Wijze, and to begin to frame a more substantive response as well.

Responding to De Wijze’s second objection is, I think, more complicated. And here there is a genuine disagreement between De Wijze and myself. De Wijze takes issue with my view that

[n]o punisher, no matter how just, is ever fully morally innocent. A punisher is in the business of making people suffer, and this business, even when the suffering is deserved (or medicinally useful), simply cannot be done without – in admittedly different degrees – dirtying one’s hands.

De Wijze worries that my position runs the risk of

undermin[ing] the seriousness of cases where we do indeed get DH and lose our moral innocence. There needs to be a threshold where punishing does indeed result in a violation, which is serious and morally tainting. […]

And De Wijze wonders whether my view does not in fact

hollow out the notion of innocence so that the term becomes unhelpful? And if this is right, does this point to a serious internal problem with developing a pluralistic account of punishment?

These are two very serious questions. But I think that part of an answer to the second question can be developed from the strategy I discussed above as to how to avoid the infinite regress regarding the dirt that may be generated when we punish dirty agents. The person who punishes someone who punished someone who punished (etc.) is a wrongdoer in a special sense (she did wrong in the context of doing right, etc.), and she strikes me as less dirty – and more “innocent” – than the initial punisher, and presumably much less dirty – and much more innocent – than the initial (non-punishing) wrongdoer. It is important that De Wijze’s sense of “innocence” here is subtle; he is not talking about legal innocence. The innocence he has in mind is similar to that which worried Melville in Billy Budd, which I discuss in the last chapter of the book.

This consideration also highlights an important distinction between this possible infinite regress and the one discussed earlier. Unlike this one, the infinite regress discussed earlier did not concern comparative innocence. The earlier regress was, as it were, purely formal. But the current possible infinite regress is anchored in something more substantive. The wrongdoer, who after all initiates the sequence anyhow, is presumably dirtier than his punisher. (Some wrongdoers can even be “dirty all over,” and this would not apply to agents with dirty hands, as I discuss in the book (227 ff.). It is not necessarily the case, and in fact I suspect it is, as a matter of fact not very often the case, that the wrong to be punished is itself a case of dirty hands.) And then if the punisher of the wrongdoing is to be punished too along the lines we are discussing, it does seem to me likely that her “wrong” is much less so, and much more qualifiedly so than the initial “unproblematic” wrong.

Whatever merit this solution may have, I do not think that it fully addresses De Wijze’s worry. And now we are back to dealing with the first question De Wijze poses as to whether I am in effect hollowing out the very notion of “innocence” (and, eo ipso, the related notions of “dirt,” “pollution,” “taints,” “guilt,” and so on) to the point of cheapening them, or at least of rendering them far less helpful. Here it is crucial to point out that I do not think De Wijze’s invocation of helpfulness is a slip into the sort of utilitarian approaches we both reject. Rather, it is to be understood in light of his pithy remark “if the guilt and moral loss spreads too widely or without end [as Zaibert suggests], it degrades the concept of DH, which now becomes so commonplace that it loses its salience and critical sharpness.”

So, I guess it is possible that De Wijze’s worry would remain pertinent even if my response above were accepted. In other words, even if the dirt, the guilt (etc.) do diminish with each iteration in the chain of punishing punishers, perhaps this is still enough “spreading” around so as to jeopardize critical sharpness. This is really a difficult issue, and I am not sure that the two points that I will make in closing may suffice. But I hope that they constitute at least the beginning of an answer.

The first point is that in light of the excesses of many contemporary, and presumably enlightened, criminal justice systems, instilling a sense of guilt in punishers (and others) has to be helpful. All too frequently punishers assume that if that the punishments they inflict are justified, then they are doing nothing wrong and their hands are entirely clean. Even perceptive critics of these dysfunctional and often downright cruel criminal justice systems seem at times too quick to take an attitude whereby by adopting this or that new take on the justification of punishment, all the suffering that these systems generate would somehow become palatable.[8] I, on the other hand, wish to underscore the reality of the default disvalue of suffering through and through. And I want to highlight the huge and often overlooked truly moral stains that punishment, even when justified, tends to generate. While I am no abolitionist, and I evidently believe that punishment is a necessity of life – both in public and personal contexts – I wish to oppose the ways in which we tend to uncritically “normalize” it by employing narrow senses of “justification.”

In Rethinking Punishment I have tried to problematize the notion of “justification.” I have drawn inspiration from a remark that in a different context Hans Welzel made as he criticized certain accounts of the criminal law whereby the killing of a person in self-defense would really be no different from the killing of a mosquito (since, after all, acting in self-defense means that our act is not wrong), a position that Welzel understandably found unacceptable.[9] Similarly, killing (or inflicting suffering on) a person as a justified punishment is not the same as killing (or inflicting suffering on, if possible) a mosquito. Clearly, the action against a person may generate moral remainders in ways that the action against the mosquito (or some other animals) may not. Executing an ex hypothesi deserving criminal should never (or virtually never) be done guilt-free. And, again, emphasizing this point seems to me, in our current political climate, not unhelpful.[10] (My discussion here is not meant to constitute an endorsement of the death penalty.)

The second (and last) point is that it is of course not this practical consideration that mainly animates me. There are theoretical reasons supporting my approach. The risk of cheapening important moral notions that legitimately worries De Wijze still needs to be weighed against other risks, above all the risk of capriciously deciding where to decree (as it were) an end to guilt. Beneath this or that threshold, there is no longer dirt, and innocence is no longer jeopardized. This is problematically reminiscent of the ways in which “threshold-deontology” may amount to the view that consequences do not matter unless they matter.[11] The fact is that punishment is a matter of inflicting suffering, and inflicting suffering has certain normative implications even when it is justified: it generates guilt, of a certain sort. This fact does not go away: it remains. But, the intensity of the guilt may be quite varied. The intensity of the guilt is diminished when the wrongdoer is extraordinarily bad – say Hitler, or Ohlendorf – or when the wrongdoer did wrong in the context of also doing right – e.g., when she punished an ordinary wrongdoer – or, even more so, when she punishes the punisher of a punisher (of a punisher… etc.). To dismiss the normative significance of inflicting suffering, even on the deserving (so as to avoid otherwise cheapening our moral notions), strikes me as perilously ad hoc. Maybe that dismissal does not quite cheapen the notion of guilt in De Wijze’s sense, but it strikes me as a cheap theoretical move nonetheless, even if in a different sense of “cheap.” At any rate, this is a problematic move I wish to resist. (For a different context in which arbitrary line-drawing becomes an issue, see my response to Ziemann, below.)

4. Martin, The Variety of Moral Taints, and Universal Lessons

It is one of the many virtues of Billy Budd, Melville’s masterpiece, that it lends itself to so many different, rich, and pregnant interpretations. Martin offers an interpretation of a crucial passage in Billy Budd that importantly – and fruitfully – differs from mine. The passage concerns the news report of Melville’s imaginary weekly naval chronicle “News From the Mediterranean,” in which the events on board the Bellipotent are recounted with remarkable flatness – a flatness that contrasts beautifully against the depth with which Melville recounts them in the course of the novel itself. The news report, moreover, is also marred by obvious falsehoods. I suggest that “for my purposes, it would have been even more useful had the report not been marred by obvious factual errors” (238). And it is at this point that Martin’s disagreement with me arises.

In Rethinking Punishment I suggest that there is something in the nature of news reports that renders them incapable of conveying the whole and complicated truth of cases such as Billy’s. My aims in this regard do not relate to literary analysis, or to a discussion of the limitation of this or that literary or journalistic genre. The reason why I turn to Billy Budd in the book is that, independently of its literary merits, its engagement with the complexity of moral reality and moral emotions and its way of plumbing the depths of the human condition allow me to avoid the drab schematism of trolley problems and other reductive approaches to punishment theory in particular, and ethics in general. As indicated in the book, Melville’s worries in Billy Budd are reminiscent of the worries of Cheshire Calhoun, of Martha Nussbaum (earlier in her career), and of Gary Watson, that I discuss in the book. The worries have to do with the way in which the complexity of moral reality seems to be sometimes ineffable. When such ineffability obtains, or when we approach it in a very high degree, then the whole business of inflicting suffering – however deserved – becomes increasingly problematic.

Martin’s interpretation is different. Quite independently of our limitations to capture moral reality in its glorious complexity, Martin suggests that it makes sense to assume that the report was “intentionally falsified” – an assumption she considers “highly plausible.” I have no doubt that there exists some textual evidence in Billy Budd to read it as a tale capturing the conflict between (non utilitarian) moral principles and (utilitarian) concerns with life’s expediencies. Though, of course, I think that there is more textual evidence to interpret the novel as I do than otherwise. Under Martin’s interpretation, intentionally falsifying the news report surely could have “served the double purpose of safety and deterrence,” of reminding the intended audience of “News from the Mediterranean” – sailors, above all – that misconduct on board her majesty’s ships will not be tolerated, and that it will be punished swiftly and severely.

One may then think that Martin’s interpretation of the news report is in line with these traditional readings of Billy Budd from which I have distanced myself. But I think that this does not quite capture Martin’s elegant move. She agrees with me on the tragic nature of the events on board the Bellipotent – a point that these traditional readings can have trouble even countenancing. But the tragedy she sees is not quite the one I see. By Martin’s lights, Vere is devastated because “he thought that Billy had to die for the sake of the safety of hypothetical future ships”; he was moved by “a commitment to the good of the universal seafearer.” The falsified news report is but a “noble lie” in the service of the greater good – just as Billy’s very execution was also in the service of a greater good. Now, tragic as this version of Vere’s devastation may turn out to be, it surely does not rise to the level of tragedy I see in the news report (and, even more, so in the imagined “accurate” news report I imagine). For my take on Vere’s devastation follows his realization that he is doomed to do wrong independently of any future (medicinal) consequences: no matter what he does or forbears from doing.

Since Martin does agree with me on the essentially tragic nature of the story, it may seem a minor point to elucidate which of our interpretations is more accurate. But I want to mention two reasons why my interpretation may be preferable. First, and repeating a point I made in Rethinking Punishment (229 ff.), it just strikes me as implausible that Vere would be so profoundly devastated if his dilemma was merely stemming from cost-benefit analyses. Why, for example, should he be described as metaphorically dead as a result of informing Billy of his fate? Granted, garden-variety calculators – as I think Martin’s interpretation suggests Vere is – can be more or less affected by some of the costs of their actions. But these costs are different from taints, both in degree of intensity and in kind. It seems to me that Melville intended to portray Vere as utterly devastated by what he had to do – a devastation that strikes me as out of place within the context of mundane cost-benefit analyses.

Second, and closely related, while Martin’s interpretation is not utilitarian, classical utilitarians can mobilize their revisionist agenda within the context of Martin’s interpretation, but not within the context of mine. For they may want Martin’s Vere (and us, in general) to understand that even in cases in which the sacrifices may appear to be too significant, or too painful, we should always remember that if we have done the best possible thing we could have done, there really is nothing to morally regret, or even feel bad about. And, if we embrace Martin’s interpretation, they may insist that Vere should get over, and “snap out” of, his devastation: feeling bad about doing the right thing is but a remnant of his (our) puerile, romantic ethics. The classical utilitarian cannot do this very compellingly if my interpretation is correct, however. After all, I see these sorts of emotions as irreducibly moral, and as intimately linked to the problem of moral remainders – a problem regarding which, again, the utilitarian has no theoretical room, for the problem assumes that these emotions may be pertinent even if we have done the best thing that we could have done.

As it turns out, while Martin agrees with me that the lessons from Billy Budd are not limited to “tragedies of this magnitude,” she nonetheless thinks that my focus on “subjective emotions,” makes it difficult for me “to draw universal lessons.” I am unsure as to what the disagreement here is supposed to be. Admittedly, part of my motivation in the book comes from a deeper appreciation – via the work of Williams and others – of the limits of theory. This approach may at times have difficulties drawing universal lessons, but this is not because it is obscurantist, or because somehow it flirts with subjectivism or relativism. In fact, some of the central theses I advance in the book strike me as theses with perfectly universal application. Take those views I mentioned before: that punishment, even when justified, is necessarily dilemmatic, that deserved suffering can in itself add value to an organic whole, that classical utilitarianism is much more prevalent in punishment theory than typically recognized and that it is incapable of countenancing the complexity of punishment, for example, are claims that are meant to hold universally. But that some of those views hold universally may not be quite what Martin has in mind when she suspects that my account may have difficulty drawing “universal lessons.”

I think that what Martin has in mind is a combination of, or may be related to, two views. The first view is that I may be somehow stacking the deck when I insist on the conceptual connection between punishment and suffering. Although Martin does not dispute that this “starting point is widely shared,” more than once she distances herself from my way of understanding the relationship between punishment and suffering. Martin in fact refers to my take on this relationship as a “foundational assumption about its [punishment’s] purpose.” On my view, however, the intimate connection between punishment and suffering is neither an assumption nor a purpose of punishment. It is, rather, a purely conceptual point: as a matter of definition, punishment seeks to cause suffering. All sorts of qualifications, of course, are needed: many of those I offered in Punishment and Retribution, and I added a few new ones in Rethinking Punishment. It is important to underscore that the point I advance here is conceptual, for it is in part this very point that explains why punishment is intrinsically dilemmatic. (This is not to say that I must recognize that the disvalue of suffering is invariant, as I explained in the context of my discussion of Dahan-Katz.)

The main locus of disagreement to this view is to be found on Hegelian or neo-Hegelian approaches that see punishment as not only not meant to be a burden on wrongdoers, but as one of their “rights.” I confess that, taken literally, I find the view absurd. The right to be punished is unlike any other right: hardly anyone would ever want to exercise it – and certainly society cannot allow its bearers to waive it at will (as they can in principle do with any other right).[12] Furthermore, if punishment were not an infliction of suffering on the wrongdoer (again: suffering is to be understood very generally, including diminishments of wellbeing, thwarting of interests, infringements of rights, etc.), then it is very hard indeed to see what is so kind, generous, or admirable about forgiveness, or about the merciful person. In an infernal reversal, the (Hegelian) denial of the conceptual point that I endorse would appear to suggest that when someone forgives someone else she is depriving her of one of her rights. Not taken literally, the view surely ceases to seem absurd, but it also ceases to be opposed to the conceptual point I have just defended. In non-absurd versions, the view boils down to saying that, in civilized societies (for the view mainly applies within the context of the state), wrongdoers have a right to be treated with respect, as moral agents, when they are punished. This may be all well and good, but it has nothing to do with an alleged right to be punished, as such.

Some may object to the conceptual relationship between punishment and suffering by pointing out that some criminals actually enjoy prison, and that even when the state knows this (peculiar) fact about these particular defendants, it does not do anything differently in order to ensure that they suffer. Thus, these objectors assume, there is no conceptual connection between suffering and punishment. This objection, which I have discussed at length in Punishment and Retribution, is rather weak. For a variety of compelling principles of political philosophy, states do not engage in Orwellian fine-tunings of the punishments of particular individuals. But this does not deny that state punishments, such as prison, are meant to cause suffering, in general – that is why they are punishments. There are reasons to tolerate the state of affairs in which this or that criminal escapes the state’s efforts to cause her suffering given her eccentric preferences, but if it turned out that prison, in general, were discovered to be not suffering-inducing, then it would simply be abandoned as a form of punishment.

The second view that may explain Martin’s worry about the lack of universalizability of my position concerns my admission that in some instances punishers would not feel nearly as tainted or as emotionally devastated as Vere did after punishing Billy – particularly in light of my insistence that the axiological conflict is always there, and that what varies is merely its emotional intensity. This is similar to De Wijze’s worry about my perhaps cheapening the meaning of guilt and other moral notions, and I would hope that some of what I say in response to him may be pertinent here as well – just as I hope that my comments here may help illuminate my response to De Wijze. Martin takes me to task for admitting that had it been the case that it was Claggart who would have been liable to punishment, Vere would not have felt nearly the devastation that he did feel when it was Billy that he had to punish.

I think that I have a story to tell here, which in spite of its generality may be helpful. In some cases of atrocious wrongdoing – Ohlendorf, Hitler, Iago, Claggart, etc. – the value generated by punishment these characters so dramatically surpasses the value of forgiveness that the emotional consequences of our failing to realize the value of forgiveness can indeed become negligible. It is still true that by punishing these characters we are failing to realize the value of forgiving them – just as it is true that by walking around India feeding the hungry Mother Teresa stepped on many ants, wounding some and killing others. (For symmetry’s sake, assume that it is conceptually impossible to walk around feeding the hungry without thereby wounding and killing some ants.) These are cases in which one of the values in tension is so superior to the other that we would think it absurd for Mother Teresa to be affected, to feel morally tainted, by the ants she has wounded or killed.

If I stand by this story, why do I think that there is something worrisome here? I think that the story may be too ad-hoc. I think that the worry is related to the same sort of reason that, understandably, has led many to be skeptical of the theoretical efficacy of something as unstable as organic wholes. As in the case of organic wholes, where I have tried to show that the instability is neither a matter of intellectual dishonesty nor, in any event, terribly mysterious, I would want to believe that something similar holds here. I have not offered a systematic account as to how to measure the values at play, or the intensity of the moral emotions that punishment generates. I have contented myself with articulating the hitherto unappreciated axiological contours of punishment, even if incipiently, in (very?) general terms. While I would be happy to see more concrete answers to these important questions, I really am apprehensive of what in the book I call “the allure of the simple,” and of the overly mechanistic functioning of the morality system. I have no reason for suspecting that the universal lessons that Martin has in mind necessitate the sort of “obligations-out, obligations-in” system that I wish to avoid, but I am worried that they may point in that direction.

One last point in defense of my position is in order, since I think it underscores that these matters are perhaps not nearly as messy as they may seem. The fact that it may be very difficult to draw a line between borderline cases does not prove that all cases are difficult (think of sorites paradoxes). Moreover, while the exact details of a certain moral taint may be very hard to determine, that some such moral taint should obtain may be relatively easy to determine. Recall my discussion of Benjamin Murmelstein (231 ff.) who participated in innumerable heart-wrenching decisions in Theresienstadt, but who claimed to experience no taint whatsoever – roughly: because he did the best he could. I do not think that it is necessary to be able to determine the exact nature of the compunction (tragic remorse, agent regret, etc.) that Murmelstein should have felt in order to be convinced that he should have felt some moral emotion.

The same holds in the case of punishment: the fact that some cases may prove to be difficult in terms of ascertaining the exact contours of the conflict of values that some inflictions of punishment generate (or of their emotional implications), not all cases are difficult in these regards. As I stress in Rethinking Punishment, many – if not most – of the people we (in the United States, for example) punish via the criminal justice system are not Ohlendorfs, Hitlers, Iagos, or Claggarts. (While the United States may be truly exceptional amongst post-industrial, democratic societies in terms of its anachronistic and harsh criminal justice system, in most countries the majority of those punished by criminal justice systems happen to be the most vulnerable members of society.) I admit that when punishing extraordinarily cruel and vile wrongdoers, the sort of emotional upheavals we may experience are of extraordinarily low intensity. But within the context of many contemporary criminal justice systems there is much to lament in the remarkable suffering we inflict upon the majority of our “criminals,” their relatives and friends, and entire communities.

So, I can almost agree with Martin in that “if Billy wasn’t an angel, but a devil, it is not clear that Vere would (or should) be tainted.” Punishing devils is admittedly supposed to emotionally affect much less intensely than punishing angels. Incidentally, nowhere do I deny that it is possible for people to be feel guiltier than they should, to think of themselves as much more tainted than they should be: pathologies are always possible. And yet, for reasons I presented in my response to De Wijze, I still think that even here there is an axiological conflict between the (admittedly huge) value of inflicting deserved suffering (on a devil), and the (admittedly miniscule) disvalue of making the devil suffer, or of not forgiving him. It is just that the intensity of the conflict is extraordinarily – indeed negligibly – low, as low and negligible as it is in the imagined case of Mother Teresa stepping on ants.

5. Ziemann, Theodicy, and Punishment Zero

Ziemann explores the connection between some of my views in Rethinking Punishment and one provision of the German criminal code. The views of mine on which Ziemann focuses relate to theodicies (which I discuss mostly on chapter one), and the provision of the German code he discusses is section 60, which mandates the non-punishment of convicted criminals in certain cases. The connection is rather obvious, since Ziemann approvingly cites Klaus Günther’s view that the substance of section 60 – (felicitously) labeled in Germany “Punishment Zero” – is none other than the “theodicean idea of poena naturalis.”

Precisely because I find the conceptual connection between punishment and suffering so obvious, I begin Rethinking Punishment by situating the discussion of punishment within the more general context of the discussion of theodicies. Theodicies, after all, seek to explain why there is evil in the world, and as I explain in the book (4 ff.), “evil” in this context can be equated to “suffering.” I find it remarkable that contemporary punishment theorists, concerned as they are with the justification of punishment – the justification of freely bringing more suffering into the world – would not have discussed theodicy more decisively (and Rethinking Punishment seeks to contribute to fill this gap). Theodicies are often theistic, focusing on the (importantly different) question as to why a benevolent God would allow suffering to exist. But, for a variety of reasons, in Rethinking Punishment I focus on secular approaches to theodicies, such as Max Weber’s. Secular theodicies seek to explain how suffering can be valuable, or, as others and I often put it, how it can add meaning to the world. More specifically, I attempt to show that the main worry of proponents of secular theodicies is the existence of gratuitous suffering – deserved suffering has not been perceived by them to be so philosophically problematic. Deserved suffering can add value to organic wholes, gratuitous suffering can rarely do so, if ever. (Throughout this discussion, I approach suffering non-medicinally: I investigate the ways in which suffering as such – and not as conducive to this or that end – may be valuable.)

I transcribe here the text of section 60 of the German Criminal Code, which in the English translation that Ziemann provides is preceded by the heading “Dispensing with Punishment”:

The court shall dispense with punishment when the consequences of the act which have befallen the perpetrator are so serious that the imposition of punishment would be obviously inappropriate. This shall not apply when the perpetrator has incurred imprisonment of more than one year for the act.[13]

These two brief sentences pack quite a lot. It is not that surprising that Ziemann quotes Albin Eser’s claim that section 60 in fact contains “a microcosm of all purposes of punishment.”

It should come as no surprise that I find the spirit that I think animates section 60 commendable. For my purposes here, however, I want to focus on an absolutely crucial word in section 60: “inappropriate.” It strikes me as almost an act of cruelty by the German legislator to have left so unclear what exactly we are supposed to understand by that word – which, adding insult to injury, the German legislator has further qualified with “obviously.” (This is not to deny that other legislators do not force interpreters into these formidable tasks of guesswork). Engaging above all with the work of contemporary German criminal law theorists, Ziemann briefly explores the different ways in which these consequences that may have befallen the perpetrator (without state intervention) can be taken to be “inappropriate.”

While Ziemann believes that “the deeper theoretical foundation” of section 60 remains unsettled, many of the commentators of section 60 that he discusses believe that in the final analysis the section has to be defended along medicinal lines. Roughly: the lesson that, via punishment, the state was supposed to teach the criminal was already taught; or the deterrent effect that the state wanted to produce has already been produced, and so on. Punishment can then be taken to be inappropriate in the sense that the medicinal goal that was supposed to be attained by means of the state inflicting suffering was already attained by the suffering that befell the perpetrator independently of the state.

I of course would prefer to focus on non-medicinal rationales. (To repeat a point I stress in the book: some medicinal consequences are perfectly legitimate, but in order to isolate the central axiological problem of punishment, we are better off ignoring them.) Among the presumably non-medicinal options that Ziemann’s overview of contemporary criminal law theory in Germany affords us, one relates to the idea of poena naturalis which Ziemann associates with Kant’s philosophy. Ziemann mentions Hobbes in this connection, but I am not sure that the passages from Hobbes that Ziemann cites perfectly correspond to Kant’s understanding of poena naturalis. After all Hobbes in these passages seems to include in his view of extralegal suffering much more than the “vice punishes itself” idea that constitutes the core of Kant’s position. (The breadth of Hobbes’s extralegal suffering may be partly hidden by the fact that he groups these varied forms under the single heading of “divine” justice.) The core of the Kantian take on poena naturalis is to be found in ancient Greek philosophy: vice is to be taken to be its own punishment in the same sense that virtue is taken to be its own reward.

I am sympathetic to what I take to be the spirit that animates section 60. I am evidently sympathetic, too, to the general broadening of our horizons when discussing the significance of extralegal suffering – and not only regarding extralegal suffering as a consequence of the commission of a crime, but also independently of it (as evidenced, for example, by my discussion of moral luck (147-176), moral brilliance, and moral imagination (233 ff.)). I am nonetheless surprised by the modest role that what I take to be the most obvious non-medicinal reason for thinking that having undergone some extralegal suffering may render legal suffering inappropriate seems to play in Ziemann’s contribution. I expected a more prominent role to the ways in which extralegal suffering may render any further suffering at the hands of the state excessive in the sense that it would constitute more suffering than the criminal deserves. While I agree that there are important differences between legal and extralegal suffering, I do wish to distance myself from the more radical view whereby the two forms of suffering “differ fundamentally.” In some contexts, suffering is suffering is suffering. As I note in Rethinking Punishment, in the contexts of punishment and theodicies, a certain looseness in our use of the term “suffering” is perfectly typical, and perfectly harmless.

I suspect that something similar to the exaggerated fear concerning the default badness of suffering that I discuss at length in the book, which leads so many generally non-utilitarian authors to (unwittingly) embrace its monistic axiology when it comes to punishment, may lead some of the German commentators that Ziemann discusses to steer clear of anything reminding them (rightly or wrongly) of an “archaic version of retributive punishment.” I can only repeat one of my central theses in the book: we should not exaggerate the risk involved in recognizing that there can be value in giving a wrongdoer the suffering that she deserves (in the sense that this deserved suffering can by itself add value to the organic whole of which it is a part). Even if deserved suffering can in these ways be valuable, it is by no means clear that therefore it should be inflicted: this is a central lesson of a pluralistic (properly or otherwise) justification of punishment.

This consideration brings me to the last two points I would like to make by way of a response to Ziemann. First, Ziemann claims that the provision contained in section 60 “has to be distinguished both from a stay of proceedings by the court and [from] acts of mercy.” I have no problem with the first of these distinctions: the principle at play here strikes me as no mere procedural consideration. But, given the conjunction of my view that within this context mercy and forgiveness can be taken to be synonyms and my view that forgiveness is little more than the deliberate, non-medicinal decision to forego deserved punishment, I must distance myself from Ziemann’s second distinction. Of course, the brief nature of this exchange prevents Ziemann’s from fully articulating what he has in mind when he speaks of an “act of mercy,” so I cannot be sure we disagree on this. For similar reasons, I cannot here articulate a full-bloodied defense of my account of forgiveness – though I do attempt it in Rethinking Punishment.[14]

Still, my distancing here is limited, for there are two relevant scenarios. The first scenario is that the German legislator mandates the forgoing of punishment because any suffering additional to the one the criminal has already experienced extralegally would be undeserved. The second scenario is that the mandate is utterly gracious in the sense that even adding extra-legal suffering to the suffering constitutive of legal punishment, this would not mean that the criminal is getting more suffering than she deserves. I would agree that the first scenario is not forgiveness: there is nothing forgiveness-like in refusing to make people suffer more than they deserve. But the second scenario is a good candidate for forgiveness. I am not sufficiently familiar with section 60 to ascertain which of these two scenarios is supposed to better capture its rationale.

The second and last point regards the elephant in the room. Ziemann concludes his contribution by discussing how Günther believes that punishment is an “act of theodicy in a secular post-metaphysical world,” and indeed as “the last remnant of metaphysics we are carrying in the modern world.” I cannot here undertake a discussion of the nature of our “post-metaphysical” world, or of why being attuned to the axiological discussion of punishment and its merciful remission may be taken to be a remnant of anachronistic metaphysics. But I can point out that whatever turns out to be the ultimate normative justification for section 60, it is very hard to see why it would cease to have appeal when we are in the presence of crimes with penalties longer than one year. It is not just that this specific length of time strikes me as problematically arbitrary (why only one and year and not two, or three years, etc.?), but that any such boundary strikes me as problematically arbitrary. (See, in connection to the problem of arbitrariness, my reply to De Wijze.) Of course, one could imagine considerations having to do with issues of political philosophy or public policy that may recommend – or even require – that section 60 not be applied tout court. Still, those considerations would presumably be medicinal, which, again is not to say anything against their importance other than the fact that they are not very helpful in our efforts to understand the fundamental axiological problem at hand. From the axiological perspective, whatever recommends “punishment zero” in cases of relatively minor crimes may recommend it too in cases of more serious crimes.

* Professor & Chair, Department of Philosophy, Union College.

Bibliography

[1] As in the book, I use here “Benthamite utilitarianism” and “classical utilitarianism” interchangeably.

[2] As noted in the book, I borrow the expression “simpleminded” from Bernard Williams, meaning: “having too few thoughts and feelings to match the world as it really is” (16).

[3] In the book, I explain why in this context “mercy” and “forgiveness” can be used interchangeably (118 ff.).

[4] For a perceptive take which I think shares the spirit of my view about how easy it is for some problems of value to be dismissed by punishment theorists (again: independently of whether they embrace deontological or teleological moral doctrines), see Jeffrie Murphy, “‘In the Penal Colony’ and Why I Am Now Reluctant to Teach Criminal Law,” Criminal Justice Ethics 33.2 (2014): 72-82.

[5] Incidentally, this is precisely the sort of purely axiological disagreement between Jonathan Dancy and G.E. Moore that I flag on 53 ff.

[6] Leo Zaibert, Punishment and Retribution, Aldershot: Ashgate (2006) 10 ff.; Rethinking Punishment 12 ff.

[7] Michael Walzer “Political Action: The Problem of Dirty Hands,” Philosophy and Public Affairs 2.2 (1973): 60-180 at 180.

[8] I have in fact criticized several perceptive criticisms of contemporary punishment institutions, whose authors nonetheless strike me as perhaps too confident that their favored justification of punishment would overcome all problems, and eliminate all moral remainders. See, e.g., my criticisms of David Boonin in “Punishment, Restitution, and the Formidable Method of Directing the Intention,” Criminal Justice Ethics 29.1 (2010) 41-53; of Victor Tadros in my “The Instruments of Abolition, or Why Retributivism is the Only Real Justification of Punishment,” Law and Philosophy 32 (2012): 33-58; and of Hyman Gross in my “Of Normal Human Sympathies and Clear Consciences,” Criminal Law and Philosophy 10 (2016): 91-108.

[9] See Hans Welzel, Das neue Bild des Strafrechtssystems: Eine Einführung in die finale Handlungslehre, Goettingen: O. Schwartz (1961): 24. For an interesting history as to the transformation and translation of Welzel’s idea see Inigo Ortiz de Urbina Gimeno, “Old Wine in New Wineskins? Appraising Professor Bergelson’s Plea for Comparative Criminal Liability,” Pace Law Review 28 (2008): 815-845, at 834. For an interesting discussion of Welzel’s idea itself see Luis Chiesa, “When an Offense is not an Offense: Rethinking the Supreme Court’s Reasonable Doubt Jurisprudence,” Creighton Law Review¸ 44 (2011): 647-704, at 673 ff.

[10] I cannot help noting that, as I write these responses, the current government of the United States is engaged in a barbaric “zero tolerance” immigration policy that amongst other horrible effects separates children from their parents in a haphazard, cruel, and potentially irreversible way. I have seen government officials adduce a remarkably narrow sense of justification (“I am just following the law” and “the law establishes this specific punishment”), which allows the (or at least makes it easy for them) to feel no compunction whatsoever about what they are doing.

[11] See Larry Alexander, “Deontology at the Threshold,” San Diego Law Review 37 (2000): 893-912.

[12] Amongst the many criticisms of the view, see John Deigh, “The Right to be Punished: Some Doubts,” Ethics 94.2 (1984): 191-211. And see also my Punishment and Retribution, op. cit.

[13] https://germanlawarchive.iuscomp.org/?p=752#60

[14] And in my “The Paradox of Forgiveness,” Journal of Moral Philosophy 6.3 (2009): 365-393.

References

Alexander, Larry. “Deontology at the Threshold,” San Diego Law Review 37 (2000): 893-912.

Deigh, John. “The Right to be Punished: Some Doubts,” Ethics 94.2 (1984).

Zaibert, Leo. “The Paradox of Forgiveness,” Journal of Moral Philosophy 6.3 (2009): 365-393.