Or else the man has a family. . . . [I]n killing him you strike the head from the entire family. And once again you punish innocents.
Victor HugoFootnote 1I. INTRODUCTION
▪ In California, someone convicted of three felonies could spend the rest of his days in prison even if two of his felonies are nonviolent.Footnote 2
▪ The Montana penal code authorizes judges to impose a life sentence upon a first conviction for distributing any quantity of marijuana.Footnote 3
▪ In Manhattan, those who commit misdemeanors, such as vandalism and disorderly conduct, are often arrested. Some are prosecuted, convicted, and sentenced. One of the reasons the city gives for pursuing petty criminals so vigorously is that it wishes to send a “zero-tolerance” message, thereby deterring those who would commit more serious crimes.Footnote 4
Aggressive anticrime measures such as these are popular with many politicians and voters,Footnote 5 but the policies also have vocal critics.Footnote 6 Some question whether the policies live up to their reputations for effectively controlling crime.Footnote 7 Many also advance objections from desert. Critics condemn laws such as California's and Montana's for mandating or permitting sentences that exceed what the criminal deserves.Footnote 8 Some critics also consider it morally wrong for Manhattan to treat the prospect of deterring serious criminals as a reason to punish petty offenders. Even if Manhattan's penalties are not excessive in themselves, critics object that the city inflicts them for immoral reasons—reasons other than giving the individual offender what he deserves.Footnote 9
These objections presuppose retributive restrictions—normative principles according to which what a convict deserves on account of his individual conduct and character restricts how states are morally permitted to treat him. Retributive restrictions are often understood as reflecting or extending Kant's famous punishment axioms.Footnote 10 The most popular type of retributive restriction is:
Minimal Retributivism (“Minimalism”) Footnote 11
The state is morally forbidden to impose on anyone a sentence more severe than he deserves for crimes he has committed.
A more constraining type of restriction is:
Exclusive Retributivism (“Exclusivism”) Footnote 12
The only reason for which the state is morally permitted to impose a sentence on someone is in order to give him what he deserves for crimes he has committed.
Exclusivism calls Manhattan's policy into question. However severely petty criminals may deserve to be punished, the fact that they deserve it is not the only reason for which public officials in Manhattan are seeing to it that punishment is imposed. Rather, these officials are also trying to control more serious crime. If exclusivism is true, then they act for unjust reasons even if the petty criminals deserve what they get.
Minimalism does not condemn Manhattan's policy, as exclusivism does, but it speaks to the policies of California and Montana. Minimalism forbids legislators to enact statutes that assign to a certain category of offense a sentence more severe than it deserves. This requires ascertaining the right upper limit for each category.Footnote 13 However, we need not settle on a specific theory of desert in order to see that minimalism has policy implications. On any of a wide range of theories of desert, no one deserves forty years in prison for distributing nine ounces of marijuana. If one of these theories is true, along with minimalism, then Montana's law is unjust. Likewise, on most theories of desert, the man whose third felony is stealing golf clubs does not deserve a life sentence.Footnote 14
Few philosophically informed writers in the past twenty years have seriously questioned retributive restrictions. Minimalism in particular is the most widely endorsed principle in modern punishment theory. Exclusivism is less popular than minimalism but still very popular.Footnote 15 Many punishment theorists today accept exclusivism as well as minimalism.Footnote 16 These restrictions play central roles in the overwhelming majority of punishment theories that have developed in the past several decades. Theories incorporating retributive restrictions include not only those explicitly identified as “retributivist,”Footnote 17 but also hybrid theories,Footnote 18 moral education theories,Footnote 19 societal defense theories,Footnote 20 and others. I refer to those who accept either minimalism or exclusivism (or both) as restrictionists.
Some people find the validity of retributive restrictions intuitively self-evident. They need no argument to persuade them that it is wrong to punish someone in excess of desert or to punish him for reasons other than giving him what he deserves. Others demand arguments for retributive restrictions, which restrictionists try to provide. The main arguments offered involve thought experiments in which the state practices telishment, pursuing its goals by deliberately inflicting severe penalties upon someone who has committed no crime whatsoever.Footnote 21 The earliest arguments of this type were presented as reductiones ad absurdum of act-utilitarianism. They now serve as staples of introductory curricula in moral and legal philosophy. Monographs on punishment theory often begin by rehearsing a telishment reductio.Footnote 22 Such reductiones Footnote 23 are now used to discredit all forms of consequentialism.Footnote 24
My interest in these thought experiments does not concern their use in reductiones of consequentialism but their use in support of retributive restrictions. In order to derive restrictions from telishment scenarios, one needs to engage in moral argument from generalization. I argue that generalization arguments of the same form can be used subversively to derive morally unacceptable conclusions from our judgments about other scenarios in which the state imposes undeserved disadvantages. I proceed from the observation that we approve of policies that inflict collateral damage, such as the ubiquitous policy of excluding the family members of inmates from prison facilities outside visiting hours. I present a generalization argument for the conclusion that these policies are unjust. I argue that if we firmly believe that these policies are not unjust, then we should put less stock in generalization arguments. We should not use them to derive restrictions from telishment scenarios. In fact, I argue, our moral convictions give us at least as much reason to reject restrictions as to accept them. Although aggressive crime-control policies are morally questionable for other reasons,Footnote 25 I conclude that theorists should not rely upon retributive objections.
II. RECENT DEVELOPMENTS IN ETHICAL THEORY
The telishment scenarios have been widely discussed,Footnote 26 so readers might wonder if they merit the extended reexamination that this article undertakes. I think they do, because recent developments in ethical theory shed new light on the scenarios, deepening our understanding of their powers and limitations. Two such developments inform my discussion.
First, in the period since the reductiones gained widespread acceptance, the method of moral generalization has come under scrutiny. Philosophers have recently disputed the assumption that the intrinsic features of actions and policies have fixed moral valences across varying contexts. Suppose the fact that an action constitutes a lie is an intrinsic feature of the action. According to the invariance assumption, if lies are wrong-making pro tanto in some contexts, then lies are wrong-making pro tanto in all contexts.Footnote 27 Philosophers have challenged the invariance assumption.Footnote 28 If it is false, then moral generalizations must proceed with caution. The derivation of retributive restrictions from telishment scenarios constitutes such a generalization, so this move warrants reconsideration.
The second theoretical development that informs my project is the rise of moral theories that combine aspects of consequentialism and nonconsequentialism in new ways. Consequentialism belongs to a larger genus of what I call beneficence theories—theories that posit a strong pro tanto reason to promote the good.Footnote 29 Consequentialists hold that this reason is always overriding, but there are also nonconsequentialist beneficence theories that hold that this reason, although not always overriding, is stronger than anticonsequentialists accept.Footnote 30
The emergence of such theories matters because philosophers have assumed that the telishment scenarios not only discredit consequentialism but simultaneously support retributive restrictions. The possibility of nonconsequentialist beneficence theories calls this assumption into question because such theories need not incorporate retributive restrictions. This article suggests that such theories could perhaps survive all versions of the reductio that have yet been devised.Footnote 31
III. THE CLASSIC TELISHMENT REDUCTIONES
I begin by reviewing the classic reductiones. They purport to demonstrate vividly and decisively why act-utilitarianism is mistaken. In standard form, a reductio posits a society in which public welfare is at grave risk for some reason. The nature of the threat varies from one version of the reductio to another. In some versions, the society has suffered a wave of serious crime and the authorities have been unable to apprehend the perpetrators no matter how hard they try. Unless the authorities can persuade the public that the perpetrators are in custody, their perceived impotence will encourage other potential criminals. In other versions of the reductio, a heinous crime has been committed, and vigilantes are prepared to commit a series of lynchings unless their desire for a scapegoat is satisfied. In either version, the opportunity arises for the authorities to frame, convict, and execute an innocent scapegoat who just happens to be a convenient target.Footnote 32
If the authorities can conceal the scapegoat's innocence and publicize his conviction, as the hypothetical stipulates, then they will deter potential criminals. It is stipulated that the expected benefits to society of deterring crime to that extent would outweigh the expected disutility to the scapegoat. Therefore, act-utilitarians must support telishing the scapegoat. However, all reasonable people, regardless of their theoretical commitments, have strong convictions that such telishment would be wrong, all things considered. Hence, the reductio concludes, we must reject act-utilitarianism.
IV. CONSEQUENTIALISM DECLINES, REDUCTIONES RISE
Until the late twentieth century, utilitarian rationales for punishment were considered so compelling that there was scarcely a market for alternatives.Footnote 33 Today, by contrast, punishment theorists (other than economists) often seem eager to distance themselves from any form of consequentialism, utilitarian or otherwise.Footnote 34 The reductio surely deserves some credit for this dramatic shift in theoretical opinion.
However, chronology suggests that the reductio cannot provide the entire explanation. The reductio first appeared in the early twentieth century, but utilitarian theories of punishment remained popular for many decades thereafter. Antiutilitarians repeated the reductio in print throughout the century,Footnote 35 but a majority of theorists did not embrace it until the last three decades. Why did it take so long for the reductio to prevail, and what accounts for its ultimate acceptance?
I think the answer lies in social trends taking place outside the academic discipline of punishment theory. By the late seventies, members of the general public had become disenchanted with rehabilitation, concluding, rightly or not, that “nothing works” to rehabilitate felons.Footnote 36 In response, they rekindled their retributive impulses, which had never really been extinguished.
Theorists, too, lost some enthusiasm for rehabilitation during this period, although they did not abandon it as thoroughly as did the general public. Theorists were distressed less by the perceived futility of rehabilitation than by ethical concerns about coercive techniques. Many theorists, along with members of the general public, were horrified by what they learned about the use on convicts of psychiatric surgery, psychoactive pharmaceuticals, and aversion therapy.Footnote 37 Many concluded that it is immoral for the state to practice coercive rehabilitation, especially if it involves exceeding desert.Footnote 38
However, most theorists have abandoned more than just coercive rehabilitation. In the seventies, many philosophers began repudiating utilitarianism and all forms of consequentialism.Footnote 39 So did legal scholars (other than practitioners of law-and-economics). Simultaneously, philosophers restored retribution to intellectual respectability as a moral rationale for punishment,Footnote 40 a development that may have further encouraged the rekindling of popular retributive sentiments.Footnote 41 Many theorists joined the general public in their renewed support for retribution and misgivings about rehabilitation. They differ, however, in one fundamental respect. Despite its enthusiasm for retribution, the general public has never abandoned its belief that punishment can and should serve to control crime. Although the public became frustrated with the ineffectiveness of crime control in the seventies, it has never wavered from the idea that this is at least one purpose of the criminal justice system. Having largely forsaken rehabilitation, the public has embraced anew the methods of deterrence and incapacitation, calling for harsher penalties and more effective law enforcement. Voters support these measures as means to both crime control and retribution.Footnote 42 Most theorists outside economics, meanwhile, have abandoned all forms of consequentialism, a trend that has no counterpart in the thinking of ordinary people.
What explains this divergence between theorists and the general public? I can only speculate. Some intellectuals associate utilitarianism with the injustices perpetrated by communist and fascist regimes in the name of the greater good.Footnote 43 Other intellectuals associate it with laissez-faire economics and/or with the political prerogative of majorities against disadvantaged minorities. Many have also been influenced by the provocative work of Michel Foucault. It is no coincidence that Foucault's emblem of the menacing power wielded by the modern penal state is the “panopticon,” designed by the grandfather of utilitarianism, Jeremy Bentham.Footnote 44
Whatever factors may actually explain the rejection of consequentialism, the ironic effect has been that most punishment theorists who consider disproportionate sentences to be unjust now reject consequentialist arguments against such sentences. Some are content to rely instead on their moral intuitions or on Kantian moral theory. However, when pressed for arguments that might persuade someone who does not share their theoretical commitments or their intuition that all disproportionate sentences are unjust, theorists usually offer variations on the reductio. I now turn to these arguments.
V. ANALYZING REDUCTIONES
In this section, I extract what I see as the core of the reductio. Then, in stages, I add layers of material to the core, producing successively more powerful versions. The core is:
Draconia 0
Innocent Ian is arrested, prosecuted, and convicted of a felony in the state of Draconia. The Draconian government sentences him to severe punishment, such as execution or years in prison. Ian does not deserve such severe punishment. He has committed no crime whatsoever.
As I use the terminology, Draconia telishes Footnote 45 Ian: it punishes him in excess of what he deserves. All reasonable people agree that Draconia's actions are objectively unjust.Footnote 46 But we can identify different features of the case as the wrong-making ones. Which of these accounts is correct has implications regarding why and to what extent the state is permitted to punish.
Act-utilitarians, for example, can easily explain what is wrong with Draconia's telishment: it lowers Ian's welfare, and Draconia 0 mentions no positive effects, so telishing Ian fails to maximize long-term welfare.
However, this act-utilitarian reply does not apply to the more interesting:
Draconia 1
The same as Draconia 0, with the specification that telishing Ian maximizes long-term welfare. Perhaps it controls crime to such an extent that Ian's welfare loss is outweighed by the gains of others.
Draconia 1 stipulates that telishment maximizes long-term welfare in this case. Yet reasonable people condemn Draconia 1, according to anti-utilitarians. If this is so, as I am assuming, then Draconia 1 reveals a conflict between act-utilitarianism and our considered moral convictions. Some utilitarians respond that Draconia's actions are not in fact unjust, but retributivists sarcastically refer to this as the “outsmarting” move.Footnote 47
Writers first introduced the reductio in order to refute act-utilitarianism, which was the most familiar form of consequentialism at the time.Footnote 48 So far, the reductio has done all it was originally designed to do. But theorists quickly extended the scope of the reductio to other forms of consequentialism.Footnote 49
A. Retributive Value Theory
Consider, first, combining act-consequentialism with a nonwelfarist value theory.Footnote 50 According to some value theories, it is nonderivatively bad for an individual to suffer undeserved disadvantage. Understand “disadvantage” as anything that reduces welfare, so any of these could constitute disadvantage: death, bodily injury, physical pain, discomfort, emotional distress, frustration of aims, deprivation of resources, etcetera. Disadvantage is bad on all plausible theories of value. But the axiologies I have in mind hold that it is worse, intrinsically and noninstrumentally, when disadvantages are undeserved. Call these retributive value theories. According to these theories, the badness of undeserved disadvantage goes beyond its negative effect on the victim's welfare. A theory that combines act-consequentialism with a retributive value theory will at least sometimes permit agents to choose an action that fails to maximize welfare if it prevents sufficient quantities of undeserved disadvantage. Michael Moore calls this consequentialist retributivism.Footnote 51
How does consequentialist retributivism evaluate Draconia 1? Although telishing Ian maximizes welfare, it may not maximize value, according to retributive value theory. The badness of telishment can outweigh the goodness of the welfare it generates. For any amount of negative welfare, there is a value theory that entails that a single telishment event is still worse. Therefore no matter how much welfare telishing Ian generates, there is a value theory such that consequentialist retributivism condemns it. Therefore there is a version of consequentialist retributivism that coincides with our considered convictions about Draconia 1.
A version of consequentialist retributivism might satisfy some anti-utilitarians if consequentialism itself was never their target and all they really wanted was an alternative to welfarism. Most punishment theorists today, however, have grander ambitions. They are not just antiutilitarians but anticonsequentialists. They can challenge consequentialist retributivism with:
Draconia 2
The same as Draconia 0, with the following specification. If Draconian officials refuse to telish Ian, some other officials (in another state, perhaps) will telish two other innocent victims. Therefore telishing Ian minimizes cases of telishment.
No version of consequentialist retributivism condemns Draconia 2, whereas our considered convictions do condemn it, according to anticonsequentialists.
We can generalize this point. However we define the good, a principle that promotes good states of affairs above all will permit telishment in some possible world. Therefore, for any value theory combined with act-consequentialism, one can fashion a reductio.Footnote 52
B. Rule-Consequentialism
Instead of (or in addition to) proposing nonwelfarist value theories, one could abandon act-consequentialism for a form of indirect consequentialism such as rule-consequentialism.Footnote 53 This is the most popular consequentialist response to the reductiones.Footnote 54
Consider a principle that directs public officials to follow those rules that if widely internalized, promote the good.Footnote 55 Does rule-consequentialism, so understood, condemn Draconia 2? Perhaps. Draconia 2 specifies that telishing Ian minimizes cases of telishment. But rule-consequentialism does not evaluate Draconia 2 by asking whether the act promotes the good. Rather, rule-consequentialism asks what rule would authorize the act in question and whether that rule, when widely internalized, promotes the good. Consider a rule that instructs officials to telish whenever they believe that doing so will promote the good. Could this rule be widely internalized by the officials in Draconia 2, and if so, would it promote the good? In practice it would not. Officials would probably misapply the rule, telishing when doing so did not promote the good. Also, in realistic societies public awareness that officials accepted the rule would generate anxiety, fear, perhaps panic, even if there were no reason to believe that officials had actual plans to telish anyone.Footnote 56 Therefore rule-consequentialism condemns Draconia 2 after all.
In response, anticonsequentialists could introduce:
Draconia 3
The same as Draconia 2, with the specifications that the Draconian officials are extremely virtuous and competent and the Draconian populace is extremely imperturbable. If the Draconian officials internalize rules that authorize them to punish whenever they believe that doing so will promote the good, then doing so really will promote the good in the long run. The officials internalize the rule and telish Ian for that reason.
Rule consequentialism does not condemn Draconia 3. This remains true when rule-consequentialism is conjoined with any reasonable value theory. However, the anticonsequentialists insist, our considered convictions still condemn it. A conflict remains.
C. Generalizing from Reductiones
I am accepting for the sake of argument that the reductiones establish that all versions of consequentialism conflict with our considered convictions. This lesson, while important, is entirely negative. The reductiones teach that consequentialism is false but they do not establish that any general nonconsequentialist principle is true. Restrictionists often move silently from the negative lesson of the reductiones to the positive embrace of retributive restrictions. They do so without giving the argument for that conclusion. They seem to help themselves to generalization arguments. These have the following structure:
1. A policy ψ is morally impermissible;Footnote 57
2. If a policy is impermissible, then any policy that shares all of its morally relevant intrinsic features is also impermissible (“generalizability”);
3. Therefore any policy that possesses all of the morally relevant intrinsic features of ψ is impermissible.
Suppose one hopes to use generalization arguments to support certain general principles on the basis of a judgment about a particular case. In order to do so, one must identify as both intrinsic and morally relevant the very features of the case that one proposes to incorporate into the supported principles. Every additional feature that one treats as morally relevant will further limit the scope of the resulting principles. Consider a generalization argument for minimalism:
1. Draconia's policy of telishing Ian is impermissible;
2. If a policy is impermissible, then any policy that shares all of its morally relevant intrinsic features is also impermissible (“generalizability”);
3. The fact that Ian's sentence exceeds what he deserves for his crimes is the only morally relevant intrinsic feature of Draconia's policy;
4. Therefore if a policy inflicts upon someone a sentence exceeding what he deserves for his crimes, then it is impermissible (i.e., minimalism is true).
The first premise of this argument identifies a single intrinsic feature of Draconia's policy as morally relevant. The principle stated in the conclusion is accordingly very general.
We can also formulate a generalization argument for exclusivism by substituting for the third premise above the premise that Draconia's infliction of a sentence on Ian for reasons other than giving Ian what he deserves for crimes he has committed is the only morally relevant intrinsic feature. So modified, the generalization argument supports exclusivism.Footnote 58
Minimalism reflects the idea that overpunishing the guilty is merely a special case of punishing the innocent. According to minimalism, when a convict who deserves one year in prison as a moral matter receives a two-year sentence, he effectively becomes a morally innocent man who gets unjustly punished during the second year.Footnote 59 Some readers may reject this equation, but I need not address them, as they reject the generalization arguments that I challenge in this article. I accept arguendo the restrictionist claim that overpunishing the guilty is pro tanto impermissible for the same reason as is punishing the innocent.
To avoid misunderstanding, however, the reader should bear in mind the possibility that there are morally important extrinsic differences between punishing the innocent and overpunishing the guilty. For example, policies permitting the former would undermine public support for the justice system much more extensively than do policies permitting the latter. I believe that such extrinsic differences have moral significance. My adversaries in this paper are those who reject the moral significance of such differences. They reason that overpunishing the guilty is unjust because it shares all morally relevant intrinsic features with an uncontroversially unjust practice: telishing the innocent. My project challenges this reasoning.
VI. TU-QUOQUE ARGUMENTS
Generalization arguments are logically valid. The major premise of the argument for minimalism is supported by our considered convictions about the Draconia cases. Using these arguments, however, exposes restrictionists to a certain kind of dialectical reply—a tu quoque.Footnote 60 Sections VI to X exploit this vulnerability. A tu quoque is a reductio of a reductio argument. Instead of denying the intuitive impermissibility of Draconia's policies,Footnote 61 a tu quoque tries to block the derivation of restrictions from these cases via a reductio of generalization arguments. A generic tu-quoque argument has the following form:
1. Policy ψ is impermissible;
2. If a policy is impermissible, then any policy that shares all of its morally relevant intrinsic features is also impermissible (“generalizability”);
3. Policy Φ shares all of the morally relevant intrinsic features of policy ψ;
4. Therefore policy Φ is impermissible.
Here is how to use a tu quoque to block the use of generalization arguments to derive restrictions from the Draconia cases. First, identify a substituend for ψ, such as Draconia's policy. Next, identify a substituend for Φ that shares all of the morally relevant intrinsic features of ψ and is not impermissible, according to our considered convictions. If a substituend for Φ exists that meets these criteria, then we must either conclude that generalization arguments are unreliable or else reject the authority of our considered convictions.
The substituends for Φ that I consider here are policies that cause systemic damage: undeserved disadvantages suffered on account of the imperfect nature of actual penal systems. Systemic damage occurs in every legal system that has ever existed. However, we do not believe that most policies that cause systemic damage are impermissible.
The standard version of the tu quoque in the literature uses the systemic damage suffered by “false positives”—individuals who are mistakenly convicted of crimes.Footnote 62 I prefer not to use this version of the argument, as it is vulnerable to objections concerning the impossibility of identifying the victims ex ante and the unforeseeability of the harm to them.Footnote 63 Instead I use a version that appeals to the systemic damage suffered by friends, close family members, and other members of a felon's community when he serves his sentence. I refer to these third parties collectively as affiliates. Bentham calls damage to affiliates “derivative evil.”Footnote 64 I prefer the term collateral damage.
It would be convenient if every felon happened to be a reclusive, unattached, unemployed, childless orphan, because then the state could punish him without inflicting collateral damage. In reality, however, felons have affiliates. For affiliates, the collateral damage of incarceration is real. It can be psychological, reputational, financial, or a combination. Someone loses his or her partner to prison. Households lose potential breadwinners. Children lose parents. Parents lose children. These losses, whether temporary or permanent, are almost always damaging. Empirical studies find that partners and children in particular often suffer very substantially in all of these ways.Footnote 65 I focus on incarceration in the remainder of this article, although execution obviously inflicts still greater collateral damage in most cases.Footnote 66
I argue that insofar as incarceration inflicts collateral damage in the real world, it shares all of the morally relevant intrinsic features of certain practices that are impermissible according to our moral convictions. This presents a dilemma: either incarceration as practiced in the United States is generally impermissible, all things considered, or (as I believe) generalization arguments are unreliable and cannot be used to support restrictions.
I am not the first to worry that those who use generalization arguments may find it difficult to excuse collateral damage. Both A.C. Ewing and Joel Feinberg express the concern in passing but decline to develop it into a tu-quoque argument.Footnote 67 Few have done so, a failure that calls for explanation. I suggest two explanations. First, to modern people prisons seem as inevitable as hospitals and schools. Prisons as we know them cause extensive collateral damage. This apparent inevitability can make collateral damage seem morally unimportant, akin to death from natural causes. Most people give little thought to the fact that the convict's friends and family members suffer when he serves his sentence. People seem only grudgingly prepared to acknowledge that punishment harms affiliates.Footnote 68
Second, we must consider chronology. Minimalism and exclusivism did not become popular, even with theorists, until the middle of the twentieth century. Before then, only much weaker restrictions had been accepted, as manifested in the expanding implementation of such legal principles as the presumption of innocence in the nineteenth century.Footnote 69 By the end of that century, however, long-term punitive incarceration was well established in Anglo-American legal systems, rationalized by utilitarian arguments. Incarceration rates continued to rise dramatically throughout the twentieth century.Footnote 70 These societal trends coincided with the development of reductiones and generalization arguments. Only then did minimalism and exclusivism become popular with theorists.
Since the 1970s, theorists have used these restrictions to criticize policies of obviously utilitarian pedigree but they have largely declined to consider whether these restrictions are compatible with the most prevalent utilitarian institution of all: the prison. It would not be surprising to discover a fundamental conflict between generalization arguments supporting restrictions and practices that inflict collateral damage, such as punitive incarceration. This is to say that our acceptance of collateral damage can be used to construct tu-quoque arguments. My version of this argument uses ordinary cases of punishment in the modern world, such as:
Nebraska
George commits a felony in the state of Nebraska. Nebraska arrests him, gives him a fair trial, convicts him, and sentences him to twenty years' incarceration under applicable law. George's wife, Martha, suffers severe financial, reputational, and psychological distress as a result, as do George's parents and children.
Although we sympathize with George's family, we do not condemn the state of Nebraska. On the contrary, we understand Nebraska to be doing justice, albeit with regrettable side effects. The fact that we see Nebraska's policy as completely permissible generates the paradox of collateral damage. Substituting Nebraska's policy for Φ in the generic tu-quoque argument yields:
Tu Quoque—Collateral Damage Version
1. Policy ψ is impermissible;
2. If a policy is impermissible, then any policy that shares all of its morally relevant intrinsic features is also impermissible (“generalizability”);
3. Nebraska's policy shares all of the morally relevant intrinsic features of ψ;
4. Therefore Nebraska's policy is impermissible.
VII. RESPONSES TO THE PARADOX OF COLLATERAL DAMAGE
Before proceeding, we also need a substituend for ψ. I begin by substituting Draconia's policy. Later I substitute other policies as needed. There are several possible responses to the collateral-damage argument so formulated. First, a restrictionist could accept it as sound and conclude that Nebraska's policy is in fact impermissible. I return to this response in Subsection XI.B. For now, I simply note that this conclusion would entail radical policy reforms.
Instead one could reject the second premise, generalizability. This is the option that I recommend in Subsection XI.C.Footnote 71 For restrictionists, however, the more appealing responses are those challenging the third premise. Most people find it obvious that Nebraska's policy is permissible and Draconia's, impermissible. Draconia's policy recalls racist miscarriages of justice in American history and human rights violations in oppressive dictatorships. Nebraska's policy, by contrast, is that of a civilized jurisdiction pursuing justice under law. Restrictionists will deny that Nebraska's policy shares all of the morally relevant intrinsic features of Draconia's. I consider seriatim some putative differences between the policies in Sections VIII to X. The first few of these turn out on closer examination not to distinguish the policies at all (Sec. VIII). Next I consider some genuine differences between the policies but find that Draconia's policies differ in these same ways from policies such as California's that are forbidden by retributive restrictions (Sec. IX). Restrictionists cannot consistently appeal to such differences to reconcile our divergent judgments about Draconia and Nebraska.
Finally, I identify some intrinsic differences between Nebraska's policy, on the one hand, and Draconia's and California's, on the other. I conclude, however, that these differences cannot help restrictionists. First, there are reasons to doubt the moral relevance of these differences. Second, even if these differences are morally relevant, I can describe other impermissible policies that share all intrinsic features with Nebraska's policy (Sec. X). I try to locate intrinsic differences between Nebraska's policy and all impermissible policies, but my search fails. My failure suggests, although it does not prove, that there are no such intrinsic differences. If there are none, then the fact that aggressive crime-control policies share the intrinsic features of Draconia's policies, which we condemn, gives us no reason to conclude that the aggressive policies are impermissible. This is because the aggressive policies also differ extrinsically from Draconia's policies in morally relevant ways. I conclude that the Draconia cases provide no independent support for retributive restrictions. Restrictionists have yet to neutralize the paradox of collateral damage.
Before continuing, I must warn the reader against a potential misinterpretation of the remainder of this article. My ultimate position entails that inflicting substantial quantities of collateral damage is justifiable, all things considered, just as common sense suggests. Although I think modern states do not take seriously enough the collateral damage that they inflict in the name of justice, my project is not a crusade to raise awareness of the plight of affiliates as the forgotten victims of the criminal justice system. Rather, I propose to use collateral damage as the basis for a new and more powerful tu quoque. I contend that no one who tolerates as much collateral damage as we do can consistently use telishment hypotheticals to support retributive restrictions.
Tu-quoque arguments are confusing because they are doubly indirect. In Sections VIII to X the reader may find herself thinking, perhaps with exasperation, that there are obviously morally significant differences between Nebraska's policy and the unjust policies to which I compare it. I will often appear to be arguing for the implausible conclusion that Nebraska's policy is impermissible. This is not my claim. I believe there are moral differences between Nebraska's policy and all impermissible policies. I simply think that some of these differences are extrinsic, not intrinsic. The question is not whether one can reconcile the justice of Nebraska's policy with the injustice of policies such as Draconia's. I agree that one can. I offer my own reconciliation in Subsection XI.C. Rather, the question is whether one can consistently extrapolate from the injustice of Draconia's policy to the injustice of all policies that violate retributive restrictions while simultaneously rejecting inferences from the injustice of various other hypothetical policies to the injustice of Nebraska's policy. This, I conclude, one cannot do.
VIII. SHARED FEATURES
In this section I consider several apparent differences between Nebraska's policy and Draconia's. In each case, the contrast proves illusory. The policies actually share the features in question.
A. Affiliates Benefit from Policy
A restrictionist might argue as follows. Martha suffers when Nebraska punishes George, but in other ways she also benefits from Nebraska's policy. First, via general deterrence and incapacitation, the policy reduces her chances of being a crime victim. Second, it increases the extent to which criminals in Nebraska get what they deserve. Suppose that Martha values for its own sake living in a society in which criminals get what they deserve and that her welfare increases when her government implements her values. In that case, Nebraska's policy benefits her in this way as well. Because Nebraska's policy benefits her, one could argue that it does not treat her impermissibly, collateral damage notwithstanding.
The problem, however, is that parallel arguments support the conclusion that Draconia's policy does not treat Ian unjustly either. Draconia's policy controls crime, thereby reducing Ian's chances of being a crime victim. It also increases the extent to which criminals get what they deserve. It is easy to overlook these facts because Ian's telishment obviously inflicts undeserved harm. One must remember that a policy of telishing individuals such as Ian has a general deterrent effect ex hypothesi. It reduces everyone's chances ex ante of being a crime victim. That includes Ian if we consider his ex ante life prospects rather than his ex post prospects as a telishment victim.
Draconia's policy also increases the extent to which criminals get what they deserve. Holding resources constant, the lower the crime rate, the more effectively can Draconia concentrate its resources on apprehending, prosecuting, and punishing the remaining criminals. Insofar as Ian values this for its own sake (as we assume in Martha's case), Draconia's policy benefits him. For these reasons, one could argue that Draconia's policy does not treat Ian unjustly despite the fact that he suffers telishment himself.Footnote 72 So we cannot distinguish between Nebraska and Draconia by comparing the benefits enjoyed by Martha and Ian.
B. Choice to Associate
Instead the restrictionist might observe that it was Martha's choice to marry a future felon. Had she not married George, his punishment would not have caused her emotions, reputation, and economic status to suffer. Perhaps Martha's choice to marry George entails that Nebraska's policy does not treat her unjustly. Ian, by contrast, made no such choice.
This argument fails. The availability of choice at an earlier time does not justify inflicting undeserved disadvantages. It would not be reasonable to expect Ian to avoid telishment by emigrating or hiding from the police, even if he knew in advance that they planned to telish him and he had opportunities to flee. Likewise, it is not reasonable to expect Martha to have declined to marry George, even if, as is unlikely, she could have predicted that her fiancée would commit crimes after marrying her. Nor is it reasonable to expect George's parents to have given him up for adoption or to have chosen not to bring him into the world simply because of the possibility that he would become a criminal. Most obviously, George's offspring did not choose their emotional attachments to him, nor did they, if dependent children, choose their material dependence upon him.
C. Shared Guilt or Benefit from Crime
One might argue that affiliates do in fact deserve to suffer when the convict is punished. Perhaps they benefited from his crimes, or knew about them and declined to report them, or acted as co-conspirators. A bank robber or tax evader may share his “take” with his family members and friends. For these reasons, one might suppose, they deserve to suffer alongside the convict.
This assumption is very common, reinforced by stereotypes of criminals' friends and “business associates.” But most crimes do not benefit the criminal's affiliates at all. Consider assault, homicide, rape, driving under the influence, solicitation, disorderly conduct, and arson. These crimes almost never benefit the criminal's affiliates. Even when affiliates benefit, they do not always know that this is the case, nor is it always reasonable to expect them to know.Footnote 73 Yet affiliates suffer when we punish convicts. Faced with this reality, we apologists for legal punishment must not ease our consciences by dwelling exclusively on stereotypes of “deserving” affiliates.
D. Responsibility of the Criminal
Of course, it is the criminal's fault that his affiliates suffer collateral damage when he is punished. When George commits a felony in Nebraska, he becomes one of the parties who are morally responsible for the resulting collateral damage to Martha. The important question is whether the state of Nebraska is also morally responsible for it. Analogously, one who commits a third felony in California is morally responsible for bringing a life sentence upon himself. However, minimalists insist that the state of California is also morally responsible for inflicting these sentences and is wrong to do so in excess of desert. If so, then Nebraska is also morally responsible for inflicting undeserved harm on Martha. George's desert does not absolve the state of moral responsibility for inflicting harm on Martha in excess of what she deserves.Footnote 74
IX. FEATURES THAT CUT BOTH WAYS
Now I examine some differences that really do make Draconia's policy worse than Nebraska's. However, I note that Draconia's policy also differs in these ways from policies that are forbidden by minimalism or exclusivism, such as California's or Manhattan's. Therefore these distinctions cannot help restrictionists.
A. Severity
Ian endures a severe sentence, as the scapegoat typically does in classic presentations of the reductio.Footnote 75 Collateral damage in the real world seems mild by comparison. Nebraska does not execute or imprison Martha; it merely deprives her of her husband.Footnote 76 Restrictionists could try to distinguish Draconia's policy from Nebraska's by the fact that Ian suffers more severe harm at Draconia's hands than Martha suffers at Nebraska's.
This will not work. The generalization arguments that support restrictions do not identify absolute severity or extreme disproportionality as morally relevant features of Draconia's act. This is to say that retributive restrictions do not just prohibit sentences that are especially disproportionate or especially severe in absolute terms. Minimalism, for instance, prohibits all disproportionate sentences, not just those that are as disproportionate or as severe as Ian's sentence. But Martha's suffering also exceeds what she deserves, so minimalists cannot distinguish between Nebraska and Draconia on the basis of severity.
B. Deception
When Draconia frames Ian, it deceives the public about his guilt. Nebraska deceives no one. One might conclude, therefore, that deception distinguishes Draconia from Nebraska in the required way.Footnote 77 However, the generalization arguments do not identify deception as a wrong-making feature of Draconia's policy. Retributive restrictions do not limit their prohibition to deceptive policies. If they were so limited, then they would not condemn the policies of California, Montana, or Manhattan, as none of these is deceptive either.
C. Means to Justice
Nebraska inflicts collateral damage on Martha in the process of giving George, a guilty convict, what he deserves. Martha's innocence gives the state a pro tanto reason not to punish George, but George's guilt gives the state a contrary pro tanto reason.Footnote 78 Restrictionists could insist that in most cases George's guilt provides a stronger reason in favor of punishing him than Martha's innocence provides against doing so. So the state has reason to punish George, collateral damage notwithstanding. By contrast, telishing Ian gives no one his just deserts. The pro tanto reason not to inflict undeserved disadvantages on Ian prevails unopposed as a reason not to telish him. Therefore, the restrictionist concludes, Draconia acts impermissibly while Nebraska acts permissibly.Footnote 79
This argument cannot serve restrictionist purposes. First, the argument is consistent with exclusivism only if we interpret exclusivism as permitting states to inflict undeserved disadvantages on someone for the purpose of giving someone else what he deserves. So interpreted, exclusivism does not forbid policies such as Manhattan's. Second, we can modify Draconia so that telishing Ian does in fact give someone his just deserts:
Gangster Dad
The same as Draconia 0, with the specification that Ian is the son of a powerful gangster who has fled to a nonextradition country, leaving Ian behind. If Draconia frames Ian, but only then, Ian's father will surrender to the authorities and testify against his subordinates, resulting in many important convictions. Telishing Ian maximizes the apprehension (or conviction, or sentencing) of actual criminals, thereby promoting the good, in this case. Draconia frames and telishes Ian for that reason.Footnote 80
Gangster Dad generates a dilemma for restrictionists who would defend Nebraska's policy in terms of George's guilt. These restrictionists claim that giving the guilty what they deserve is a pro tanto right-making feature with the potential to override the pro tanto wrongness of inflicting undeserved disadvantages. If this is true, then according to generalization arguments, the guilt of the gangster and his subordinates gives Draconia a reason to telish Ian, and that reason is strong enough to outweigh the reason against telishment provided by Ian's innocence. This entails that Draconia's actions are permissible in Gangster Dad, yet our considered convictions condemn them. So this way of distinguishing Nebraska from Draconia itself generates a conflict with our convictions.Footnote 81
X. GENUINE DISTINCTIONS
Finally, in this section I identify some features that distinguish Draconia's policy from Nebraska's but not from other policies, such as California's, that violate restrictions. Invoking these features, however, causes new problems for restrictionists.
A. Confinement
Draconia severely confines Ian's physical movement, whereas Nebraska merely excludes Martha from the prison grounds, leaving her as free as ever to roam the rest of the earth. That, restrictionists could suggest, is why we condemn Draconia but not Nebraska.
B. Stigma
Some writers emphasize the expressive function of punishment.Footnote 82 When the state telishes someone, it expresses the same public message as when it punishes, whereas collateral damage has a very different meaning. A conviction publicly stigmatizes the convict—it denounces him, condemns him, expresses disapproval, ostracizes him, shames him, etcetera.Footnote 83 Restrictionists could insist that Martha suffers few, if any, expressive harms, while Ian suffers many. The expressive component of telishment makes it worse than the equivalent nonexpressive harm (e.g., loss of liberty). Therefore restrictionists might conclude that they can consistently hold that telishment but not collateral damage is unjust.
This reasoning proceeds from a dubious premise. At most, I would agree that affiliates suffer stigma to a lesser degree than does the convict himself. Even this strains credulity. Some affiliates suffer more stigma than does the convict, living in the outside world as they do and having to contend with contempt from neighbors.
But at least the state does not publicly denounce affiliates. What the state “says” about a convict when it incarcerates him is much more condemnatory than what it concurrently “says” about his affiliates. Of course, the state knows that affiliates feel the effects of punishment, but it takes no official position with respect to those effects. The state is not passing any sort of judgment on affiliates. It is not sending them any individualized message.Footnote 84 Nor is the state communicating to other citizens any message about the affiliates. It is not announcing to us that they are shameful or worthy of avoidance or deserving of anything in particular. We might draw such conclusions on our own, but the state does not intend for us to do so.Footnote 85 This distinction marks a genuine, intrinsic contrast between Draconia and Nebraska.
I have responses to the arguments from stigma and confinement, but they are best presented after addressing a more powerful argument to which I now turn.
C. Double Effect
Restrictionists might observe that Draconia intends to harm Ian, whereas Nebraska does not intend to harm Martha; it merely foresees that punishing George will harm her. This, too, looks like an intrinsic difference between the policies, and one that could have moral significance. One version of the doctrine of double effect holds that an action taken with the intention to cause a particular harm is more difficult to justify than an action that is merely foreseen to cause such harm.Footnote 86 Several commentators suggest that restrictionists need double effect in order to answer the false-positive version of the tu quoque mentioned in Section VI.Footnote 87 Despite the rich literature on double effect, no one has explored its usefulness to restrictionists as I now do.
In order to determine whether restrictionists can use double effect to deflect my version of the tu quoque, we must specify Nebraska's reasons for acting. Nebraska foresees that its actions will harm Martha, but does it intend to harm her? I grant that harming Martha is not one of Nebraska's final ends. The question is whether Nebraska harms Martha as a means to an end. Consider, first, Nebraska's incarceration of George. Obviously, its purpose in doing so is to harm George, not Martha.Footnote 88 The harm that Martha suffers from George's incarceration is merely a foreseeable side effect, not a means to an end. This is an intrinsic difference between Draconia's harming of Ian and Nebraska's harming of Martha. If that difference is morally significant, as double effect dictates, then perhaps restrictionists can consistently assert that Draconia's policy is impermissible while Nebraska's policy is permissible.
Restrictionists must remember, however, that double effect is a controversial doctrine in normative ethics.Footnote 89 antirestrictionists could always mount a frontal assault upon it. This article is not the place to undertake such an assault. I would, however, make the following observation. The only point of using a telishment reductio is to persuade readers who are not already committed restrictionists. It makes no sense for a restrictionist to appeal to double effect as a shield against the tu quoque unless those same readers are more committed to double effect than they are to restrictions. Supporters of double effect who were already restrictionists anyway never needed to be confronted with reductiones in the first place. The reductiones were originally designed to persuade all but the most committed consequentialists. A restrictionist who uses double effect has narrowed her audience to certain deontologists.
I need not press this point for two reasons. First, even assuming that double effect is defensible with respect to individual human agents, some have argued that it does not apply to states as moral agents.Footnote 90 If double effect does not apply to states, then restrictionists cannot use it to rebut the tu quoque. I find these arguments persuasive but I do not rehearse or rely upon them. In the rest of this section I assume arguendo that double effect is valid and applies to state action.
D. Coercive Separation of Affiliates
I think I can formulate a version of the tu quoque that survives these concessions. This is possible because in addition to incarcerating George, Nebraska actually threatens Martha herself with physical force. Nebraska even uses force against her if she becomes uncooperative. It prevents her from assisting, rescuing, or protecting George and from visiting him outside legally defined times of day. Picture a prison guard forcibly escorting Martha off the prison grounds at the end of visiting hours. She begs for five more minutes with George, but the guard continues to drag her toward the exit. Surprisingly, much of Martha's distress results not from George's own incarceration but from Nebraska's standing threat of force against her if she tries to interact with George in the proscribed ways. It is Martha, not George, whom the guard subjects to physical coercion.
Policies that forcibly separate spouses from one another demand moral justification. Everyone has a fundamental human right to marry and to remain in close, continuous, physical proximity to his or her spouse.Footnote 91 This right is not absolute but it is extremely strong. When Nebraska excludes Martha from the prison, it infringes her moral rights of free movement, association, and marital companionship. Such infringement has an especially negative impact on Martha's interests. George plays a unique role in her life. There is no adequate substitute for his company. Spending time alone with George may be one of the most important goods in Martha's life. Nebraska forcibly infringes Martha's right to access this good.Footnote 92
It is easy to overlook the fact that states coerce affiliates, not just convicts.Footnote 93 But the fact that Nebraska forcibly separates Martha from George makes possible an extension of the tu quoque. If generalization arguments are valid, then we can derive a principle that counterintuitively condemns Nebraska for its treatment of Martha. Consider the following hypothetical:
Chaperonia
In the state of Chaperonia, married couples have no privacy. State employees supervise spouses at all times. It is a crime for spouses to spend time together without state supervision. Chaperonia enforces this policy because some spouses use their time together to plan and execute criminal conspiracies, which the state aims to discourage.Footnote 94
Chaperonia's policy is impermissible. What are its intrinsic wrong-making features? The simplest possibility is that Chaperonia's policy is wrong because it prevents spouses from spending any (unsupervised) time together, which infringes their rights, when they do not deserve this. If that is why Chaperonia's policy is impermissible, then there is a generalization argument for the conclusion that Nebraska's policy is also impermissible. Nebraska prevents Martha from spending time with her husband. She has done nothing to deserve this.Footnote 95 If we are confident that Chaperonia's policy is impermissible and that Nebraska's policy is permissible, then we must either reject generalization arguments or else find alternative intrinsic wrong-making features of Chaperonia's policy that distinguish it from Nebraska's.Footnote 96
A restrictionist could suggest the following. An action can cause harm as a means to an end or merely as a side effect or aspect of the action. F.M. Kamm defends a principle of permissible harm that permits the latter but forbids the former. Her principle states in part:
[I]t is permissible to cause lesser harm as a means to a greater good just in case the lesser harm is an effect or aspect of the greater good itself that one produces. One may do (greater) good even though (lesser) harm will come of it. One may not . . . do harm in order that good will come of it as a causal effect of the harmful event.Footnote 97
Chaperonia separates spouses as a means to an end. Restrictionists who agree that this is why Chaperonia's policy is wrong could deny that Nebraska separates affiliates from inmates as a means to an end. They could argue that Nebraska does not coerce Martha “in order that good will come of it as a causal effect.” Rather, they could claim, Nebraska coerces Martha only as an effect or aspect of a greater good.
This claim seems plausible at first, but it is mistaken. Real states exclude affiliates from prisons in order to serve various state purposes. This is not intuitively impermissible. Of course, incarcerating George has many effects and aspects that Nebraska does not intend. George cannot water his houseplants while he is imprisoned. The death of his plants is merely a side effect of confining him. But Nebraska's coercion of Martha is not a mere side effect of George's confinement. It is an additional means to Nebraska's ends. Nebraska's incarceration of George does not simply “cause” Martha to be excluded from the prison as it causes George's houseplants to die. Rather, having decided to incarcerate convicts, Nebraska makes an additional policy choice to have its prison guards coerce affiliates. This may be a good choice, but it is a choice.
Similar reasoning applies to the suggestion that excluding Martha is merely an “aspect” of punishing George. The truth of this proposition depends upon what George's punishment encompasses. The main component of his punishment is his incarceration. Martha's exclusion is an aspect of George's punishment but it is not an aspect of his incarceration. Nebraska could limit its threats to inmates themselves while allowing affiliates unlimited visitation privileges. This might be bad policy, but Nebraska could do it while still incarcerating George. Surprisingly, Martha's exclusion is not an aspect of George's incarceration.
So why does Nebraska exclude Martha? One plausible reason is that it causes George to suffer more than if Martha were allowed to visit him at will.Footnote 98 Excluding affiliates makes for more effective punishment in most cases. Suppose Nebraska excludes Martha as a means of increasing George's misery and loneliness. In that case, Nebraska could exclude Martha as a means to the end of depriving George of her companionship, as a means to the end of making him suffer, and this, in turn, as a means to the end of giving him his just deserts. If the only intrinsic wrong-making feature of Chaperonia's policy is that it separates spouses from one another as a means to an end, and if generalization arguments are valid, then Nebraska's policy is wrong, too.
Alternatively, one might suggest that part of George's punishment is living with the knowledge that Martha suffers as a result of his crime. This is a different way of suggesting that Martha's coercion is merely an effect or aspect of the greater good of punishing George. In truth, most convicts do not deserve to have their affiliates suffer as a supplement to their own sentences. But let us accept arguendo that George deserves it.Footnote 99 Even so, George's desert cannot justify coercing Martha. Some premodern states punished family members instead of or in addition to the convict himself.Footnote 100 Consider the following:
Homicidal Dad
The same as Draconia 0, with the specification that Ian is the beloved son of a convict who has murdered several children. In order to approximate more closely the suffering that Ian's father deserves, Draconia telishes Ian (in addition to punishing his father by conventional means).
We condemn Draconia's actions in Homicidal Dad however much Ian's father might deserve to have his own son suffer. We think Draconia is wrong to coerce Ian as a means to the end of punishing his father. Accordingly, the fact that George deserves to have Martha suffer does not justify Nebraska's coercion of her.
Now consider another reason Nebraska might exclude affiliates from prisons: the goal of maintaining prison order. By excluding affiliates, Nebraska is able to maintain better control of its inmate population and reduce criminal activity within its prisons. But if Nebraska excludes affiliates for this reason, then it excludes them, again, as a means to an end. Nebraska keeps Martha out of the prison by force, even if she just wants to visit George and does not plan to do anything illegal, immoral, or disruptive. Nebraska could instead simply prosecute affiliates after the fact if they actually engaged in disruptive or criminal conduct while visiting (e.g., assisting escapes). But maintaining the same level of order under this policy would be much more expensive. Nebraska excludes affiliates because it believes (quite plausibly) that allowing them greater access to inmates would greatly increase the cost of maintaining its desired level of order within its prison system. Because Nebraska cannot identify the individual visitors who pose real threats to prison order, it chooses to exclude everyone.Footnote 101 Therefore there is still a generalization argument against Nebraska's policy. Chaperonia, too, separates spouses from one another in order to control crime. If Chaperonia is wrong to enforce its policy in order to control crime, then, under generalization, so is Nebraska.
For the reasons I give above, restrictionists should admit that Nebraska excludes Martha as a means to an end. They could, however, observe that Martha's psychological distress is not a means to Nebraska's end but merely a side effect or aspect of her exclusion. This is true but irrelevant. The distress of the separated couples in Chaperonia is not a means to an end either. It is merely a side effect or aspect of their separation, which is a means to Chaperonia's end of controlling crime. Chaperonia's policy is still impermissible, all things considered. Under generalization, so is Nebraska's.
Instead, restrictionists might observe that although Nebraska excludes Martha as a means to an end, it does not use her as a means to an end. But again, Chaperonia does not use married couples by separating them; it simply infringes their moral rights. Using someone is not the only way of wronging him.
In fact, I need not even insist that Nebraska excludes Martha as a means to an end, much less that it uses her. I need not insist that Nebraska's policy shares any features with Draconia's. I am not presently arguing that if Draconia's policy is wrong, then Nebraska's is wrong. In this section I am defending a different conditional at a higher level of abstraction: if, as restrictionists claim, the fact that Draconia's policy is wrong entails that California's is wrong, then the fact that Chaperonia's policy is wrong entails that Nebraska's is wrong, too.
The Chaperonia case also illustrates why restrictionists cannot defend Nebraska's treatment of Martha, as I suggest in Subsections X.A and X.B, by emphasizing that Nebraska does not physically confine her or publicly stigmatize her as a criminal. We also condemn policies, such as Chaperonia's, that do not involve stigma or confinement but merely undeserved separation from one's affiliates. Chaperonia severely wrongs married couples by depriving them of marital privacy. It does so without confining anyone or falsely announcing that anyone has committed a crime.
Finally, restrictionists might observe that Nebraska criminalizes all attempts to assist, rescue, protect, or visit inmates or otherwise to interfere with the state's penal operations. Therefore Martha commits independent criminal offenses when she attempts such interference, in response to which positive law authorizes Nebraska to coerce her.Footnote 102 This is true. However, positive law cannot render morally permissible a restriction on liberty if that restriction infringes moral rights. Infringements of moral rights require moral justifications, not legal. That is why Draconia, Chaperonia, and California cannot justify their policies simply by embodying them in positive law. Therefore the moral difference between Nebraska's policy and Draconia's or Chaperonia's cannot be that Nebraska enjoys statutory authority to coerce Martha. The justice of the statute as applied to Martha is in question.Footnote 103
My point in Sections VIII to X has not been that Nebraska's policy is unjust. Nebraska may have no alternative to the familiar affiliate-excluding system of incarceration that delivers equally effective punishment at reasonable cost.Footnote 104 Affiliate exclusion is good policy in that case, but it remains a means to an end. Moreover, if appeals to cost justify coercive affiliate exclusion, then they could also in principle justify policies that violate retributive restrictions, as I suggest in Subsection XI.C.
Sections VIII to X fail to identify adequate bases for distinguishing between Nebraska's policy and all impermissible policies—bases that do not also enable us to distinguish between Draconia's policy and some of the policies prohibited by retributive restrictions, such as aggressive crime-control policies. The paradox of collateral damage remains.
XI. VIABLE SOLUTIONS TO THE PARADOX
For each set of features of Nebraska's policy that I consider above, I identify at least one policy that shares those features and that we condemn as unjust. Of course, I am not considering all the features of Nebraska's policy. It may have some intrinsic features that I have overlooked, such that its complete set of intrinsic features is not shared by any intuitively unjust policy. If Nebraska's policy has such features, then restrictionists can resolve the paradox of collateral damage. Until they identify these features, however, I suggest seeking alternative solutions.
A. Outsmarting
One could, as always, discount our conviction that Draconia's actions are even pro tanto impermissible. This “outsmarting” move denies the major premise of the generalization arguments from Subsection V.D. I neither advocate nor criticize outsmarting here, but it remains a logically acceptable option.
B. Radical Prison Reform
Another solution involves accepting the lesson of Sections VIII to X: for any morally relevant intrinsic features of Nebraska's policy, we can formulate at least one impermissible policy that shares these features. Suppose also that this policy is impermissible in virtue of its intrinsic features and that generalization arguments are valid. That would entail that Nebraska's exclusion of Martha is also impermissible. Embracing this conclusion is a response to the paradox that I postponed from Section VII. This response concludes that the collateral-damage argument is sound—that it is, in fact, impermissible to punish anyone using methods that inflict collateral damage on his affiliates. This conclusion has radical policy implications, some of which I now describe.
In the United States, spouses have a constitutional right to associate with one another. However, federal courts have upheld virtually complete restrictions on the exercise of this right by inmates and their spouses.Footnote 105 As legal justification, courts invoke the state's “valid penological objectives” and the “exigencies and operational considerations of our penal system.”Footnote 106 No state permits spouses and inmates to cohabit. If the collateral-damage argument is sound, however, then justice requires the state either to empty its prisons (putting all convicts under house arrest, perhaps) or else to grant unrestricted visitation privileges to family members and friends. It must allow family members, if they so desire, to cohabit in prison with any remaining inmates. At the very least, the state must offer affiliates reasonable monetary compensation for their losses or permit inmates to work for wages and send the money to their families.
I neither criticize nor endorse these radical proposals. I invite resolute restrictionists to embrace them. I simply note the radicalism of such reforms and the fact that no restrictionist known to me believes that the state is morally obligated to implement them.Footnote 107
C. Abandoning Generalizability
I now turn to a somewhat less radical and more promising solution to the paradox: stop using generalization arguments. Suppose we assign moral relevance to the extrinsic features of cases. We might then rationalize our divergent convictions about the policies of Chaperonia and Nebraska by appealing to extrinsic features that distinguish the cases. Such an appeal might go something like this. When a publicly financed enterprise becomes more expensive while holding revenue constant, the state must reallocate funds from other projects if the enterprise is to continue. The more access to inmates Nebraska allows affiliates to enjoy, the more expensive it becomes for Nebraska to maintain equally effective prisons. One purpose of prison is to subject each inmate to a certain total quantity of unpleasantness as his sentence dictates. But every day someone spends in prison has social costs. Running prisons costs money. Inmates cannot contribute much to the lives of people on the outside nor enjoy what life on the outside has to offer. Because Nebraska separates inmates from affiliates, a week in prison is much more unpleasant than it would be without separation.
Also, if Nebraska allowed affiliates unlimited interaction with inmates, the cost of maintaining prisons per unit time would rise substantially. Many inmates are prospective criminals who would use extended visits from affiliates to undermine prison order. Prisons would have to invest more heavily in maintaining discipline, preventing escape, etcetera. By separating inmates from affiliates, Nebraska can deliver more unpleasantness to inmates at the same social cost. Nebraska can then use the recovered resources to finance a more effective law enforcement system, thereby bringing more deserving criminals to justice and reducing the infliction of undeserved disadvantages on crime victims. Although Nebraska's policy inflicts collateral damage, this is justifiable if it is actually necessary to these goals.
By contrast, Chaperonia's reasons for separating married couples are insufficient under realistic sociological assumptions. Most spouses (outside prisons) are not prospective criminals. Chaperonia inflicts undeserved disadvantages on all married couples for the sake of preventing a relatively small number of undeserved disadvantages that some of those couples would otherwise have inflicted upon others, had Chaperonia allowed them the opportunity to conspire. On these assumptions, Chaperonia's policy increases the total number of undeserved disadvantages suffered in Chaperonia. In this respect, Chaperonia's policy differs extrinsically from Nebraska's on the basis of the total number of undeserved disadvantages suffered if the policy is implemented. If this is a morally relevant feature, then we can draw different moral conclusions about the two policies without denying that they share all intrinsic wrong-making features.
What about Draconia's policy? As many have observed, such policies would not promote the good in realistic, modern societies.Footnote 108 Imagine that a legislature in the United States today granted prosecutors and judges the legal authority to telish whenever they believed beyond a reasonable doubt that doing so would promote the good (however defined). It is extremely unlikely that such a policy would actually promote the good, because virtue, knowledge, and rationality are too limited in the real world. We cannot trust law enforcement to use such powers to promote the good. Some will try to promote the good but will fail due to inadequate knowledge and rationality. Less virtuous officials will use the power to promote their own interests at the expense of the greater good. Even if we could trust state officials with such power, ordinary people would continue to fear otherwise. They would lose faith in government and refuse to cooperate with law enforcement. They might even agitate for revolution or emigrate. Therefore telishing the innocent is impermissible, all things considered, in modern societies.
XII. MORAL PHENOMENOLOGY
Some readers, however, will be dissatisfied with accounts of what is rotten in the state of Draconia that appeal to extrinsic features. Call these contextual accounts. Contextual accounts seem not to reflect the phenomenology of our reactions. Our outrage seems to flow from what Draconia does to Ian—violating his rights, ignoring his innocence, treating him unfairly. In our moral phenomenology, these facts seem sufficient to render Draconia's actions wrong. However, according to contextualism, this appearance is misleading: these facts do not actually suffice to make Draconia's policy wrong. In order to make Draconia's policy wrong, these facts must be “enabled”Footnote 109 by extrinsic factors. Our phenomenology may not reflect this requirement. It may suggest that there is something intrinsically wrong with telishment, whereas contextual accounts seem to deny this, insisting that telishment is only extrinsically wrong.
This reaction to contextualism reflects a subtle misunderstanding. Contextualism is consistent with the claim that some features of the Draconia cases are intrinsically wrong-making pro tanto.Footnote 110 Intentionally inflicting undeserved coercion is one such feature. Contextualism simply denies that this feature is always wrong-making, all things considered. This sets contextualism at odds with retributive restrictions. I suggest that our phenomenology reflects the fact that telishment is intrinsically wrong-making pro tanto, but we mistake features that are intrinsically wrong-making pro tanto for those that are wrong-making, all things considered.
Confusion sets in, I think, because so much happens at once in the Draconia cases. Contextualism holds that Draconia's actions belong to a type that is intrinsically wrong-making pro tanto. Moreover, as everyone agrees, Draconia's policy is wrong, all things considered, at least in realistic contexts. I think this coincidence can mislead even very careful thinkers. It is extremely easy to confuse intrinsic wrongness with wrongness, all things considered. Draconia's policy is a token of a type that is intrinsically wrong-making, but only pro tanto wrong-making. My position entails that Draconia's policy has intrinsic features that are pro tanto wrong-making, even when the policy is permissible, all things considered (i.e., in extremely unrealistic contexts). This distinguishes my position from consequentialist “outsmarting” positions that deny that Draconia's policy has even pro tanto wrong-making features if it promotes the good. If contextualism is correct, then this policy is not intrinsically impermissible, all things considered, but only impermissible, all things considered, in certain contexts. Although it is easy to misperceive it as belonging to a type that is intrinsically impermissible, all things considered, my arguments suggest that it does not belong to such a type.
If my arguments are correct, then some common restrictionist attitudes come into question. Restrictionists believe that Draconia's telishing of Ian is intrinsically worse than Nebraska's depriving Martha of George. I think the moral difference between these policies is an extrinsic matter. I myself have become comfortable with this unorthodox conclusion. The claim that Nebraska's policy is intrinsically less objectionable than Draconia's now strikes me as an unexamined conviction of convenience. Middle-class people tend to have little sympathy for the affiliates of incarcerated felons. We know that separation is unpleasant for the inmate and we want prison to be unpleasant, even at great cost to innocent bystanders. We also believe that our security and comfort depend on prisons and we understand intuitively that equally effective and secure prisons would be much more expensive to maintain if they did not exclude affiliates. So we very much want to believe that there is nothing even pro tanto impermissible about coercive affiliate exclusion. For all of these reasons, we want Nebraska's policy to be intrinsically less objectionable than Draconia's.Footnote 111 But both policies deprive innocent individuals of something irreplaceable and fundamental to their flourishing. Martha's fate may not be as tragic as Ian's, but it is tragic nevertheless. Perhaps we should not feel as uncomfortable about what Nebraska does to Martha as we do about what Draconia does to Ian, but we should feel uncomfortable about both.
Again, my position does not entail that causing collateral damage is impermissible, all things considered (see Subsec. XI.C). Nor does it entail that telishment is permissible in realistic societies. In the real world, telishment is almost always wrong, while inflicting collateral damage is often justifiable. But these are contextual judgments, not to be mistaken for the notion that inflicting collateral damage is intrinsically less wrong than telishing.
We can, however, imagine highly unrealistic societies such as Draconia 3 in which a telishment policy really does promote the good (e.g., by minimizing undeserved disadvantages suffered). Similarly, we can stipulate an unrealistic version of Chaperonia in which the couple-monitoring policy actually promotes the good because so many married couples have criminal plans. Some readers will insist that these policies are still impermissible even in these highly unrealistic societies. I cannot prove them wrong. However, if such a reader believes that the policies of Draconia and Chaperonia are impermissible even if they promote the good, and she also believes that Nebraska's policy is permissible, then she bears the burden of identifying morally relevant features that distinguish Nebraska from the other cases. Until she does so, I think we must choose between (1) our approval of Nebraska's policy in the real world, and (2) the inclination to insist that the policies of Draconia and Chaperonia are impermissible in all possible worlds. I am more committed to the former position than the latter. Others may have different priorities.
XIII. AGGRESSIVE CRIME CONTROL
The main implication of my arguments is that one cannot generalize from our moral convictions about cases such as Draconia's and Chaperonia's even if our convictions are reliable. Restrictionists emphasize that California's “three-strikes” policy shares all morally relevant intrinsic features with Draconia's policy. My arguments entail that even if those features render Draconia's policy impermissible, they may not render California's impermissible. Even if one is certain that telishing Ian in Draconia 3 is impermissible and one cannot identify any morally relevant differences between Draconia's policy and California's policy, one cannot infer that California's policy is wrong. The fact that infringing someone's right to visit his spouse is unjustified in Chaperonia does not entail that such infringements are never justified, as in Nebraska. Likewise, the fact that infringing someone's right against punishment in excess of desert is unjustified in Draconia does not entail that such infringements are never justified. Extrinsic conditions matter. For example, California's policy has not undermined public confidence in the criminal justice system to nearly the same degree as Draconia's would (if implemented in realistic societies). That may be a morally decisive extrinsic difference between the policies.
My arguments also suggest that restrictions that are weak enough to permit widely approved forms of collateral damage will not forbid policies such as California's. Much as Nebraska defends its policy of affiliate exclusion, California could argue that its three-strikes policy conserves resources that it uses to finance more effective prosecutions of deserving criminals. The desert of these other criminals then becomes a justification for the policy. This argument may fail from a false empirical premise—California may not make proper use of the recovered funds. But if the reasoning is invalid, then so is Nebraska's appeal to the social costs of allowing affiliates greater access to inmates. If Nebraska's position is tenable, then the fact that a policy violates retributive restrictions does not suffice to render it impermissible.
Similar conclusions apply to Manhattan's policy. Manhattan punishes some criminals for reasons other than giving them what they deserve, but Nebraska likewise coerces Martha for reasons other than giving her what she deserves. Either both policies are wrong for that reason or neither is wrong for that reason, according to generalization arguments.
Many of us, myself included, are dismayed by the spread of aggressive crime-control policies. They often seem shortsighted and cruel. We are tempted to seize the moral high ground in opposition. My fellow philosophers may not relish the difficult and uncertain work (some of it empirical) that would be required to establish that these policies inflict more undeserved disadvantages than they prevent. They want to condemn them a priori, using retributive restrictions. I conclude with some regret that they are not entitled to that intellectual shortcut.
However, my conclusions do not preclude nonretributivist objections to these policies, such as the following: (1) they are ineffective; (2) they compromise respect for the rule of law; (3) they express vindictive attitudes, which civilized states should disavow; (4) they systematically disadvantage low-income individuals and members of ethnic and racial minorities; (5) they exacerbate ethnic and racial tensions. I endorse none of these objections here. I list them only to illustrate the possibility of other challenges to aggressive crime-control policies—challenges that do not require retributive restrictions.
Although I conclude that retributive objections are philosophically weak, I recognize their effectiveness in political rhetoric. The general public understands the claim that life in prison for stealing cheap videotapes is “too harsh.”Footnote 112 Appeals to this sentiment may in fact be more persuasive to the average person than arguments from ineffectiveness or distributive justice, even if the latter are superior philosophically. If aggressive crime-control policies are objectionable for whatever reasons and retributive arguments help to mobilize political opposition, then I do not object to making such arguments in public whether or not they are sound. I have addressed theorists seeking intellectual clarity, not political actors with predetermined goals.
If my arguments have merit, then how should theorists proceed? First, I suggest that we should not be so quick to reject on principle any theory that justifies punishment as a way of promoting good states of affairs, broadly defined to include those in which people get what they deserve. Nor should we oppose theories just because they tolerate undeserved punishment. Finally, we could develop beneficence theories that diverge from consequentialism just to the extent necessary to survive the reductiones but not so far as to generate conflict with our convictions about the permissibility of collateral damage. Such theories would permit aggressive crime-control policies only if they were truly necessary in order to minimize undeserved disadvantages. Such theories would not as a matter of principle forbid punishment for nondesert reasons. Nor would they forbid the state on principle to punish in excess of desert, although they might permit disproportionate punishment only to the extent that we think the state is permitted to inflict disproportionate collateral damage.
XIV. CONCLUSIONS
Anyone who questions retributive restrictions today invites ridicule by punishment theorists as uninformed, naïve, or worse. The telishment hypotheticals shock the conscience. Retributive restrictions seem to reflect our moral intuitions, but I argue here that they mischaracterize the injustice of telishment. They entail that telishment is intrinsically impermissible, all things considered. I have concluded that telishment is intrinsically wrong-making pro tanto but not intrinsically impermissible, all things considered. Many of us, including mainstream punishment theorists, will resist this qualification. We do not want to believe that the injustice of telishment depends in any way on extrinsic contextual factors. We want its injustice to derive exclusively from its blatant disregard for individual innocence. That disregard becomes the focus of our outrage and anxiety when we contemplate telishment.
However, a successful tu quoque reveals that our deepest moral commitments are more complicated than our outrage indicates. We strongly favor punishing the guilty even when doing so causes substantial collateral damage. We do not really take proportionality as seriously as we seem to take it in the heat of our reaction to cases of unjust telishment.Footnote 113 Retributive restrictions embody an absolutism that we may find ennobling in the abstract but that we do not uphold consistently.
Retributive restrictions are humanitarian pieties with great intuitive appeal. Knowing the evils of which governments are capable, it is comforting to picture these principles shielding us from state power run amok. They strike a reassuringly categorical note. But retributive restrictions may be revealed as dogmas sustained mostly by our collective horror at the evils perpetrated by historical governments in the name of the general welfare. We may discover that these restrictions conflict with the rest of our moral beliefs. The more disquieting this possibility, the more important it is to contemplate it.Footnote 114
