1 Introduction

In his deeply original new work of legal and political philosophy, Gideon Yaffe tackles a surprisingly difficult question: why, exactly, should we punish children who commit crimes more leniently than adults who commit the same offenses? The orthodox answer appeals to kids’ underdeveloped capacities. In the first part of his book, Yaffe dismantles the orthodox view. There are, in fact, countless children whose capacities match those of adults we deem fit for full punishment. Further, those who do suffer from sufficiently diminished capacities—who commit crimes impulsively, or without grasping their consequences—already deserve a mitigation of culpability directly furnished by such facts; no further appeal to their status as children is (or should be) necessary. Yet, Yaffe contends, our judgment persists that it is the status of children as children that grounds the pervasive conviction that they deserve a break when it comes to criminal punishment. What, then, can justify such a view?

Yaffe’s central thesis offers an arresting answer: children deserve a break because they can’t vote. In a nutshell, because children are not complicit in the law’s demands—i.e., because they lack the legal entitlement to co-author its content—the strength of their legal reasons to comply with the law is correspondingly weaker than that of enfranchised adults. And because the extent of a criminal’s culpability is indexed to the strength of the legal reasons that she disregarded when committing crime, and the severity of retributive punishment is in turn properly indexed to criminal culpability, it follows that kids deserve correspondingly less punishment than enfranchised adults who commit the same offenses. These claims constitute the core of Yaffe’s argument.

Yet the argument invites an obvious rejoinder: why not simply enfranchise children, thereby granting them legal reasons that are the same strength as enfranchised adults, and thereby permitting similarly severe punishment? As Yaffe recognizes, it will not do to reject the enfranchisement of children on the grounds that they, as a category, have underdeveloped capacities; this strategy would rely indirectly on the very orthodox rationale that Yaffe discredits at the outset. So instead Yaffe offers an argument as to why it is justified to deny children the vote, one reliant on the twin democratic values of self-government and political equality. He argues that part of what it means to live in a self-governing society is to exert (more precisely, to possess the entitlement to exert) influence over the law into the future. This we rightly accomplish, among other ways, through our entitlement as parents to shape the political values of our children. Yet if we were to give our children the right to vote, it would give parents (in contrast to non-parents) too much of a say over the law, thereby violating our distinct commitment to political equality. So how do we achieve both self-government and political equality? We do it by denying kids the vote—not because they categorically aren’t up to the task, but because we must deny them it in order to co-satisfy the two demands of democratic government.

The sheer number of elaborate moves in this fascinating and richly argued book makes it impossible to deal with the argument as a whole here. Indeed, an implicit methodological achievement of the book is that it makes clear just how interdependent various concerns across political, moral, and legal philosophy, in fact, are. Just this one question—why give kids a break?—demands a theory of legal reasons, a theory of criminal culpability, and a theory of retributive punishment, not to mention a further normative democratic theory that specifies and justifies the central ideals of democracy itself and what they require. This elaborate convoy of premises testifies to the impressive scope of Yaffe’s architectonic. It may also seal its fate.

In what follows, my aim is to identify a series of doubts about the argument, which taken together considerably loosen the grip of its appeal. While the tenor of these remarks is critical, the intention is constructive: to draw attention to the elements of the argument that I think would benefit from greater fortification. My discussion comes in three phases. In Part II, I air an initial worry about the linchpin of the theory: enfranchised citizens who commit crimes are complicit in their own punishment, which explains their increased culpability. I show that this view is implausible outside of the confines of the (unpopular) Rousseauian political theory on which Yaffe relies, and so is likely to be rejected by most theorists of democratic authority today. Then, in Part III, I focus on the argument for why kids should not be enfranchised. I argue that Yaffe’s claim here—that a self-governing society is one whose citizens are entitled to exert an influence over its future—is dubious. Further, I argue that, even if this claim is correct, denying politically competent kids the vote is neither a narrowly tailored nor a fair means of achieving this end. I thus suggest that we should enfranchise politically competent kids, and that we can do this without compromising political equality if we recognize limits on parental authority that are already popular among political liberals.

Finally, in Part IV, I undercut the motivation of the book by questioning whether kids deserve a break after all. Is the claim that they do really the “undeniable normative fact” Yaffe insists it to be (p. 10)Footnote 1? I don’t think it is. So I will suggest that we should both enfranchise (many) kids and punish (many of) them as seriously as we should punish adults. Crucially, this is not to say that we should punish them as harshly as we do punish adults. Instead, we should radically rethink how harshly we punish everybody. More specifically, we should radically rethink why we punish. Pace retributivists like Yaffe, reflecting on why we punish children should make us rethink whether retributivism is the right penal justification for anybody. I close by suggesting that once we reject retributivism, the claim that we shouldn’t grant children a break becomes far more plausible.

2 Democratic Complicity and Criminal Culpability

The linchpin of the book is that “those who have a say over the criminal law are complicit with the punishments inflicted in accordance with it in a deeper way than anyone who does not” (p. 153). Because of this complicity, Yaffe argues, those who have a say acquire weightier legal reasons to comply with those rules than those who lack such a say.

It is important to notice that Yaffe cannot plausibly be using the ordinary notion of complicity, as involving the (culpable) causal contribution to the action (usually the wrongdoing) of another (Gardner 2007; Lepora and Goodin 2013; cf. Kutz 2007). In this causal sense, mere entitlement to influence a law is plainly insufficient to establish that one is complicit with the law. Imagine that there is an active debate among citizens on a proposed criminal statute on which a referendum is held. Can it really be that those who vote against the statute, and campaign against it on account of its alleged injustice, are thereby complicit in their own punishment when—after the statute’s enactment—they violate it and are subsequently convicted? Suppose they are gay citizens, and the statute criminalizes gay sex. Can it really be that they are as complicit as those who voted for the statute and then violated it? Yaffe’s conception of complicity answers affirmatively.

I will not press this worry further here. So let us assume arguendo that there is a further sense of complicity—even if not the ordinary sense—that applies to those with a say over rules, such that they are complicit in those rules simply in virtue of their say over them (even when they vehemently oppose them).Footnote 2 Call this further sort of complicity membership complicity. The normative upshot of membership complicity, according to Yaffe, is that members—those with a say—have stronger reason to comply with those rules than they otherwise would. This is so, Yaffe stresses, even when a member disagrees with the norm he violates: “Thanks to the fact that he had a say over the norm, that is, he is estopped, as it were, from insisting on the norm’s invalidity” (p. 153). But why does having a say over rules mean that one has (or at least could have) greater reason to comply with them?

It will not do, in reply to this question, to stipulate that we have constructed our legal institutions such that it simply is the case that those with a say have stronger legal reason than those without a say. Positivists may grant that we have done this, just as they can grant that the strength of our legal reasons flow contingently from other choices we have made about the statutory, expressive, and institutional dimensions of the law’s specification and enforcement (pp. 138–139). But the crucial normative question concerns why it is morally justified to make it the case that the strength of citizens’ legal reasons is indexed to how much say they have. So let me turn to that argument now.

Yaffe pumps the reader’s intuitions on this with a toy example outside of the legal context, involving a tiddlywinks club, where “[t]hose who opt in pay dues between $0 and $100. The amount they choose to pay determines their degree of say in club policy” (p. 146). We are then to imagine a dues-paying member, X, flouting a club rule, subsequently reprimanded by officials. Yaffe explains:

[W]hen he asks why they think the bylaws have any authority over him, they answer that the bylaws have authority over him because he had a say in their creation. The reply is forceful. It silences a particular objection that X might be raising, an objection that might be voiced by saying “That’s not my rule!” When X has a say over the bylaws, this is simply not true; it is, to some degree, his rule. (p. 147)

The same, Yaffe thinks, can be said to the democratic citizen who asks why the criminal law has authority over him: it has authority over him because he has a say over it.

Crucially, Yaffe concedes that we could imagine a different sort of tiddlywinks club, in which the authority of the bylaws is merely a function of the consent of members who joined the club (p. 147). The point is simply that, for some sets of rules, it is plausible that authority is a function of having a say, but “only if those facts generate a reason for him thanks to the fact that he has a say over them” (p. 147). And so similarly, Yaffe thinks that it is possible to identify “multiple sources of political authority” (p. 148), such that, in some cases, having a say is irrelevant: “My point is only that in many cases, the answer to the question of why a particular fact provides a legal reason for a particular person is that the person had some say over the setup thanks to which such facts generate legal reasons” (p. 148, emphasis added).

But how do we know that our institutions of criminal law—or political authority generally—constitute such a case? Yaffe’s answer involves declaring his allegiance to what he terms a “limited Republican theory of political authority,” according to which, “in the tradition of Rousseau, the state acts authoritatively with respect to a given person thanks to the fact that the person has a say over the law permitting the state’s action with respect to him” (p. 145). Yaffe specifies: “I am operating with the view that some states have authority to do some things to citizens thanks to the fact that citizens have a say over the law authorizing those state behaviors. Further, the criminal punishment of citizens in the United States today, and many other countries, is an example” (p. 146).

In other words, it is precisely because Yaffe assumes a political theory according to which state authority over citizens is a function of their say that he concludes that the law is less authoritative over those without a say. This assumption does an enormous amount of work in the argument. If the assumption is false, the argument does not deliver the desired conclusion. And so it is important to flag that it is a controversial assumption. Of the prevailing theories of democracy’s value and authority on offer in contemporary political philosophy, the Rousseauian thesis that democracy is justified because only democracy is reconcilable with our moral freedom is not a major contender. Indeed, the more general claim that democracy is valuable, and commands legitimate authority over us, simply because it enables us to realize an important form of autonomy and self-government, has been subjected to substantial criticism (Kolodny 2014a, p. 208ff). “So much the worse for contemporary democratic theory,” Yaffe might reply. But it is important to note the costs of such a move: namely, radically narrowing the audience of the argument.

To illustrate why Yaffe’s claim is difficult to accept outside of the Rousseauian view, let me briefly run through three theories of democratic political authority—each more prominent than Rousseau’s in contemporary political philosophy. First, instrumentalists about democracy argue that citizens should have a say over the law, to the extent that they should, simply because doing so secures better (i.e., more just) public policy than alternative arrangements (e.g., better protecting individual rights) (Arneson 2004; Estlund 2008). How do instrumentalists explain why democratic institutions have legitimate authority? After all, merely demonstrating the value of democracy does not suffice to explain why it is authoritative (Kolodny 2014a, p. 197). Here one might appeal to an instrumental conception of practical authority, such as Raz’s (1986), according to which states have authority over us just in case they improve our conformity with the reasons that antecedently apply to us. Or, following Immanuel Kant and John Rawls, one might appeal to the natural duty to comply with just institutions (Rawls 1999, pp. 99–100; Waldron 1993; Stilz 2009). Or one might combine the Razian and natural duty view, as Jonathan Quong does, to argue that institutions have authority just in case they improve our conformity with our duties of justice (Quong 2011). Regardless of the particular path one takes, what’s crucial to notice is this: it is consistent with instrumentalism to vary the amount of entitlement to a say that citizens have, just in case doing so better serves just outcomes. For example, we might think that taking certain decisions out of citizens’ hands—giving them to a Central Bank, or a Supreme Court—thereby improves the predicted justice of outcomes. This is true even for issues that concern the procedure and substance of criminal law, much of which is adjudicated by the Supreme Court. Even J.S. Mill’s suggestion to give educated citizens votes of greater weight is not ex ante morally out of bounds, even if it turns out to be unwarranted (Estlund 2008, p. 215ff). And yet there is no reason to think that those with less say have less reason to comply with the law, just by dint of having less say. Instrumentalists about democracy—an important and significant constituency—cannot therefore accept the linchpin upon which Yaffe’s view rests.

Consider, second, the popular claim that democracy is justified because it constitutes a non-instrumentally valuable response to citizens’ disagreements about public policy—either because it recognizes citizens’ equal moral status (Christiano 2008), or because it constitutes a fair mode of decision-making (Waldron 1999). What this family of views demands is that citizens have an equal say. The legitimate authority of democracy depends on citizens’ possession of this equal say, and it explains why citizens have a (presumptive) obligation to obey the law (and indeed one of similar strength): in light of their reasonable disagreements about public policy, they ought nevertheless to respect each other as equal moral agents. Resolving disagreements democratically is the way to accomplish this.

But what if citizens don’t have an equal say? Yaffe holds that the strength of legal reasons diminishes as the amount of say diminishes (p. 149), since diminishing say means that the law has diminishing authority. But I don’t think this could follow for this sort of view. Here’s why: if citizens are denied equal say, what follows is that the legitimating conditions of democracy are not fully satisfied. In other words, if the legitimate authority of the state inheres in its granting of an equal say to all, the failure to grant that equal say implies that the legitimate authority is compromised. Now, this need not mean that the legitimate authority disappears; it need not be binary. To say that the legitimate authority is compromised means that its normative force diminishes. But why think that it diminishes only for some—those granted less than an equal say—but not for others? Consider a state such as apartheid South Africa. It would be puzzling to think that because white South Africans had considerable say over the law, the law thereby was strongly authoritative for them, yet because black South Africans had no say, the law lacked as much legitimate authority (if any). Rather, the upshot is that the law’s legitimate authority was compromised for all. After all, if our obligation to comply with democracy traces to the way it enables us to respect others as equals, the strength of that obligation attenuates for all to the extent that it fails to do so.Footnote 3

Consider, third, what is perhaps the most fashionable democratic theory of the day, according to which “democracy is a particularly important constituent of a society in which people are related to one another as social equals, as opposed to social inferiors or superiors” (Kolodny 2014b, p. 287). On this view, “relations of social equality are partly constituted by precisely that equal opportunity for influence” over the law that democracy, properly constituted, affords. The justification for the authority of democracy is the moral imperative to relate to others as social equals: “If I were to disregard the democratic decision … I would be, by depriving others of that equal influence, relating to them as a social superior” (Kolodny 2014b, p. 315). The legitimate authority of democratic decisions is a function, then, of the way that they partly constitute relations of social equality (see also Viehoff 2014). But when some citizens are granted less of a say than that to which they are entitled, the political structure no longer helps to constitute relations of equality, and so the basis of the duty to comply with that structure’s commands is undercut—regardless of who one is.Footnote 4

I have argued that Yaffe’s core argument cannot be straightforwardly justified within three of the most prominent theories of democratic authority in contemporary political philosophy. It is unclear why defenders of these theories would demand that legal institutions be designed such that those with a say have weightier legal reasons to comply with criminal statutes than those without a say. One challenge for Yaffe’s view is to explain whether his main argument is compatible with the leading views on democratic authority among scholars today, and if so, how.Footnote 5

3 Children Enfranchisement, Parental Rights, and Self-Government

Suppose we accept Yaffe’s core claim that kids are owed a break because they suffer from “diminished citizenship” (p. 156). But this simply raises the question: should they suffer from diminished citizenship? Or should we instead enfranchise them, such that they have as much say over the law as adults?

Yaffe’s argument for rejecting the enfranchisement of children begins by appealing to value of self-government: “[O]ur commitment to self-government requires that we … adopt some mechanism through which today’s citizens can have a say now over the law in place after they are dead” (p. 176). How should this be done? “This is made possible,” Yaffe suggests, “by giving today’s citizens a say over something that will persist after they are dead, and which will in turn structure the way in which influence is exerted over the law: their children’s values” (p. 176). It is for this reason that he thinks that children must be denied the vote: not everybody has kids, and we don’t want to give those with kids too much say.Footnote 6 Thus, “[o]ur commitment to equality places us under pressure to deny the vote to children while their parents have a say over their behavior and their values” (p. 175).

Notice that Yaffe does not deny that children have bona fide interests in voting—i.e., interests in possessing a legal entitlement to exert influence on the law. Indeed, he goes as far as to say that “a commitment to self-government … places us under pressure to grant the vote to kids” (p. 174). This is paramount. We tend to think that adults with the capacity for full citizenship have a weighty interest in possessing that capacity—so weighty, we think, as to ground a moral right. So Yaffe’s view, then, amounts to the claim that adults’ interests in possessing an entitlement to exert influence into the future is sufficiently great as to outweigh the morally significant interests that children have in voting.

Note also that Yaffe need not think that all kids have an interest in having a say, or at least not a morally significant one. It is likely that only some kids do. To make this idea more specific, we need a notion along the lines of what Claudio Lopez-Guerra calls the franchise capacity, understood as the possession of some minimal set of “intellectual and moral powers” necessary for exercising one’s say. This capacity involves “faculties to understand and value the act of voting—what an election is about, what the options stand for, and so on” (Lopez-Guerra 2014, p. 72). On the basis of his review of the evidence, Lopez-Guerra suggests that “at ten years of age all normal children have the capacity to understand the idea of electing representatives and to adopt a position of their own, however rudimentary, on both the morality of the process and the alternatives at a given contest” (Lopez-Guerra 2014, p. 81). This is all contestable, of course—the average age might be somewhat lower, or higher, or simply vary from individual to individual. The point is simply that there is some significant subset of kids, probably the majority of teenagers, who have the requisite capacities such that they possess a weighty prima facie moral claim to enfranchisement. (I will call such kids competent kids.)

Yet this raises two fundamental concerns. First, we certainly would not be prepared to sacrifice that interest (or, if the interest is weighty enough to ground a moral right, infringe that right) simply for the sake of furthering some other interest, at least not without a compelling justification. Yaffe grants as much, endorsing Hillary Rodham Clinton’s claim that “age categories should be open to scrutiny for some of the same reasons well established suspect classifications are,” such that state action that limited kids’ political participation “should be required to demonstrate a compelling government interest” (p. 181, quoting Rodham 1973, p. 512). Second, we would certainly not be prepared to sacrifice such a weighty interest unless there were no alternative means of achieving our end that did not involve this sacrifice. In other words, the issue is whether the solution that Yaffe defends—denying kids the vote—is narrowly tailored to accomplish its purpose at minimal cost. I will argue that Yaffe fails to attend adequately to either concern.

First, what is the compelling justification for denying kids the vote? Yaffe’s answer is that self-government involves our right to exert influence over the future, after we are dead. I confess that I am not clear on why Yaffe believes this is such a central feature of self-government. He appears to think that it just obviously belongs to the concept, such that any adequate conceptual analysis of the term “self-government” will generate this requirement.Footnote 7 But this is not obvious. Even conceptually, to say that certain individuals govern themselves is just to say that they govern themselves, not that they hold an entitlement to exert influence on completely different individuals who happen to inherit their society. Imagine a society in which we lay eggs, to hatch only after we all die. The possibility of influencing these offspring would be very limited. Would we thereby conclude that we don’t live in a self-governing society, even though we continue to be the authors of the rules that govern us? That would be counter-intuitive.

Science fiction isn’t necessary to make the point. Thomas Jefferson famously suggests as much. Consider Jefferson’s letter to James Madison on September 6, 1789, arguing that constitutions should be continually rewritten by each generation, contending that it is “self evident ‘that the earth belongs in usufruct to the living;’ that the dead have neither powers nor rights over it” (Jefferson 1789). No doubt, there is much to be said on each side of this argument. My point for present purposes is simply that the matter isn’t nearly as obvious as Yaffe believes, and we must hear more before accepting that we have found a compelling justification for denying kids the vote.

Second, even if the self-government argument could be shored up, it is a further question whether the solution Yaffe provides is narrowly tailored to its purpose. Is there some better way to achieve the joint aims of self-government and political equality without denying competent kids the vote? Here it is instructive that Yaffe makes the following point:

An influence over one’s children’s values is not, of course, the only mechanism through which those who today have a full say over the law have a say over future law. Stare decisis, for instance, serves as a mechanism through which anyone who has a say over today’s law also has a say over tomorrow’s. But, still, age thresholds for enfranchisement are one important mechanism. (p. 179)

But if this mechanism comes with an enormous cost—overriding the weighty interests of competent kids—it is natural to wonder: why is it an essential part of the package?

Consider an alternative mechanism for influencing the values of future generations: compulsory education, either provided directly by the state or by private schools required to comply with state-mandated curricula. This is obviously a central way in which we can shape the values of the next generation. Children typically spend as much time interacting with their teachers over the course of their teenage years as they do with their parents. If we were to rely on education, rather than the parent–child relationship, as our primary mechanism for influencing the next generation, this would be consistent with giving competent kids a say over the law.

Relying on education as the primary mechanism would have a further added benefit: fairness to non-parents. If Yaffe is right that our prerogative of self-government entails that we get to exert influence upon the values of our society’s future members, this is presumably a prerogative that we should all enjoy equally. But relying on the parent–child relationship as a central mechanism for influencing the future is unfair to non-parents. Yaffe confronts this worry, contending that “in so far as the childless have real and meaningful opportunities to have children, opportunities that they may have forgone, they have a say over future law” (p. 180). But this is too quick. Bearing children is extraordinarily costly—so costly to those who don’t want children (itself a perfectly legitimate preference) as to constitute a serious obstacle to the exercise of the entitlement to a say. To tell those whose conceptions of the good involves projects incompatible with having kids that they must abandon those projects, and instead have kids if they want to exert influence over future law, is—it seems to me—unreasonably demanding. If we abandon the parent–child relationship as the main mechanism for influencing the future, we avoid this problem.

Thus I am unconvinced that Yaffe’s solution is narrowly tailored: we can achieve the twin demands of self-government and political equality by giving competent kids the vote, and influencing the future through education. What would be so bad about that? Nothing inherent in the ideas of democratic self-government and political equality, I suggest, militates against this option. Yet Yaffe is determined to reject it. But why? I think that latent within Yaffe’s view is a certain implicit commitment to the value and proper function of the parent–child relationship, independent of its involvement in democratic concerns of self-government. He seems to believe that it is simply undeniable that parents should have the sort of relationship with their kids that involves a legal entitlement to control or at least influence the people they become. “It’s just obvious,” he writes, “given some basic facts about the dependency of children on their parents, that we ought to have a system in which parents have a say for at least some period of time over their children’s behavior and values” (p. 175).

But I want to flag that, while this may be “just obvious” to Yaffe, it is not obvious to many other philosophers who have explored this issue. Liberal political theorists have widely condemned the claim that parents enjoy a capacious right to decide what moral values to inculcate in their children. For example, political liberals have argued that, even if parents enjoy some general right to shape the education of their children, this right does not include the inculcation of morally unreasonable political views, such as racist or homophobic doctrines (Quong 2011, pp. 301ff). But even within the domain of reasonable views—e.g., inculcating a mainstream religion broadly consistent with liberal and democratic values—many liberals have resisted the idea that parents may simply install their convictions into their children’s minds. As Joel Feinberg puts it in his canonical defense, children have a “right to an open future” which parents must respect (Feinberg 2007). The role of parents, on this view, is to develop children’s autonomous capacities, including their awareness of different conceptions of the good, so that they can autonomously choose for themselves (see Gutmann 1987; Callan 1997; Levinson 1999). In the most developed recent version of this view, Matthew Clayton argues that we should view parents as officials exercising state-authorized power just like we view police officers and presidents, and so insist that they comport their conduct in accordance with the strictures of Rawlsian “public reason,” which “prohibits adults from enrolling children into particular comprehensive [religious and philosophical] doctrines” (Clayton 2006, p. 112).Footnote 8

We need not settle here the exact scope of parent’ rights to shape their children’s values; no doubt, many will think that Clayton’s view goes too far. The point is simply that any plausible view will recognize some constraints on parents’ legally protected prerogative to shape their children’s values. The salient question, then, is: are these constraints sufficient to allay the worry that parents will have too much say on voting day if their kids are empowered to vote?

I suspect that the answer is yes, that we can identify minimal constraints that are sufficient to allay this worry. Supposing competent kids could vote, it is plausible that parents should abide by the following two constraints. First, they should not instruct their enfranchised kids on who to vote for, nor try in other ways to get them to form specific intentions on what to support in the voting booth. Second, parents are entitled to educate their children to hold general political values—including the general values of liberal democracy—only if they endow their children with the tools and opportunities with which to reflect critically on those values.Footnote 9 These are abstract constraints, and much more would need to be said to render them specific and plausible.Footnote 10 But the rough idea is that if parents widely and faithfully abide by such constraints, non-parents would have significantly weaker reason—probably no reason at all—to complain about political inequality between parents and non-parents.Footnote 11

In sum: provided that parents adhere to moral constraints that are already widely defended in liberal political philosophy, we could enfranchise politically competent kids without thereby compromising political equality. The rationale for denying these kids the vote—and thus, if they commit crime, punishing them leniently—is thereby undercut.

4 Rethinking Retributivism

In the previous section, I tried to show that Yaffe’s arguments for denying competent kids an equal say are not as compelling as they first appear. But, if that’s right, it leads to a potentially unwelcome conclusion: we should enfranchise competent kids, and thereby punish them as harshly as we think enfranchised adults should be punished. This, Yaffe believes, would affront what he takes to be an “undeniable normative fact” (p. 10), that kids deserve a break just by dint of their status as kids. This conviction has the status of a fixed point in Yaffe’s reflective equilibrium. But there is reason to wonder, in light of the aforementioned doubts, whether we should regard the argument’s conclusion as beyond reproach in this way.

So I want to close by imagining the unthinkable: a world in which (many) kids are punished as harshly for their crimes as are adults. Crucially, this world need not be one in which kids are punished as harshly as we in fact currently punish adults (Duff 2002). Rather, the question is: whatever we imagine is the morally fitting response to adults’ wrongful criminal conduct, are we morally disturbed when we then imagine the same response befalling (ex hypothesi) equally culpable children who commit the same offenses? If so, then this is puzzling, precisely because of the fact of equal culpability. So, what, then, could explain our negative intuitive reaction to the idea of children languishing in prison cells, even if we lack that reaction when such a fate befalls equally culpable adults?

To explain what I think is happening here, it is important to recall that Yaffe is explicitly operating with a retributivist view of criminal punishment (p. 119ff). Retributivists believe that punishment is justified because it is deserved. There are many versions of this view, to be sure. But the central insight is that the infliction of suffering on wrongdoers is valuable for its own sake—that is, such punitive suffering has intrinsic value. Many critics have objected that this is both barbaric and mysterious (e.g., Tadros 2011, p. 60ff). How could it be valuable for its own sake when people suffer?

I think the reason we react with horror when we think of children languishing in prison—even those who have committed serious criminal offenses—is that we are capturing this deep moral insight: it is barbaric and senselessly cruel to inflict suffering on children simply for the sake of it. Because we are disposed to be highly sympathetic to children, we grasp this insight clearly in their case. My conjecture is that, in contrast, we have been conditioned to be unsympathetic to the claims of adult criminals, such that it is much more difficult to find ourselves repulsed by the infliction of pointless suffering upon them. This phenomenon—being systematically blinded or insensitive to certain moral claims—is obviously pervasive in human history. My hypothesis is that we are systematically blinded to the moral claims of adult criminals, yet we still, somehow, see more clearly when reflecting on child criminals.

If it is never intrinsically valuable when children suffer—as I think is intuitive—it follows that the retributive punishment of children is never justified. But if there is no difference in criminal culpability between children who commit certain crimes and adults who commit the same offenses, what could justify subjecting adults to retributive punishment? There is, I believe, no such justification.

But this does not mean that we cannot justify punishment. Indeed, once we turn to non-retributivist justifications of criminal punishment, I think the idea that we punish children and adults similarly becomes far easier to stomach. Consider, for a start, R. A. Duff’s influential communicative theory of punishment (Duff 2001). On this view, criminal conviction enables the political community to communicate to the wrongdoer that she has violated an important public norm. The assigned punishment—be it a community service order, or a set of probation requirements, or in rare circumstances incarceration—serves as an opportunity through which the wrongdoer can reflect on what she has done, and, it is hoped, communicate her apology to her victims and the wider political community, and undertake an effort to reform herself. The more serious the wrongdoing is, the more onerous the punishment should be. It is plausible that we could endorse similar sorts of punishments for children and adults under a communicative penal regime. In fact, many conscientious parents who punish their own children may already have in mind something quite like the communicative rationale.

Or consider the related view that a central purpose of criminal justice is to enable the rehabilitation of offenders. Indeed, we already tend to think that rehabilitation is the right penal approach for juveniles. Elsewhere I have defended the thesis that, when agents commit culpable wrongs, they incur a moral duty to identify the source of their malfunctioning moral capacities and remedy it—i.e., to reduce the likelihood that they will act wrongly again (Howard 2017). There is no reason to think that the duty that befalls a 17-year-old murderer to do this should be any less stringent than the duty that befalls an 18-year-old murderer.

Consider, finally, a deontological deterrent model. According to Victor Tadros’s much-discussed theory, wrongdoers are permissibly subjected to punitive harms as a means of deterring others from committing offenses (Tadros 2011, p. 275ff). This is permissible because wrongdoers incur duties to their victims (and possibly others) to remedy the harm they have caused through their crimes by protecting them from future harms—which they can best accomplish by submitting to a regime of general deterrence. On Tadros’s view, the stringency of the duties is indexed to the amount of wrongful harm initially inflicted (Tadros 2011, p. 345). Crucially, one’s age makes no difference to that.

In this closing section I have made two claims. First, the widespread negative reaction that we experience when children are subjected to the same quantum of pointless backward-looking suffering as adults who commit the same offenses should not be interpreted as the idea that kids deserve a break relative to adults. Instead, it should be interpreted as reflecting a deep moral insight: imposing suffering on children is never valuable for its own sake. If that’s right, as I’ve argued, it is likely never intrinsically valuable for anybody. Second, I have argued that, once one reflects on non-retributivist theories of punishment, it becomes wholly palatable to think that children and adults who commit the same offenses with the same mental states should be punished similarly. Kids, in sum, don’t deserve a break qua kids, even if individual kids do for other reasons.