“This man has committed treason; his blood is bad; there is traitor's blood in him; that must be wiped out. And in the blood feud the entire clan was wiped out down to the last member. And so, too, will Count Stauffenberg's family be wiped out down to the last member.”
—Heinrich Himmler
A Sikh murders a Hindi, so Hindus burn down the nearest Sikh neighborhood. The consequentialist case against such collective punishments is usually straightforward: Since the murderer is only one Sikh out of millions, his expected punishment is near-zero.
In economic terms, the well-being of the Sikh community is a public good. If “helping the Sikh community” is a flimsy impetus for philanthropy, “harming the Sikh community” is a flimsy deterrent for crime. Indeed, collective punishment is barely more effective than punishing a random scapegoat. “If you commit murder, we will run a national lottery and execute the loser” should scare you about as much as “If you commit a murder, you could be struck by lightning.”
Unless the relevant collective is your family. Family punishments - also known as “kin punishment” - or, in German, Sippenhaft - have been common throughout history. In modern times, totalitarian regimes like Nazi Germany, Soviet Russia, and modern North Korea have successfully used family punishment to put fear into the hearts of their most courageous opponents. Such tactics work because, as evolutionary psychology predicts, even the most selfish human beings normally care intensely about their blood relatives. “If you rob the Mafia, they’ll come after fellow Sikhs” just sounds odd. “If you rob the Mafia, they’ll come after your family” sends chills down your spine.
Now suppose you’re a consequentialist. How do you justify making criminals suffer? The party line is: “Making criminals suffer is, in itself, always bad. The only available moral justification is deterrence: The suffering you save must exceed the suffering you inflict.” My twitter readers even buy this story for Hitler himself:
Now notice: You can easily justify family punishments using the same logic consequentialists routinely use to justify individual punishment: “Sure, jailing a murderer’s son seems unjust. But the real question is whether the suffering you save exceeds the suffering you inflict.” Family punishment is especially promising when the offender is (a) hard to catch, (b) already expects to receive your maximum punishment for earlier crimes, (c) doesn’t mind the standard punishment too much, or (d) cares deeply for his family members. Truly, it’s the ideal deterrent for murder-suicide and terrorism: “You’re not scared of the maximum punishment we can inflict? Think again, miscreant!”
At this point, I can hear legions of consequentialists desperately rationalizing “the fact” that family punishment always fails a cost-benefit test. My one-word reply: “Seriously?!” Yes, family punishments give perverse incentives to criminals who hate their own families. But regular punishments give perverse incentive to criminals who fear life “on the outside.” Yes, family punishments reduce the incentive of criminals’ relatives to avoid crime: “If I’m going to get punished for my brother’s crimes anyway, I may as well become a criminal myself.” But family punishments also amplify relatives’ incentive to police each others’ criminality: “If I’m going to get punished for my brother’s crimes, I’d better make sure he obeys the law.”
What’s clear is that when rulers enforce family punishments, the ruled shiver with terror. Don’t tell me these tactics don’t work.
To be clear, I oppose family punishment. Why? Because I am a retributivist. As the Stanford Encyclopedia of Philosophy explains:
The concept of retributive justice… is best understood as that form of justice committed to the following three principles:
that those who commit certain kinds of wrongful acts, paradigmatically serious crimes, morally deserve to suffer a proportionate punishment;
that it is intrinsically morally good—good without reference to any other goods that might arise—if some legitimate punisher gives them the punishment they deserve; and
that it is morally impermissible intentionally to punish the innocent or to inflict disproportionately large punishments on wrongdoers.
That’s right, I think that inflicting suffering on vicious wrong-doers is morally praiseworthy even in the absence of any deterrent effect. Deterrence is a happy byproduct, but the fundamental point of punishment is to balance the scales of justice. In my Hitler-on-an-island thought experiment, for example, I maintain that the morally correct action is to make Hitler suffer as much as possible. Though you can’t make him die twenty million times, you should try your best to approach that ideal.
Horrifying? Perhaps. But the same principle tells us to leave Hitler’s baby alone, even if the deterrent effect is large. Since the baby is innocent, even a mild punishment is morally forbidden. The same goes for the rest of Hitler’s family. Even if Hitler murdered your whole family, you do wrong if you murder his whole family. (Except, of course, in the bizarre scenario where every single person in Hitler’s family merited death for their own actions).
What if harming Hitler’s baby was the only way to save the world? As a moderate deontologist, I reluctantly endorse this implication. But only in dire hypotheticals with little real-world relevance. The same goes, of course, for letting the heinously guilty off with a slap on the wrist. If you can end World War II a year early by giving Hitler amnesty for his crimes, you should.
Though once you’ve ended the war and got Hitler in chains, the morally best approach is to make him maximally suffer - and claim he died of a heart attack. The reputational harm will be trivial - and his millions of victims deserve no less.
To expand on a standard disclaimer: My position on criminal justice is not endorsed by George Mason University, the GMU Econ Department, or (as far as I know) any of my favorite colleagues. Indeed, I suspect that all of the above strongly reject my position. My challenge for readers: Propose any alternative to retributivism that precludes family punishment. Or even a strong moral presumption against family punishment. I wish you luck in this intellectual quest, because you’ll need it.
Subscribe to Bet On It
Caplan and Candor
I don't see how this is a challenge for consequentialism. If there's a really good case that punishing families is going to lead to a lot better prevention, then maybe it's worth doing - but the kinds of crimes that have been hard to prevent are often ones where the criminal is estranged from their family, or has no family, and punishing the family is going to be extra ineffective in those cases.
If anything, it seems to me that retributivism opens the door much *more* to family punishment. Retributivism raises a question that consequentialism doesn't, namely the one about which entity it is right to take retribution out against. The most common answer in a modern liberal society is going to be the individual whose body is a continuant of the body that committed the crime, but if we ever develop Star Trek style teleportation, or if we are convinced of the possibility of a rehabilitated person really being a new person, then we might change our mind. Historically, retributivists probably actually endorsed the view that the family was the body that is relevant. There's nothing the least bit in tension with the core retributivist idea to say this, and so family punishment should be a core question for retributivism, in a way that it's unlikely to be for consequentialists, who can appeal to empirical facts about how much criminals care about their families to nip this in the bud.
Family punishment is incompatible with individualism, which—at least under modern conditions—is the best social system. Family punishment forces families, rather than individual people, to be the primary units of social agency. Families will decide what is to be done, producing less socially beneficial actions. Your family will be able to veto your plans, just as community groups now veto developers’ plans for the construction of housing.
By the way, the advocate of family punishment must specify how *extended* the “family” supposed to be. A consequentialist would say: the proper size of the unit for responsibility and punishment is that which makes for the best consequences overall, in the long run. And the likely answer is that the best sized unit is *a single person*.
Pursuing still further the direction from individualism to familialism, we find nationalism--the whole nation as "family."
(Question: if a more or less extended family were the chosen unit, what would happen to people without families?)