A thought experiment on retributive punishment and why it's bad
The argument
Years ago I posed this thought experiment, but at the time it was poorly written and filled with unnecessary hoops. I wanted to re-present it in a more polished form. Essentially my argument is as follows:
By retributive punishment, I mean punishment conducted wholly or partly for the purpose of retribution, not incapacitation, deterrence, or rehabilitation.
Even if retributive punishment is acceptable in the abstract, it is unacceptable to create a system that administers retributive punishment if even a single innocent person is caught by it.
All legal systems will catch some innocent people.
Ergo it is always impermissable for retribution to be a factor in sentencing in the legal system.
1 is a definition and 4 seems to follow validly from what came before. 3 seems straightforward, so the question is, how do I justify premise 2? With this thought experiment.
The thought experiment
Suppose you’re the king of the afterlife Your angels have bought you ten thousand very wicked people to be punished with a stint in a suitably nasty purgatory. Make no mistake, these are thoroughly wicked people. They didn’t just back the wrong religion or jaywalk, they’ve done things like defrauding vulnerable old ladies, cold-blooded child murder, torture, war crimes, and so on and so forth. They were not punished for these things in life. Punishment in this afterlife is the last opportunity to correct their wrongs.
The world is cut off from this afterlife, punishing them will serve no deterrent effect. The dead can do no more wrong, punishing them will not affect incapacitation. Heavenly social science says that punishing them is not the best way to rehabilitate them either. No, there’s only one reason to punish them- retribution.
Because you believe in retribution, you are about to hit the button to consign these beings to years and years of suffering, flipping a trapdoor underneath them that will send them to torments below for however long you consider just. However, an angel hurries up to you:
“Sire, due to an administrative kerfuffle, one innocent person is among these guilty people, and we cannot determine which one it is. If you flip the trapdoor, you’ll send the innocent person to a period of horrible torment.”
Would you punish one innocent person to punish 9999 guilty persons -just for the sake of punishment?
I submit that, no, you wouldn’t. If there were utilitarian reasons why the one innocent person needed to be punished, you might accept it as a grim necessity- but for the sake of punishment qua punishment? Goodness no.
This doesn’t just apply to the afterlife
I’ve set this scenario in the afterlife, so as to cut off considerations of deterrence etc., but I believe nothing would change if we set it on earth with the right modifications Mutatis mutandis.
Suppose you’re assigning sentences on mundane earth. You have ten thousand prisoners to be sentenced- all with, miracle of miracles, identical circumstances, and crimes. The considerations of deterrence, rehabilitation and incapacitation jointly indicate a sentence of 3 years would be appropriate. However, considerations of retribution make you lean towards a sentence of four years.
Once again, you receive a bit of evidence telling you that one of the accused- only one mind- is innocent, however, you do not know which.
I submit that, in these circumstances, it would be correct to sentence them all to 3 years, rather than apply retribution to a single individual that doesn’t deserve it. I think (I hope) that most people share my intuition on this.
Now you might say that this is an indication that we should only punish retributively where we’re really sure. It’s true that sometimes we feel sure. Maybe even sometimes we’re rational in feeling sure. But many times people have been convicted and sentenced by juries and judges that felt subjectively sure but turned out to be wrong. Thus, on an institutional level, we’re never really sure that anyone is guilty. Ergo retribution should never be a factor in sentencing.
Admittedly, sometimes when people are talking about retribution they don’t really mean retribution for the sake of retribution, they mean something more like “severity we need to add to stop the general public from rioting at the leniency of the sentence”, a utilitarian argument. It should be clarified that my argument doesn’t really apply to that consideration, although in many cases I would question the wisdom of giving into this tendency on other grounds.
I feel like this depends on whether or not you value retribution as an end in itself - in the abstract we probably don't, we're all good consequentialists here in the abstract real of ideas, but think seriously about whether your "retribution is wrong" attitude would change if one of those people had tortured someone you care about personally.
I have an admittedly cynical view that one of the goals of the criminal justice system is satisfying the desire for vengeance such that the majority of people feels no need to take the law into their own hands, because that is absolutely the kind of thing people do when they don't feel like the legal system is able to do it for them. To placate these people, the legal system probably ends up harsher than it might otherwise be if left purely to the wisdom of the ruling class or even the preferences of the average person. Which isn't to say that we can't have a less punitive criminal justice system, it's just that I can't see a completely non-punitive system working without a significant change to society.
The example doesn't seem to be doing anything more than just restating the premise. Effectively its just assuming "in a situation where a punishment had no non-retributive value but nonzero costs, you wouldn't apply it" which is the exact thing you're trying to prove.
(Perhaps there's also a slight element of haggling over the cost, or perhaps a hint of the dust-specks-vs-torture argument, but both of these would just be distractions from the point you're trying to make)
Here's an alternative version that (at least to me) points in the opposite direction:
You have the magical ability to make people step in dog poo.
Last week somebody stole your bike. You clearly saw the person who did it.
Today you see them again. You're 99.99% certain it's the same person.
Stepping in poo would have no deterrent value - nobody would associate it with the theft. (You don't want to make your ability public because otherwise you'd be subjected to all sorts of uncomfortable scientific testing.) It wouldn't keep them away from stealing other bikes. And there's no way it would rehabilitate them into the sort of person that doesn't steal.
Do they end up with sticky smelly shoes? If so why is the situation different from the one in the post?