If Free Will Isn't Real, Should We Punish Less?

If Free Will Isn't Real, Should We Punish Less?

Our White Papers & Research section today features a fascinating series of studies about free will and punishment. In several of the studies, when faced with specific crimes, participants recommended far shorter prison sentences when they had been given information that bolsters a "mechanistic understanding" of human behavior.

This result doesn't surprise me. The urge to punish often depends on a belief that the person being punished is blameworthy -- and people often take an explanation of behavior as a justification for it and a reason not to punish it. This is one reason (though not the only one, and certainly not the best one) that evolutionary-psychology theories about criminal violence encounter so much resistance.

But if free will indeed doesn't exist, should we take that as a reason to punish less? Of course, this question itself has to do with the decisions we make, which may not be decisions at all. But that aside, I'm not sure we should. Retribution, deterrence, and incapacitation are all reasons to punish -- and with science explaining more human behavior every day, we need to think through how free will factors in to each one.

The argument against free will goes something like this, at least to my non-philosopher, non-scientist mind. Even with our limited tools today, we can do a much-better-than-chance job of predicting how people will behave and where they will end up in life. People's actions depend on things like their personalities, their genes, their circumstances, the broader societies they live in, and so on. If we had a full dataset of the world -- if we knew everything about every hormone, every neuron, every piece of DNA, every atom in every piece of matter in the universe -- we could predict everything. Even if the universe isn't "predetermined" in the sense that someone actually came up with a plan ahead of time, it's predetermined in the sense that the state of the universe today, coupled with the natural behavior of every particle in the universe, determines the state of the universe tomorrow.

When it comes to retribution, this can matter -- if you're the one with retributivist urges and are capable of stepping back to think about free will. In that case, like the people in the study, you'll lose a reason for wanting harsh punishment. Someone didn't freely choose to cause harm; in fact, things couldn't have gone any other way. However, I suspect that this process is much easier when it comes to empathetic retributivist feelings: It's more likely to play out if a stranger has been harmed than if you or a family member has. When someone hurts you, you want blood, period.

Humans developed retributive instincts for pretty obvious reasons. As Steven Pinker notes in The Better Angels of Our Nature, if you have a reputation for getting even no matter the cost, people will be less likely to mess with you to begin with. But the problem with carrying out vengeance on an individual level is that cycles of violence get started; the person who harmed you (and/or his family) doesn't think he did anything wrong, or doesn't think your counterstrike was proportionate, and goes after you again. So, in modern societies, the government steps in and gives out a fair, final punishment, nipping these cycles in the bud.

In other words, the retributive aspect of punishment isn't just society holding someone blameworthy -- it's society exacting retribution so that an aggrieved individual doesn't feel compelled to. Whether the punishment satisfies (or at least mollifies) the aggrieved is important, and it doesn't necessarily matter if the criminal behavior was predetermined if the urge to punish the criminal behavior was predetermined too. Perhaps we should give less weight to retribution than we do today, but even if free will doesn't exist we shouldn't ignore retribution entirely.

What about deterrence? I'm not sure free will really factors in here when it comes to our sentencing decisions (such as they are). Science shows us that behavior and consequence tend to go together -- if a lever dispenses a pellet, the rat becomes more likely to push the lever. Sure, the entire experiment may be predetermined, but until we bring ourself into the Minority Report age, that's of little use to us. We have to rely on the broad patterns that connect punishment to deterrence.

Finally, incapacitation. Here, a mechanistic view of human nature might even strengthen the case for punishment. Say Josh and Roger are hanging out together when Josh insults Roger's wife. Roger breaks a beer bottle, stabs Josh in the neck, and leaves him to bleed out on the floor. With a full dataset of the world, we could have predicted this. But what information in that dataset would have been most informative?

Sure, the situation mattered: If Josh hadn't insulted Roger's wife, Roger probably wouldn't have stabbed Josh in the neck. But also important was who Roger was. If Josh had insulted a different man's wife, that man might have done nothing, settled for a verbal altercation or fistfight, or just left. If Roger is simply the type of person who responds to insults with a broken beer bottle to the neck -- and, importantly, if that trait is stable, a separate question bearing on the possibility of rehabilitation -- the case for incapacitation is stronger than it would be if Roger had freely chosen that course of action among many equally plausible alternatives.

If Roger has free will, we can be optimistic about getting him to make different choices in the future; if human behavior is mechanistic and the machine that is Roger tends toward violence, our best option might be to just lock him up.

I wonder if an experiment presenting arguments along these lines, rather than focusing on retribution, would find more support for punishment among people who'd been told free will doesn't exist. When it comes to punishment, determinism can cut both ways.

Robert VerBruggen is editor of RealClearPolicy. Twitter: @RAVerBruggen

Show commentsHide Comments

Related Articles