Being Fair to Kant and Bentham
In my last few posts I suggested that classical retributivist theories ran into trouble when it came to explaining the way people think about punishment. I made the old point that the social meaning of punishment varies across social groups, so there is no single unambiguous form of punishment that will seem deontologically just – conceptions of justice will always be contingent on the meaning of a sanction, and that meaning will depend at least in part on values that vary across communities.
Things didn’t look much better when it came to neoclassical deterrence theory (which seems to be on the decline these days; there just aren’t many folks like Posner out there any more).
In this post I admit to being unfair to both classical theories of retributivism and deterrence. Sort of.
First, the unfairness. Let’s start with retributivism. My complaint was that deontological theories (in their classical, unreconstructed forms) don’t describe the real world very accurately. This turns out not to be such a terrible thing as long as you take these theories to be normative rather than positive theories. Kant doesn’t tell you how he thinks the world does work, but rather how he thinks it should.
The neoclassical deterrence theorists are in a bit more trouble as they try to have it both ways. First, neoclassicism tells us, we should assume that people behave as if they are rational. If it turns out they don’t, well then, they should. As Richard Posner put it not all that long ago when responding to emerging research showing that people don’t behave like neoclassical rationalists: “Even if [deviation from neoclassical assumptions about cognition and behavior] has biological roots, it should not be impossible to educate people about it. Behavior therapy has enabled many people to overcome their fear of flying, which I suspect has more tenacious biological roots.”
So again, you have a normative rather than a positive theory. Do I share Posner’s normative vision? Not really, but it’s a free country and he can promote neoclassical rationality all he likes.
Problems arise for neoclassicists, however, whenever there is slippage into positive argument. One way to try to resolve this problem is to endogenize social meaning, but this creates a funny way of talking about things: Humans have complex utility functions that include social meaning satisfactions, and the demand curves for those satisfactions vary across groups. Doesn’t this just mean that the kind of punishments people prefer depend on the values they hold?
So what do we do in the face of evidence indicating that some people favor retribution, some deterrence, some rehabilitation, and that most people care about all this and more in varying degrees? One could de-universalize the theories, take retributivism, combine it with a socially-attentive theory of deterrence, and then mix in some other concerns. But pretty soon it starts to sound like ordinary talk about crime and punishment in which people care about diverse things.
In the legal academy, the people who talk in this strangely straightforward way often talk about “social norms” and “expressive” theories of punishment. Thanks in large part to the work of Jean Hampton, Joel Feinberg, Robert Nozick, Dan Kahan, and others, this approach is gaining ground. But while socially-attentive theories have made substantial headway, deterrence and retributivism talk, as Dan Kahan has noted, still “dominate mainstream theorizing,” and “proponents of these accounts frequently deride the expressive theory or, even more contemptuously, ignore it altogether.”
One reason for this might be that the camps are talking different languages. The classical normative theories starting from different first principles aren’t likely to agree on the terms of debate (that’s a basic function of their starting from different first principles), but at least they recognize that the opposition is proceeding from first principles. Still, in order to engage one another somewhere, they are forced onto the common ground of shared experience, and so their normative arguments about the way the world should work bleed into positive arguments about how people actually think and behave. Perhaps, then, when they encounter expressive theories talking about how people think and behave, they attempt to reverse engineer a normative theory and find it lacking a coherent normative first principle. Perhaps they think: “What a shabby theory!”
Of course, though, that would be evaluating the theory with a mismatched metric.
So what would a theory that grows out of empirical findings rather than normative first principles look like? It would have to account for differences in the values people hold. Look at some of the (relatively rare) studies of attitudes towards punishment, and you’ll see that people care about plenty of things: offender accountability, deterrence, desert, rehabilitation, public costs, secondary social effects, and so on. Moreover, they value these things differently depending on their varied experiences and value-commitments. I suspect that it will come as no surprise, for example, to find that studies show that conservatives tend to give more weight to retribution, progressives to rehabilitation and deterrence. Nor will it surprise you to hear that, interviewing folks from around the country over the last few years, I’ve heard people talk about punishment in very different ways.
Admittedly, though, our empirical measures of this variance have been, to date, fairly rough – mostly polls with a few simple experimental studies thrown in. Thankfully, there are researchers doing the (hard, complicated, and slow) work of teasing apart the factors that shape our punishment preferences. I’ll start describing some of this work in my next post. (And, again, let me solicit references to empirical research into punishment preferences. I’m sure I don’t know all of what is being done out there!)
So there you have it. Was I unfair? Kind of.