Toward a More General Theory
of Punishing and Policing
Today, we deploy the general in order to individualize—inorder to make a better judgment in the case of “any given man.”1In order to devise the right punishment “for each man about to be paroled.”2We place people in categories and assess risk factors in order to know the individual better. The actuarial did not grow out of a desire to disregard the individual. It did not grow out of a preference for the general over the particular. To the contrary, it grew out of our lust to knowthe individual. To know thisindividual. To tailor punishment to the particularities and probabilities of each man. This is, paradoxically, the triumph of the individualization of punishment. Today, we generalize to particularize.
Tragically, though, we are blinded by our lust for knowledge, for prediction, for science. We ignore the complexity of elasticities in order to indulge our desire for technical expertise. We make simplifying assumptions. And we model—incorrectly. What we need to do, instead, is to rigorously analyze the actuarial in much the same way that it analyzes us. And when we do that, we will be surprised to learn that its purported efficiencies are probablycounterproductive—that they may increase rather than decrease the overall rate of targeted crime; that they may produce distortions in our prison populations with a ratchet effect on the profiled groups; and that they may blind us to our own sense of justice and just punishment.
Carol Steiker captures this dilemma well by analogy to “the old chestnut about the man looking for his car keys under a street lamp yards away from his car. When asked why he is looking all the way over there, he replies, 'Because that's where the light is.'”3As Steiker suggests, the appropriate question may be “Why are so many in the world of criminal justice clustering under the street lamp of predicting dangerousness?”