Greene and Cohen contend that the law is only superficially compatibilist. The law only requires the capacity for rational behavior as a condition of criminal liability. That is compatible with the causal determination of our actions. But, according to Greene and Cohen, we accept this condition for incompatibilist reasons. The reasoning person, as we think of it, is a non-physical mind, distinct from the physical brain (the view that mind and brain are distinct entities is called dualism).
Neuroscience, according to Greene and Cohen, will undermine this dualist picture of persons by explaining ever more of their behavior as the product of physical causes. As a result, they believe, we will abandon the retributivist parts of our practices of punishment. The retributive ideal is that punishment is reserved for the guilty: those who freely chose to do bad. In place of retributivism, they believe, we will take up a consequentialist approach to antisocial behavior.
Greene and Cohen claim to show that we are tacitly incompatibilists with their example of Mr. Puppet. The idea is that Mr. Puppet would not be held responsible for his actions because they were so clearly formed by the scientist. But, they reason, if we aren’t willing to hold Mr. Puppet responsible for behavior that was caused, we should not hold anyone responsible for their behavior as everyone’s behavior is caused.
I claimed that most people draw a distinction between human and non-human causes. Instead of holding Mr. Puppet responsible, most people would hold the scientist manipulating him responsible instead. There is no one to hold responsible when the causes are more diffuse, as in everyday life. Therefore, the argument goes, the fact that we are willing to excuse Mr. Puppet does not mean that we are committed to excusing everyone whose behavior is caused.
Greene and Cohen claim that we will turn away from retributivism once neuroscience can make extremely accurate predictions about behavior (see Greene and Cohen, 1781).
Aaron noted that there are many physical systems whose behavior cannot be accurately predicted. If I were them, I would say that we will abandon retributivism when neurologically based predictions of human behavior are nearly as good as predictions about the behavior of physical systems, even if those predictions are imperfect. That would strongly suggest that human action is just as subject to causal influence as physical objects are.
James questioned whether we would really move from a retributivist to a consequentialist view of punishment. After all, he noted, it is nearly impossible not to believe in free will. If so, perhaps the great social change Greene and Cohen anticipate will not happen even if the neuroscience comes to be fantastic.