Feinberg tries to isolate punishment from fees and other “mere penalties”. He argues that punishment has a special function that these other sanctions lack. Specifically, it expresses attitudes of resentment and indignation.
Punishment, according to Feinberg, has an expressive function. That is his analysis of what the practice of punishment involves; it is a descriptive claim. To make good on it, he gives us four cases in which punishment accomplishes something that is not explained by either the retributive or the utilitarian justifications for the practice.
I think these four cases fall into two categories. One concerns who punishes, the other concerns how punishment communicates. We spent a fair amount of time talking about these cases. We spent a lot of time on whether the first and second categories were distinct. All of this is interesting, but none is really essential to his argument. He just needs one.
With this analysis in hand, we can do some pretty nifty things. For instance, we can determine whether the law has been correctly applied. Legal punishment has to meet fairly high standards that fees do not. So when the government wishes to defend its actions that harm someone, it sometimes claims that they do not involve punishment in order to avoid having to meet the higher legal standards.
Feinberg claims that, with his analysis in hand, we can tell when at least some of these claims are false. If the government’s actions express resentment and indignation, they constitute punishment.
He also thinks we can explain why strict liability rules are thought to be OK for fines but not for imprisonment. And he thinks he has identified the feature of punishment that calls for justification: the expression of resentment and indignation.
We had trouble seeing the distinction between fines and punishments. Fines just struck us as mild punishments.
I said that I thought most of his objections to retributivism were matters of implementation. They don’t show that the goal of retributivism is mistaken but, at most, that its implementation would be imperfect and difficult. That’s the way it is with just about every public policy goal, so I don’t see that these considerations are especially damaging to retributivism.