Claire Finkelstein's primary conclusion is that "retributivism fails to justify the use of death as punishment." (1) By "justify," Finkelstein seems to mean no more than "show to be morally permitted." I shall "justify" the death penalty in a somewhat stronger sense. I shall show it to be not only sometimes morally permitted but, all else equal, something that reason sometimes recommends.
Finkelstein seems to understand her conclusion as a universal truth, true everywhere and always. She reaches it by refuting a number of common retributivist arguments for the moral permissibility of the death penalty and drawing from those individual refutations a provocative secondary conclusion: "a desert-based theory of punishment is particularly ill-suited to [justifying the use of the death penalty]." (2) If her list of retributive arguments is complete (and her refutations sound), she is entitled to those two conclusions. If she has missed even one sound retributive argument, she is not. I believe that she has missed at least one. It is that argument that I present here. If it is sound, the death penalty can be justified retributively and both her conclusions are false.
I Two Kinds of Retribution
Finkelstein begins her discussion of retributivism with an oversight that may explain how she missed the argument I am about to make. While recognizing that the literal form of lex talionis--the criminal deserves in punishment the same harm that the crime caused--is not a popular position even among staunch retributivists, she still seems to suppose that all forms of retributivism must proportion punishment to the harm done (at least for the major intentional crimes). The retributivist "would seek to inflict on the perpetrator by way of punishment the nearest morally permissible form of punishment to the act the perpetrator committed." This is the "moral equivalence" ("nearest") form of lex talionis. (3)
All forms of the lex talionis use the harm the crime does as a major component when determining equivalence. All forms of the lex talionis, therefore, have trouble with the enormous diversity of crimes characteristic of any sophisticated legal system. Too many crimes do no harm--in any useful sense of "harm." Any plausible retributivism must be able to assign penalties to many "harmless" crimes, everything from attempted murder to reckless driving, from failure to place a tax stamp on a liquor bottle to conspiracy to commit embezzlement. For that reason, a number of retributivists have tried to understand punishment as taking back (or canceling) the unfair advantage the criminal gets from the crime as such (or, at least, the value of that unfair advantage) rather than as returning something like the harm he did. Here is an example of the method by which that sort of retributivism would assign penalties to crimes:
1 Prepare a list of penalties consisting of those evils (a) which no rational person would risk except for some substantial benefit and (b) which may be inflicted through the (relatively just) procedures of the criminal law.
2 Strike from the list all inhumane penalties.
3 Type the remaining penalties (by evil imposed), rank them within each type (by amount of that evil), and then combine rankings into a scale.
4 List all crimes.
5 Type the crimes (by interest protected), rank them within each type (by degree of protection), and then combine rankings into a scale.
6 Connect the greatest penalty with the greatest crime, the least penalty with the least crime, and the rest accordingly.
7 Thereafter: type and grade new (humane) penalties as in step 3 and new crimes as in step 5, and then proceed as above.
The harm a crime actually does plays no part in this assignment of (maximum) penalties. The "harm" in prospect, if the evil in question is "harm" in any interesting sense, does have a part, but one constrained by considerations of type and grade, the architectonic of a system of criminal justice. …