Retributivism varies in its details but typically holds that we are justified in making wrongdoers suffer (or be punished) in proportion to their wrongdoing.
There is a lot being crammed into the "suffer (or be punished)" phrase here. It is not in fact a retributivist principle that we are justified in making wrongdoers suffer in proportion to their wrongdoing; the principle is that we are not justified in giving wrongdoers a punishment beyond what is proportionate to their wrongdoing. These are not remotely the same things, and the former is not even remotely reasonable -- for one thing, it is impossible to control how people suffer in general with this precision (the penalized can themselves affect how much they suffer, for instance, and there can be completely unintended causes of suffering). When retributivists talk about proportionality, they are talking about punishments, not suffering; this is true even when they think of punishment as a particular kind of suffering. Kobler tries to deal with this in his main argument, but the argument itself is problematic:
Some retributivists deny that they must justify the side effects of punishment because those side effects are not themselves punishment. Rather than get wrapped up in debates about the meaning of punishment, I find it more helpful to speak of the justification of punishment practices. Retributivists who purport to only address the intentional inflictions of punishment will have little to say about real-world punishment practices as they invariably include side- effect harms.
If there are any retributivists who are not making this denial, they are conceding far more than retributivism itself requires. By the very act of saying that he is not going to get wrapped up in debates about the meaning of punishment, Kobler has already begged the question against the retributivist; it follows directly from retributivism that you cannot assess these matters at all without determining exactly what the punishment is. And the last sentence is simply false -- from the fact that retributivists will say some aspects of punishment practices are purely incidental to punishment itself, one cannot conclude that they say that all or most aspects of punishment practices are incidental, which is the only way they would 'have little to say about real-world punishment practices'. Very obviously, punishment plays quite a large role in 'punishment practices' (this is why they are called 'punishment practices'), and retributivism certainly has something to say about that. As for the rest, any side-effect harms, they may well be worth taking into account, but the retributivist is going to say that these, precisely because they are not part of the punishment itself, should be considered not under the theory of punishment but under some other part of the theory of justice, and how they relate to the issues of punishment will be handled by a more general theory of justice. There's no reason why retributivists have to include everything under punishment just because subjectivists want to do so. It's entirely OK to say that our punishment practices touch on other aspects of justice beyond punishment alone, and that our theory of punishment should confine itself to the actual punishment in our punishment practices. Precisely one of the reasons why retributivism spontaneously gets defenders every generation is that it is a fairly modest and simple theory of punishment; it does not demand a lot from punishment itself, and it does not overload the theory of punishment with questions beyond those directly concerned with penalty and desert.
Kobler also seems to think (and many critics of retributivism do likewise) that retributivists are committed to saying that every wrongdoing must be punished as it deserves. No actual retributivism holds this. Every retributivist I have ever come across has recognized that (1) there are wrongdoings we just practically cannot punish, either because we never know enough about them, or because we cannot establish the relevant desert, or because we don't have the means to punish them; (2) there are cases where a wrongdoing deserves to be punished but either practical necessity or higher good (like that of the peace of the entire community) justifies pardon or commutation or just letting people off with a warning or even at times ignoring it altogether. These don't fall under the theory of punishment but under more fundamental parts of the theory of justice which consideration of punishment presupposes. What retributivism involves is not a demand to punish people according to their deserts no matter what; it rather says that punishment is justified by desert and punishment can only be justified by desert. Proportionality is a common way of understanding part of that second clause, but it doesn't rule out situations where you might punish someone with a punishment much lighter than they deserve.
There is another problem with these kinds of arguments, although on this point, in fairness, it's a problem that does plague a large number of contemporary retributivisms. This is the notion that 'proportionality' indicates some sort of mathematical or quasi-mathematical measure. (We find the same problem haunting discussions of proportionality in just war theory.) In reality, 'proportionality' in this context means the proportionality of means to ends, where the general end here is 'addressing the wrongdoing insofar as it deserves to be addressed', which will in turn get specified further in light of what the actual wrongdoing was. This is something that can be reasoned out as to kind and as to quality of response, but it is not the sort of thing for which precise measures usually exist, particularly of the consequentialist sort that the people often assume. The reason people (including some retributivists) slip into thinking of proportionality as mathematical has less to do with core requirements of retributivism and (probably) more to do with the odd tendency of contemporary philosophy toward clunkheaded literalism. But 'proportionality' in these contexts is very old, and it is the proportionality of means to end, as found in theory of action; derived, indeed, from the even older mathematical usage, but only derived. And whatever its derivation, it has had its own meaning since the ancient Greeks. It's true that this means you can't reduce everything to a handy formula; but as no such magical formula could exist, and retributivism doesn't predict that it would, there's no problem here for retributivists.
