[Punishment in the appropriate measure] proclaims that what [the perpetrator] did was wrong, that the victims did not deserve
their victimization, and that they were important enough to be worth
So (to fill the argument out), retribution is that component of punishment that serves the goal of expressing moral outrage about the perpetrator's wrongful conduct.
I think this line of argument begs the whole question: Moral outrage is what whets the retributive juices in the first place, and if retribution is illicit, then my desire (as a victim of a criminal wrong) to have the state express my moral outrage retributively is also illicit.
Now, even an upstanding nonretributivist, qua student of public policy, would have to be attuned to the fact that public resentment in response
to criminal wrongdoing is a brute psychological fact--the kind of fact that good
policy making requires taking into account. Thus, even nonretributivist policy makers might opt to
channel resentment through some extra measure of punishment. On this view,
incarceration isn't an expression of an authentic duty owed to the victims in light of
their (presumably justified, or at least understandable) resentment so much as
it is a pragmatic accommodation of (or an accession to) psychologically
However, note that this argument takes us far afield of any classically "retributive" theory of justice. The
reason nonretributivist legislators would mete out Kleiman's sort of "expressive punishment" (as I'll call it) here isn't that victims
deserve to have expressive punishment meted out on their behalf.
Rather, it's that that's the only way we can think of to assuage popular resentment and
rehabilitate the victim. If that's right, then what does the justifying
work in this scheme of expressive punishment isn't the retributive idea itself,
but some independent restorative or restitutional principle.
1I'm assuming for sake of discussion that failing to accede to public resentment in this way is a greater wrong than failing to shield the perpetrator from whatever extra margin of punishment such accession entails. (And I'm probably assuming a lot more than that.)
This NYT article reports on erstwhile Rep. Randy Cunningham's (R) resignation after having plead guilty to taking "at least $2.4 million in bribes."
I haven't followed the Cunningham story, but judging from the article, he seems to have struck the right kind of tone (namely, atoning) in his post-plea statement. For whatever that's worth.
But the fun part about the article is the almost surreal litany it provides of Republican operatives targeted in independent investigations currently underway: Frist, DeLay, Scanlon, Abramoff, Libby, Cheney, Rove. I mean, wow.
Now, smoke doesn't necessarily mean there's fire, but this isn't a case of an "independent" counsel affiliated with and appointed by one party to investigate the activities of members of the competing political party. This is a case where three truly independent investigations by career prosecutors have ended up focussing on prominent members and affiliates of a single party. At the very least, such a result should eviscerate the talking point (still aped in the "liberal" media) that both parties are beset by the appearance of corruption. A difference in this degree makes for a difference in kind.
UPDATE: Dangit! If only I'd spelled it "Ballance" in the title...
[P]sychology experiments reveal that people are often satisfied by [the] empty form [of an explanation]. For instance, when experimenters approached people who were standing in line at a photocopy machine and said, "Can I get ahead of you?" the typical answer was no. But when they added to the end of this request the words "because I need to make some copies," the typical
answer was yes. The second request used the word "because" and hence sounded like an explanation, and the fact that this explanation told them nothing that they didn't already know was oddly irrelevant.
Slate's Paul Boutin limns physicist Lawrence Krauss' complaints about string theory:
Unlike relativity and quantum mechanics, it can't be tested.... When I asked physicists like Nobel Prize-winner Frank Wilczek and string theory superstar Edward Witten for ideas about how to prove string theory, they typically began with scenarios like, "Let's say we had a particle accelerator the size of the Milky Way …" 
The string theorists blithely create mathematical models positing that the universe we observe is just one of an infinite number of possible universes that coexist in dimensions we can't perceive. And there's no way to prove them wrong in our lifetime. That's not a Theory of Everything, it's a Theory of Anything....
Or as Wolfgang Pauli once said (perhaps making a slightly different point), "That's not right; that's not even wrong."
In any case, interesting times for science. On the one hand, and if the string theorists are to be believed, the central ideas of physics have trudged so far ahead of the data that the new physics legitimately could be called metaphysics (not that there's anything wrong with that...), at least in that disputes turn more and more on "a priori" (given the background data, i.e.) arguments rather than new empirical discoveries.1 On the other hand, mathematics--the paradigm of a priori "science"--waxes "experimental" (using computer routines to generate data that probalize or deprobabilize this or that conjecture). Weird, wild, wacky stuff.
1 On this score I'm not sure Krauss and the string skeptics identify a serious problem for speculative cosmology: The "discovery" of new arguments even during what you might call the empirical era has been at least as important as the discovery of new data. At all events, the background data at this point constrain theory significantly, and if a new theory can significantly unify and explain heretofore disparate theories, well, that ain't chopped liver.