Distinguishing finality interests between convictions and sentences

As I await more responses to why the Ninth Circuit’s Carrington ruling should be troubling (query here), let me spotlight key distinctions between finality interests regarding convictions (especially state convictions) and regarding sentences (especially federal sentences).  These distinctions are, in my view, critical to a complete understanding of the dynamics of retroactivity doctrines.

Teague, the 1989 SCOTUS case defining modern retroactivity doctrine, was about whether and when a new federal constitutional rule ought to disrupt final state convictions.  Structural concerns about finality are at their zenith here or else every long-ago convicted state defendant (even those who finished their sentence decades before) might run to federal court seeking to have a long-ago conviction wiped off the books.

Carrington, in sharp contrast, concerns whether a new federal constitutional rule ought to allow reconsideration of a federal sentence still being served.  One might argue that there are no finality concerns because the on-going sentence under challenge is not final.   Moreover, the result of the challenge won’t wipe out a long-ago conviction, it will only (perhaps) alter an on-going sentence.

Some recent capital rulings help put the conviction/sentencing distinction in sharp relief.  After Atkins and Roper declared certain defendants ineligible for the death penalty, few asserted that we should still execute all the now-ineligible death row inmates that were sentenced to death before these rulings.  If “sentence finality” is so important, we should still be execute all the mentally retarded and juvenile capital defendants sentenced before Atkins and Roper.