I was going to reply to you upthread in the specific comment thread we were having, but instead will reply here.

I agree with much of what you've been saying throughout this comment thread. I think that we're probably about 75% in agreement about the goals of and preferred approaches to criminal justice.

Two areas where I suspect we differ slightly:

1. I do see some utility in retribution besides simple restoration of "karmic" or "cosmic" balance.

Whereas I agree with you that on a personal scale, I don't find much utility in punishing so as to "balance scales" or what have you, I also understand that you and I are in a distinct minority. Particularly among the Abrahamic religions (which comprise an overwhelming majority of both the global and U.S. populations), retribution is integral to the concept of criminal justice in those belief systems. Therefore, not only is it inevitable that a criminal justice system - a political creation - will reflect the beliefs of a majority of a given population's members, it should do so. For failure to do so leads to the pervasive sense that the justice system doesn't work which leads to both a) more criminality; and b) all manner of vigilantism and lawlessness.

In other words, it's important that we change people's beliefs about criminal justice first, THEN we change criminal justice.

2. An area which, IMHO, you are overlooking is the fourth goal of criminal justice. You referred to three upthread (a great comment, BTW), but in my criminal justice studies I was always taught four - with the fourth being general deterrence: the notion that it's important for a society to signal at-large that crime will be punished, and specific crimes will be punished in a specific way.

So while you've been talking a lot about recidivism, or preventing specific criminals from re-offending, there is a view that the justice system should seek to avoid offending in the first place by sending such clear signals.

I'll concede that the justice system shouldn't be the primary method through which we seek to prevent crime from happening in the first place. We should focus on root causes of crime: economic inequalities and deprivations, substance abuse, child abuse, mental health problems, etc., even beyond - into the newer scientific frontiers of understanding the human brain and genetic predispositions to various biological traits that might correlate with criminality.

However, that doesn't mean that we should exclude from the criminal justice system the objective of preventing crime in the first place. It can and should be part of a multi-faceted approach that includes the aforementioned societal changes. Because while the empirical evidence shows that severity of punishment does not reduce criminality, the prospect that criminality will be apprehended and punished does appear effective at reducing it.

In other words, giving people "Get out of jail free" cards, by which we completely forego privation of liberty in lieu of therapeutic methods could, and likely would result in increased criminality. It simply wouldn't be seen as a real punishment.

I think your approach rests on the assumption that rapes happen almost by accident. That they don't really know better. Everyone is inherently good or wants to be good, and some just stray from the path. If they were confronted with their victim's pain and you explained why it was wrong, they'll have an epiphany and refrain from doing it again. I think by and large they already know how wrong it is and they just don't care. Some perfunctory counseling isn't going to change their disposition, any more than "a better education system" is going to teach conspiracy theorists "critical thinking" skills that disabuse them of their beliefs. Something inside them is just fundamentally broken and it's probably beyond our abilities to fix right now.

Now, if you want to argue for some kind of early intervention system I think that'd be a great idea. Trying to predict and treat these kinds of behaviors in schoolkids before they crystallize could be worthwhile.

The approach that most probably reduces recidivism and reoffence rate is elimination. It’s the most hardline and brutal approach, but also the most effective. It likely wouldn’t be accepted or legalized in most modern legal systems, but I also doubt that American society in general is prepared to enact change based on the best available research, at least not in the current generations.