But what should we do when rehabilitation repeatedly fails for a particular individual?
Bit of the UK currently have a persistent prolific offender strategy. There's no national clear definition of a PPO, but there are estimated to be about 8,000 nationally. (In a country of about 70m). I've seen estimates that about half these people are in prison. (One old definition of a PPO was someone over 18 who is convicted of 6 or more recordable offences in the past 12 months. But that definition is old, and people now look at the impact to the local community rather than number of crimes.)
The PPO strategy has several layers.
Offenders are identified, and provided with large amounts of support to rehabilitate. This includes education, housing support, drug and alcohol treatment, and so on. (I don't know whether it looks at undiagnosed brain injury, which are common in prisoners).
When they're in the community they are closely monitored. Any criminality is rapidly detected, prosecuted and convicted, and punishment is rapidly applied. Most of these people will be out of prison "on licence", which makes returning them to prison easier. This makes it very clear that criminal behaviour is not tolerated. Sometimes the justice system does not make that clear!
The aim is to reduce the amount of offending the person does, but also to reduce the fear of crime in the local community.
Put the person on supervised release, put an ancle monitor on the person and put the person to community work, until they become a net benefit to society.
Other solutions may apply, but at least the above is better than three strikes.
That brings to mind an interesting question. Why should prison time be proportional to the crimes committed?
If the purpose of prison is rehabilitation, then shouldn't the time be however long it takes to rehabilitate the prisoner? There might be some correlation between that and the crimes committed, but I would expect that to be a weak correlation. There will be many people who are easy to rehabilitate whose first crime was very serious and many people hard to rehabilitate whose first crime was relatively minor.
Society only obeys the rule of law as long as there is some consensus that the law is fair and just. Basic fairness demands that breaking a rule should result in a punishment.
The punishment is only fair if it is proportional to the offense. Therefore prison time should be proportional to the crimes committed.
While punishment for crime certainly appeals to our instincts, should we not consider more "humane", and possibly more effective crime reduction measures (rehabilitation over punishment)?
The original discussion was about alternatives to three strikes (which explicitly both of us oppose). Notice the lack of value judgement of rehabilitation vs. punishment.
Secondly we were discussing the concept of proportionality. I expressed the rationality of society's need for proportional punishment. Notice how I did not present my personal views regarding rehabilitation vs. punishment.
So far we have only established that I believe that the punishment for a crime should have an upper bound which is proportional to the offense.
Now, whether the purpose of the prison system should be punitive or rehabilitative is another discussion.
If the system that is supposed to be rehabilitating people is failing sizable classes of individuals (which is the problem -- not a very few isolated individuals -- that broad three-strikes laws, particularly the pre-reform California version, address), one might consider that the problem is with the design of the system for rehabilitation, and ask how to make that work better, rather than progressively subjecting people for whom it doesn't work to it for longer, and then giving up, the way three strikes laws and other ways of taking post convictions into account in sentencing do.