Although sometimes I feel bad when my passing comments inspire prolonged reflection, I do find it gratifying when a mathematician agrees with me:
I am hesitant to apply the label witch-doctor to doctors who study and attempt to heal minds, but the label may be valid. First, a quick case-study.Yeah, it is. But you mean before the fact. The case is even worse than that: it's impossible to distinguish between them even after the fact, except in one case: the rare case where a 1B engages in murder or suicide. A 2A can't prove he is not a 1A; and a 1B who doesn't end up hurting anyone looks just like a 2B.
If a person comes before a mental health examination for anti-social tendencies (with or without any noted predilection towards weapons ownership), the possibilities are:
(1A) The person is a danger to himself and others, and the examiner decides that he must be locked up.
(1B) The person is a danger to himself and others, but the examiner decides that he should not be locked up. (This could happen through several modes. Two possibilities are that the examiner misjudges the level of danger, or the examiner misjudges the examinee as not being dangerous.)
(2A) The person is not a danger to himself and others, but the examiner decides that he must be locked up . (Here the examiner erroneously diagnoses a non-dangerous person as dangerous.)
(2B) The person is not a danger to himself and others, and the examiner decides that he should not be locked up.
Given that this is a prediction of future actions based on present observations, and that the future actions cannot be compared to a control-case in a lab, I can agree that such determination is much closer to the activities of a witch-doctor than the activities of a scientist.
A determination that a person is likely to be a danger to himself and others requires a lower level of proof than the determination that he certainly will be a danger to himself and others in the future. A prediction that a person certainly will engage in pychopathic murder is a prediction that requires omniscient foreknowledge. Absent such certainty, it is very hard to distinguish between cases (1A) and (2A) given above--or between cases (1B) and (2B).
Which means what? It means that if you weight the system to 'prevent another Virginia Tech,' it will learn to treat all cases from a pro-lockup perspective. You can't really prove the guy was wrong to lock you up, so he has nothing to lose; but if he didn't lock you up and you happened to go on and do something bad, he's liable at least for criticism, and possibly for legal difficulties.
Thus, there's a strong economic incentive for a psychologist to strip liberties on any occasion they're asked to do so; and no economic disincentive, given that there is no standard of proof that can "prove" sanity or stability. There may be a moral or ethical disincentive; and then again, there may not be.
Inalienable rights aren't, sadly, in a practical sense -- our government has busied itself finding ways to alienate them almost from the moment it proclaimed them. Putting the power to alienate a man from his rights in the hands of people who have no reason to do anything else, and no final, scientific and falsifiable standards against which their decisions can be challenged, is no way for a liberty-loving people to act.