my problem with Szasz though is different. He seems to accept the basic assumptions of mental illness that our society does, and therefore never frees himself from the problems of those assumptions. There is no scientific basis for these assumptions, as they form, effectively, the threshold questions under which one can try to make neurochemical inquiry into mental health. These assumptions are:
1. Mental illness, to the extent it exists, is objective and quantifiable.
2. Social context is irrelevant to mental health.
3. Therefore mental health is an individual issue.
Of course a libertarian will accept these assumptions because they are based on the assumptions of personhood behind that movement. But what if both of the first two are wrong? What if mental health is very much subjective and what if social context is an important factor in mental health? If so, then, not only is the science based on bad assumptions but so is Szasz's rebuttal to it, and mental health becomes more an art than a science.
None of this reaches the question of coercive therapy. The fundamental questions are who, when, and how (both regarding the coercion and the therapy). Obviously there are times when this is needed. But we should not ignore the dark side this has had throughout history.