In case there was any doubt that psychiatry is on the march (from Psychiatric Times June 2007):

The mass murders at Virginia Tech [sic: there was only one mass murder] could lead to harsher laws restricting [mentally ill people's] rights... Perlin, professor of law at NYU, predicted that several states will try to change the basis for involuntary commitment from danger of harm to self and/or others to the need for psychiatric treatment. [emphasis mine, but really, does it need emphasis?]

Mr. Perlin said he expects the U.S. Supreme Court will be asked to rule on such a statute's constitutionality within 10 years.  "I am already counting the votes."

Me, too: Scalia, Thomas, Roberts, Alito, Ginsburg-- strange bedfellows, indeed, but Scalia and Ginsburg spend every New Year's Eve together-- against; Souter, Kennedy for; the rest is anyone's guess (Stevens may not even be there.)

The second article, from the same issue, accidentally describes the crux of the psychiatry/violence dichotomy.  In "Mental Health Staff Can't Sue If Injured By Patient," the writer explains how it is rare, and generally discouraged, for staff to sue or press charges against a patient who is violent and injures them.

Patients who attack mental health professionals in hospital settings are rarely prosecuted and usually cannot be sued for civil damages [said] Ralph Slovenko, Ph.D. at the annual meeting of the American College of Forensic  Psychiatry.

...Authorities usually take the position that it would be inconsistent to prosecute a person who has already been hospitalized for reason of mental illness...

... a New Jersey Court ruled [that] to convict a mentally ill person for displaying symptoms of mental illness... could not be justified constitutionally or morally.

Anyone disagree?  Choose carefully.  Here's the problem: the exemption from prosecution isn't for the insane behaving insanely, or the schizophrenic exhibiting psychosis; the exemption is for any patient. "Patient" in this context, is defined as anyone who is in the psychiatric hospital.  In other words, it's not a label based on pathology; it's a label based on geography.

This may surprise many people, but psychiatrists hospitalize non-mentally ill people all the time.  Any resident will lament how often they are confronted with the malingering drug user who fakes suicidality to gain admission.  Well, if you admit him-- strike that, if he has ever been admitted-- then he is a de facto patient.  If he kills you, he is automatically in a different legal status than if he murdered you in a supermarket.  For example: no death penalty.

It goes without saying, of course, that even the presence of mental illness shouldn't free one from prosecution.  That's why we have the legal construct of insanity.

Here's the clincher:

The situation is analogous to the "Fireman's Rule" in tort law, he said. A firefighter... cannot sue the owner of a burning building for injuries sustained in firefighting.

... I assume because a firefighter must have a reasonable expectation of fire-related danger.  Fine.  But if the firefighter, while fighting the fire, gets shot in the face by one of the meth-lab workers inside that the owner of the building is employing to make methamphetamine, is there no basis for a suit?  Does reasonable expectation of a certain level of danger extend to, well, to volitional acts of violence that have nothing to do with the physical structure that the violence happens in?

The reason I mention these two articles together is because they are the same. "Mental illness" is a term so vague and empty that it is dangerously useless.  Reducing one's responsibility, or restricting their freedom, based on such an arbitrary term is, well, insane.  Doing both at the same time is a tacit acceptance of classism; that some have the responsibility to rule, and some have the responsibility to be ruled.

Oh, I know:  everyone hates George Bush because he has no respect for civil liberties.  Ok.