In my column yesterday, I noted proposals to protect the public from armed lunatics like Jared Lee Loughner by making it easier to lock them up before they become violent. New Republic blogger William Galston, for example, argues that "the law should no longer require, as a condition of involuntary incarceration, that seriously disturbed individuals constitute a danger to themselves or others." Instead of this standard, which was adopted in response to the Supreme Court's conclusion that a psychiatric diagnosis by itself is not sufficient grounds for confining people, Galston says "a delusional loss of contact with reality should be enough to trigger a process that starts with multiple offers of voluntary assistance and ends with involuntary treatment, including commitment if necessary." To enforce this new standard, he says, "those who acquire credible evidence of an individual's mental disturbance"—including "parents, school authorities, and other involved parties"—"should be required to report it to both law enforcement authorities and the courts, and the legal jeopardy for failing to do so should be tough enough to ensure compliance."

In short, Galston wants a system that compels Americans to keep a close eye on their odd relatives, friends, neighbors, students, and employees, reporting them to the authorities when their strange ideas escalate into "a delusional loss of contact with reality." That distinction may prove hard to draw. The evidence of Loughner's mental illness, the "warning signs" that pundits say indicated he was dangerous, consisted mainly of the eccentric opinions he expressed in college classes, conversations with friends, online discussion threads, and YouTube videos. Many of the things he said on subjects such as grammar, mathematics, lucid dreaming, and monetary policy were inscrutable or demonstrably false. But if that were enough to signal a break with reality justifying involuntary commitment, our mental hospitals would be overrun.

The fuzzy line between Loughner's ideas and his illness was illustrated by a quote from one of his friends. "He was a nihilist and loves causing chaos," the friend told The New York Times, "and that is probably why he did the shooting, along with the fact he was sick in the head." Was Loughner's nihilism a symptom of his illness, a cause of it, or an independent factor motivating his crime?

As difficult as such matters are to disentangle after the fact, it is even harder to say ahead of time which of the country's many oddballs and malcontents will convert bizarre ideas into homicidal actions. In retrospect, every strange or off-putting thing that Loughner did or said marked him as a dangerous madman, including not just overtly crazy stuff like his video linking Pima Community College to genocide and torture but borderline behavior such as singing to himself, talking out of turn, pestering teachers about grades, smiling or laughing inappropriately, and making weird comments in class. But if you read the records concerning his disciplinary troubles in college without thinking about the crime he eventually committed, it is not hard to see why administrators and the police might have seen him as a pain in the ass with psychological problems rather than a public menace. The psychiatrist E. Fuller Torrey, one of the critics advocating laxer commitment rules, says that even among people diagnosed as schizophrenics, only 10 percent ever become violent. So assuming that Loughner qualifies for that label, a policy of protecting the public by detaining people with similar symptoms would sweep up nine harmless people for each future criminal.

Although they are routinely called upon to say whether people pose a danger to themselves or others, psychiatrists are notoriously bad at it. "Over thirty years of commentary, judicial opinion, and scientific review argue that predictions of danger lack scientific rigor," notes University of Georgia law professor Alexander Scherr in a 2003 Hastings Law Journal article. "Scientific studies indicate that some predictions do little better than chance or lay speculation, and even the best predictions leave substantial room for error about individual cases. The sharpest critique finds that mental health professionals perform no better than chance at predicting violence, and perhaps perform even worse."

The current system of involuntary commitment rests on predictions of dangerousness that are appallingly inaccurate. Abolishing the requirement of dangerousness would avoid that embarrassment at the cost of imprisoning even more people who pose no threat to others. Yet Galston insists that individual rights should not stand in the way of a mental-health dragnet that promises to take potential mass murderers off the street. He warns that "the rights-based hyper-individualism of our laws governing mental illness is endangering the security of our community and the functioning of our democracy." His lack of concern about the civil liberties issues raised by his proposal is reflected in the glib "warning label" at the top of his essay: "This article will make civil libertarians unhappy. Read at your own risk." Similarly, Time blogger Joe Klein, who regrets that "we no longer lock up the mentally ill," concedes "there will be all sorts of civil libertarian objections to this." Like Galston, he seems to view those objections not as legitimate concerns that need to be addressed but as obstacles to overcome.