Psychiatry is either in a state of complete disarray, or it's on the verge of triumph over mental illness.
The science behind the discipline is either rigorous or completely fallacious.
Compared to other medical disciplines, psychiatry faces the daunting burden of having to regularly prove to the universe that it exists, that there is something besides air in the open boat, that its practice can be useful, and that the significant lack of known knowns about mental illness do not correlate to a reality where there never will be knowns that will be known better.
What psychiatry is is a serious question, and one worth grappling with. The controversy over the profession's diagnostic standards manual is illuminating and well worth the thought that thinkers give to it. But I don't think it serves the mentally ill well to label the profession that serves to help them as intellectually bereft and scientifically dangerous.
Consider the case of a widely prescribed pill, one that doctors hand out like candy, with fairly minimal side effects. It has made billions for its manufacturer. This pill is designed to treat a discrete medical problem that doctors know a lot about. They understand the way the pill interacts with the body, they know how it metabolizes, and what systems it affects. And yet, for reasons unknown to docs, it works about half the time.
I write, of course, of Viagra.
The scientific world has a pretty good handle on what causes erections. And men be damned, they can't get a pill to work more than half the time. Imagine all of the sex that's missed!
One of the foundational criticisms of modern psychiatry is that the cures, which are designed to address symptoms because the underlying biology remains a black box, don't work for most patients. Or, they work weakly. Considering how far we are from addressing the question of consciousness, and since psychiatric symptoms are often (though not exclusively) symptoms relating to cognitive (thinking) or valence (emotional) distortions, it should be encouraging that psychiatry — the doctors, the drug manufacturers, the patients — have managed to come this far.
Psychiatric patients know intuitively what even the head of the National Institute for Mental Health seems to want to minimize, which is that most psychiatric illnesses are bio-psycho-social disorders. Overreacting to the criticism that psychiatry is not scientific enough, the profession has put a lot of energy into sounding more like, say, cardiology. They do this at the expense of understanding. Biology surely plays a role in mental health. But just like the problem of consciousness runs into the "what makes red RED" problem, and solves it, at least temporarily, by noting that we conscious beings live in a real world with other people and other things that affect us daily, mental health is not going to be understood in the same way that we understand heart conditions. (Actually, stress — external stress — can exacerbate a variety of cardiac conditions as well).
Treating mental illness requires good drugs, good talk therapy, and — a big "and" — significant attention to the chronic stress and potential threshold stressors that worsen pre-existing conditions or diatheses. I think psychiatry is still kind of afraid to talk about illness as something that lives in a community as well as a body, but as many people who have been successfully treated by the profession will tell you, their symptoms did not abate until they figured out how to change the conditions that triggered them. Doing that often requires resources, expertise, and, often, money, as well as relative job and relationship stability. We may be a ways away from concrete biology, but if the goal is to restore function to people who cannot function, then that's where we ought to look.