Breaking out of the straitjacket

We know what mental illness looks like - but what is it exactly? That question prompted philosophers Dominic Murphy of the California Institute of Technology in Pasadena and Stephen Stich of Rutgers University at New Brunswick, New Jersey, to brave a radical rethink. In a chapter of the book Evolution and the Human Mind, they argue that the mind is a set of structures with specific functions and that mental illness happens when those structures fail in a particular environment, or when the structures themselves "break". Here, Murphy explains how their theory contrasts with prevailing views of mental illness as a biological disease in the brain and why there may be nothing "wrong" with psychopaths

What was your view of human nature that let you take a fresh look at mental illness?

We started with the fairly standard evolutionary psychology view that the mind consists of a large stock of modules, each crafted by natural selection to do certain jobs. Our idea is that the mind is a set of structures with evolved functions and that mental illnesses can best be understood as failures of those structures. This is very different from the idea dominant in psychiatry, that mental illness is a brain disease. A lot of psychiatrists traffic in molecular explanations and tend to miss the cognitive picture of the flow of information between those modules.

How do these structures go wrong?

Either something goes wrong with the system itself or a system that was functioning perfectly well ends up in an environment it wasn't designed for. Anxiety disorders and phobias look like good candidates for the second category. We know that we find some things very scary in the sense that we become frightened of them after minimal contact. And if you look at the way the phobias are listed in the psychiatrist's diagnostic bible, DSM [Diagnostic and Statistical Manual of Mental Disorders], they're specific phobias about blood, heights, snakes and so on. The most common phobias do seem to map quite well onto fears that would have made perfect sense in our evolutionary past. And unlike other disorders, the form of the phobias fits the evolutionary functions quite well. So, for example, we faint in response to loss of blood: fainting prevents you going into shock because the physical system closes down.

What about when the system's gone wrong?

The mind will also have other general-purpose components, and there may be some very important disorders that we need to attribute to central reasoning systems. An example I like is the Capgras delusion, where people think that somebody close to them, usually a spouse, has been replaced by a replica. There are two problems here. First, there's something wrong with the face-recognition module: it's giving you a match but not the emotional response that confirms this is your spouse, so the experience is of seeing an identical twin to your spouse. There's also something wrong with the way these people are thinking. It may be that they have a higher upper boundary for permissible beliefs. That might help explain Cotard's delusion, too, in which one forms the belief that one is dead. The big difficulty is that we don't understand central reasoning systems at all: some researchers believe they are beyond the reach of current science. If that's right then we can't have a good information-processing explanation of delusions or other irrational states - and a lot of psychiatry will struggle to advance much beyond its present state.

Where does all this leave DSM?

DSM is designed to be atheoretical. It doesn't talk about causal theories of particular disorders. It doesn't say that there isn't a theory of what a mental disorder is, but it is designed to avoid causal theories. People sometimes use the phrase "Chinese-menu approach" to characterise DSM: to meet a certain diagnosis, you have to have two symptoms from list A, two out of four from list A, three out of six from list B plus either C or D, and so on. I think that philosophy of science suggests that you really can't have a satisfactory classification without it being based on some causal understanding of what you're classifying.

It sounds almost like medieval taxonomy...

Yes. All those lists! It's also a document written by dozens of semi-autonomous committees, each dealing with a particular condition and none of them worrying about the big picture. It's not clear that the pieces even fit together terribly well. As a scientific document, DSM might be better off ignoring traditional or common-sense views about what a mental disorder is and looking to the other sciences of the mind for a guide to mental functioning. Then you could have a document that details all the ways our mind/brain can break down, including disorders of the visual system, all the agnosias and so on. But they're not considered "mental" illnesses by psychiatrists! These false distinctions result from the divorce between psychiatry and neuroscience. This is why we've tried to produce a cognitive neuroscience approach that has an evolutionary interpretation - and to apply it consistently.

At least that provides a theoretical framework.

That's right. You shouldn't expect this approach to have an enormous short-term impact on psychiatry, but it's interesting that Chris Frith at the Institute of Neurology in London and other people in the US are working on models of schizophrenia in which the cognitive neuroscience approach to information-processing malfunctions is increasingly important. And there are interesting cognitive theories based on the idea that panic disorder is a misinterpretation of bodily sensations. Hence when you stand up too quickly you get that characteristic head spinning, and you interpret this as a symbol of imminent disaster - I'm about to die or I'm going crazy.

Will some mental illnesses "disappear"?

Take personality disorders. Some research suggests these are the result of people being "designed" to be antisocial or manipulative, and that in a game-theory approach to evolution, if enough people are behaving "normally" there might be a niche for antisocial, histrionic individuals to exploit other people or treat them selfishly or indifferently to gain advantage. If that is right, then we have to say that although these people are hell, there's nothing wrong with them. Antisocial personality disorder in particular is probably overdiagnosed. People who work with psychopaths complain that this diagnosis doesn't distinguish between career criminals and young guys who join a gang - and a hard core of genuine psychopaths or sociopaths who are designed to be this way.

Might they have been more adapted at another time in our evolutionary history?

They might, though they might be well adapted to larger, more anonymous, contemporary societies. A lot turns on whether there are many successful psychopaths or sociopaths out there who don't come to the attention of the courts or the health services. The psychopaths we know about tend to be very impulsive and self-destructive and often spend their reproductive years incarcerated, which is the big problem for this view. But many researchers, including ones who don't take an evolutionary line, think there might be rational and successful psychopaths in the general population. Hervey Cleckley's The Mask of Sanity was the standard text on psychopathy for years. Cleckley included in his case studies the psychopath as scientist, the psychopath as physician, the psychopath as psychiatrist, the psychopath as man of the world, and so on. He thought there was a genuine psychopathic type and that you had to look carefully to watch the mask slip occasionally. If you apply our concept of mental illness then these people are functioning as Mother Nature intended because Mother Nature intended them to be bastards!