Click to enlarge

"This magazine is flush with tight smart writing."
Washington Post

























Oliver James

Why Do We Fight?

The Anatomy of Violence: The Biological Roots of Crime
By Adrian Raine (Allen Lane/The Penguin Press 496pp 25)

Since the publication of the map of the human genome by the Human Genome Project (HGP), genes have, at least so far, been shown to play astonishingly little role in explaining why one sibling's psychology is different from another's. This is an incontestable fact. No molecular geneticists can dispute it since the vast majority of the many hundreds of studies performed have shown that less than 1 per cent of the variance of any psychological trait - be it intelligence, mental illness of all kinds or personality - is explicable by differences in DNA. The most that almost any studies find explicable by DNA is around the 5 per cent mark.

The conclusiveness of these findings has created a quandary for the psychiatric and psychological establishments. Because twin and adoption studies have led them to expect much higher heritabilities, they assume that the genes must be there somewhere. The holding position has been to explain their absence as 'missing heritability' or a 'DNA deficit'. But they have looked pretty much everywhere that the genes could plausibly be, using highly sophisticated techniques and huge samples.

Some are now pinning their hopes on finding that hundreds or thousands of variants within individual genes make tiny contributions which, added together, produce the difference. Having called for even larger samples (requiring more money), they continue to draw a blank. Others are searching in the 'junk gene' areas on chromosomes, maintaining that the needle is somewhere in the haystack. But off the record, many molecular geneticists admit that they doubt they will find anything more. What will it take for them to admit this in public?

For a scientist like Adrian Raine, the HGP evidence must be very painful: he has devoted four decades to painstaking study of the brains of different criminals, especially the violent kind. His assumption throughout has been that genes primarily cause electro-chemical differences in the brain that predispose individuals to violence. While he has always included a role for environmental factors, they have occupied a secondary position.

Like many recent scientists-turned-authors who have been confounded by the findings of the HGP, Raine's book struggles to reconcile the facts with his thesis. Though there is little doubt that the brains of criminals do differ in a few respects, at least in some cases, from those of non-criminals, he has to bob and weave around the evidence in order to avoid the increasingly unavoidable: that it is non-genetic factors that explain electro-chemical differences between people. While he does make a case for nutrition and levels of metals as causal factors, he largely ignores the massive body of evidence that early and later parental care is the critical independent variable. A book about the causes of violence whose extensive index does not include the words 'abuse' or 'trauma' suggests an author who has either not read the evidence or is ignoring it. I have interviewed over 150 violent men and found only two who had not been severely physically abused in childhood, yet Raine barely mentions this factor in his book.

His introduction indicates his bias. He asserts that 'the dominant model for understanding criminal behaviour has been, for most of the twentieth century, one built on sociological and psychological models'. This is parsimonious with the truth. He appears to have forgotten the hugely influential claims made by Hans Eysenk and Cyril Burt, and the 1994 book The Bell Curve by Richard Herrnstein and Charles Murray, which maintained that genetically inherited low intelligence, especially among African-Americans, explained much criminality.

Starting with the claim that geneticists have been fighting an uphill battle against a predominantly nurturist environment is a trick every genetically minded author seems to play. A recent example was the muddled book by the twin-study scientist Tim Spector. Raine, like Spector, contradicts his own contention that the dominant culture is against him. Only one page after bemoaning the dominant nurturist model, Raine states, 'Thinking of human behaviour from a biological perspective is no longer controversial - you can hardly open a newspaper or magazine today without reading about a new breakthrough in how genes and the brain shape our personality.' Like Spector, he presents himself as a scientific crusader battling against cultural orthodoxy, and at the same time wants to argue that the scientific establishment is so much on his side that his views are uncontroversial. This latter position is rapidly cracking, as the recent criticisms of the latest version of the psychiatric bible the Diagnostic and Statistical Manual of Mental Disorders have shown. The British Psychological Society's Division of Clinical Psychology issued a statement that there is simply not enough scientific evidence to support the claims of a genetically caused neurological deficit in the mentally ill.

Raine's problem is that the HGP evidence does not support his claims about the importance of genes. Having debated the evidence with him on a Newsnight programme devoted to violence, I know him to be a seeker after truth. But his book seems to ignore evidence that does not suit his case, and occasionally his selectivity borders on distortion of key contrary evidence.

The crucial example is a 1984 study by Sarnoff Mednick of 14,000 Danish adoptees. Using property crime statistics, it found a significant relationship between rates of conviction for adopted children and their biological parents, but no such relationship between the children and their adoptive parents. This suggests a role for genes. However, as Sarnoff makes absolutely clear, there was no such relationship between convictions for violent crimes. Just so there can be no misunderstanding, I quote Mednick's statement in the abstract of his scientific report: 'A statistically significant correlation was found between the adoptees and their biological parents for convictions of property crimes. This was not true with respect to violent crimes.'

Raine wheels out this research as part of a review of a hundred twin and adoption studies, which he claims lead inexorably to the conclusion that 'genes give us half the answer to the question of why some of us are criminal and others are not'. But note that he is talking about crime here, not specifically violence. As he well knows, the twin and adoption studies that have been performed show no role for genes when violence is being measured. Yes, aggression in general and antisocial behaviour are found to be genetic, if you trust those studies (which you should not, as I shall explain in a moment). But there is a blatant sleight of hand here in moving from violence in particular to crime in general.

Raine knows that Mednick's evidence contradicts his thesis. Over the course of our half-hour with Jeremy Paxman, I was at pains to make this point about Mednick's study and he accepted it. Nor is this the first time I have encountered equivocation about the Mednick study - I had a similar run-in about it with Steven Pinker on Radio 3.

If Raine could face it, he would accept the findings of the HGP and abandon twin and adoption studies as evidence of genetic determinism. In the case of twin studies, it has been pointed out for years that, insofar as identical twins are seen to act more similarly than non-identicals, it could be because they are treated more similarly - the evidence strongly suggests this. Taken with the HGP findings, it is time for the likes of Raine to face the truth about the role of genes and, indeed, for university departments to stop teaching twin studies as evidence for genetic effects.

Yet Raine believes the HGP supports his argument because a 'warrior gene' has supposedly been identified and associated with violence. He ignores the evidence that the initial findings have been repeatedly unreplicated (only the replications which support his argument are mentioned). At one point, in this endeavour, he hits an embarrassing and illuminating problem. It turns out that Maori males in New Zealand are more likely than Caucasians (56 per cent vs 34 per cent) to have the alleged warrior gene. Perhaps because such a finding might lead to accusations of racism, Raine sweetens the pill by pointing out that 77 per cent of Chinese males have the gene, though they are far less violent. Without a backward glance, he does not seem to have noticed that he has all but disproved the role of the warrior gene: if the gene was so important, surely rates of violence would be far greater in China?

Adrian Raine's book is interesting for only one thing. It is part of a growing literature by establishment psychiatrists and psychologists who have simply refused to believe the evidence of the HGP. As a statement about the true causes of violence, it is largely worthless.


Exclusive from the Literary Review print edition. Subscribe now!


Oliver James trained and practised as a child clinical psychologist. He is the author of Affluenza: How to Be Successful and Stay Sane and Love Bombing: Reset Your Child's Emotional Thermostat.


Royal Literary Fund


John Murray