The articles on genetics and crime by Adrian Raine, Krebs and Kacelnik, and Michael Rutter (É«ºÐÖ±²¥S, February 10) give an excellent picture, while the comments by other contributors add interesting points. The only jarring exception is Steven Rose's ad hominem remarks that violence research "is framed within a reductionist and deterministic paradigm which seeks the causes of social problems in individual biology, fostered by a political philosophy . . . which rejoices in the privileges which come with inequality in wealth and power and rejects steps to diminish them".

Such old-fashioned Marxist-Leninist fundamentalism does not fit well in a discussion of scientific findings, and leads one to ask how Rose comes to know the political philosophies of the researchers concerned - I know most of them, and cannot really recognise the picture he paints.

Any sensible approach to psychological problems must begin with the realisation that man is a biosocial animal, and that his behaviour is governed by both biological and social determinants, the relative importance of which varies from time to time, and from situation to situation. How a person acts in specific situations will be governed by heredity and by past learning and experience; it is hard to understand how John Hapgood's "free will" comes into the picture. If an action is not determined by heredity and past learning, how is it determined? Is it random, a kind of Brownian movement in the cortex? Until we are told, and given experimental proof of "free will" we might be better served by looking at more robust evidence. In doing so we have to come to grips with the problem of causality, already recognised by David Hume 250 years ago, and complicated beyond belief when considered in the context of human behaviour. What we are dealing with are risk factors, ie factors correlated with behaviour but not necessarily directly causally related to it.

Most alleged causes of violent and criminal behaviour (genes, poverty) obviously cannot act directly to produce such behaviour. All that genes can do is to produce polypeptides, which compose proteins, including structural, transport, and catalytic proteins (enzymes). Research has to focus on a long line of intermediary products, such as hormones, neurotransmitters, etc. Low serotonin, low cholesterol, low mono-amine oxidase and high testosterone are correlated with violent criminality; what we have to do is to investigate the reasons for their apparent success in producing such behaviour.

É«ºÐÖ±²¥

Also needed is research into the interaction of such biological factors with social ones; there are reasons for suspecting such interactions to be synergistic. Similarly, poverty as such cannot produce violent criminality; indeed, the opposite may be true. In the early days of the Weimar republic, extreme poverty and unemployment were widespread, to an extent unknown in England, yet violent criminality was minimal. During the past 30 years, the living standard in England has risen dramatically, but so has violent crime. No one has even begun to trace the causal relations, if any.

Does, or should, genetic research provide a legal excuse for unpunished aggression and violence? One could argue that biologically people are variously predisposed to such violence, and that those most predisposed have a better excuse than those least predisposed. One could also argue that those most predisposed require much longer periods of detention to safeguard prospective victims. These are legal and ethical questions scientists are not well equipped to answer. They are more likely to respond to demands directed at finding an answer to the practical problem of reducing this predisposition.

É«ºÐÖ±²¥

Contrary to the belief that if something is genetically determined, nothing can be done about it, the evidence suggests that knowing how the genetic factors work may enable us to circumvent this action.

Phenylketonuria is an obvious example; this is a disorder leading to oligophrenia, due to a single recessive gene. Children so affected are unable to convert phenylalanine into tyrosine, with the incomplete breakdown of products poisonous to the nervous system. Giving infants a diet free of such substances sidesteps genetic influences and abolishes the mental effects associated with phenylketonuria. Perhaps we can sidestep genetic factors predisposing people to criminal violence similarly?

Recent research into violence in prisons in the United States has shown that this can be reduced by 40 per cent by simply giving prisoners vitamin and mineral supplements; the effect in several different studies was astonishing. The É«ºÐÖ±²¥ Office and prison governors in the United Kingdom have shown little interest in this research. One might think that here we may have a useful tool for reducing violence generally; perhaps if we gave all youngsters with vitamin-mineral deficiencies supplements at school, this would lead to a general decrease in violence. At least we might expect an experiment along these lines to be done at some inner-city schools particularly troubled by violence. (We have found by blood analyses that about one-third of apparently well-nourished children are suffering from vitamin and mineral deficiency to a significant degree. The proportion among inner city children would be expected to be larger).

Clearly there is no single cause of violent criminality. We must deal with complex, interacting biological and social factors jointly responsible. Research should be directed to determine the mode of interaction, but particularly to discover practical ways of reducing violence. Special pleading and ad hominem arguments about the motivation of researchers are unlikely to further such research.

É«ºÐÖ±²¥

HANS EYSENCK

Professor emeritus of psychology

Institute of Psychiatry

University of London