Enough is enough or vujà dé

Not yet another debate on nature, nurture and IQ! In one of the two excellent concluding commentary chapters, Earl Hunt expresses the sentiment as "vuja de", his invented twist around of deja vu. The intended meaning is roughly: "Take it away, I don't ever want to see it again". Many readers will feel the same with chapters that parade again polarising arguments of an all-too-familiar kind. As Hunt puts it: "One of the nice things about acrimony in psychology is that, if you missed it the first time, you can always watch the reruns". On the one side, Tom Bouchard dismisses opposition to behaviour genetics as no more than pseudo-analyses or pseudo-arguments. On the other side, Douglas Wahlsten and Gilbert Gottlieb argue that partitioning population variance into genetic and environmental determinants is uninformative with respect to individual development. Sandra Scarr, in a lively but pugilistic chapter, pits behaviour genetic theories of intelligence against socialisation theories. Not surprisingly, her conclusions imply a first-round knock-out for behaviour genetics.

Nevertheless, it would be wrong to dismiss the book on the grounds that it is simply a rehash of a many-times-served meal. To begin with, there are some ideas and findings that will be new to many people. For example, Scarr gained the generous cooperation of Deborah de Baryshe to fit alternative models to her dataset with Gerald Patterson, which had dealt with the role of antisocial behaviour and ineffective discipline in the genesis of poor academic achievement. The re-analyses, in contradiction of the published conclusions, suggest that neither plays a substantial causal role; instead it appears that parental achievement is the prime driving force - a finding interpreted as suggesting genetic mediation. It is unexpected, too, to find Arthur Jensen urging the possible importance of random epigenetic effects - what have sometimes been called the "third force" outside nature and nurture. The idea is that development is programmed to work in a probabilistic fashion and hence it involves random variation deriving from multiple physical microenvironmental events before and just after birth.

Stacey Cherney and his colleagues use data from both twins and adoptees to model the development of intelligence between one and ten years of age. The findings indicate that, although the non-shared environment (meaning that which operates differentially to make siblings different from one another) has important effects at any one age, it does not drive the developmental process. Continuity in intellectual development, it seems, is mainly shaped by genetic influences because those that are influential at one age are, by and large, the same as those that matter at later ages. For the same reason, however, shared or common environment effects (those making siblings similar) also help shape development. Steven Reznick's twin findings are interesting, too, in showing a different pattern of effects for the use of language and for its understanding, with very minor genetic effects on the latter during the first two years of life. Robert Sternberg's account of trying to match styles of teaching to patterns of cognitive skills is fascinating but frustrating because it provides only a summary sketch of the preliminary findings available in 1992.

What should we conclude from all of these ideas and findings and where do we go from here in taking the field forward? In many ways, the two commentary chapters that aim to address both questions are the most interesting. Their answers have in common the well-based conclusion that genetic effects on individual variations in intelligence are substantial. Irwin Waldman suggests that they account for somewhere between 40 per cent and 80 per cent of population variance and Hunt correctly argues that there is no point in trying to be more precise than that. There are no theoretical, policy or practice implications that stem from whether the "true" heritability is 40 per cent or 80 per cent. Either way, major changes could be brought about by environmental alterations. But, beyond that, there is not, and cannot be, any figure for "true" heritability; any figure is population-specific and crucially dependent on the range of environments and genetic backgrounds included in the population. That is well understood by geneticists but nongeneticists still get unnecessarily bothered by whether the figure is "right". Within very broad limits it does not matter.

Nevertheless, that does not mean that variations among population groups are unimportant. On the contrary, they can be very informative. Both Hunt and Waldman highlight the fact that almost all genetic findings stem from predominantly white samples with only a tiny proportion of individuals from high-risk environments. That is obviously the case with adoptees because adopting parents are selected on the grounds that they will provide a high-quality upbringing for the children. Of course, selection is far from perfect but the findings do show a serious under-representation of high-risk environments among adoptee households. That is not necessarily the case with twin samples but, in practice, it has usually turned out to be so. The lack of evidence from studies of ethnic minority groups is important, not because it is likely that their genes are very different, but rather because of the evidence that genetic effects are often greatly influenced by environmental context. There is good evidence that this is so with medical conditions such as diabetes and coronary artery disease; it also applies to characteristics such as height (as shown, for example, by the difference between people of Japanese origin living in California and in Japan). The same applies to schizophrenia where the rate is not raised in the West Indies but is raised in people of Caribbean origin living in this country. Living conditions can make a major difference to traits subject to strong genetic influences. Studies that compare populations living in very different circumstances can be highly informative in the light they throw on the environmental modulation of genetic effects.

Both the commentary chapters argue the crucial need to go beyond statistics to determine how genes bring about their effects and how the dynamic interplay between nature and nurture operates. Waldman has a range of suggestions on how this might be done. The book loses greatly in this connection, however, from the fact that most of the chapters were apparently written several years ago. Thus, Scarr's chapter gives the high-profile book by Robert Plomin and Gerald McClearn as "in press" despite its publication in 1993. Molecular genetics is a fast-moving field and the volume suffers from being badly out of date. It has little to say on how molecular genetic findings could prove highly informative on the routes of genetic mediation. There is plenty to be said on this topic, but readers will have to look elsewhere to find it.

There is an equally serious lack of discussion of what might be involved in the interplay between nature and nurture and of how the research questions can be tackled. Hunt argues that the dispute is between the scientific views of behaviour geneticists and the humanitarian views of those who espouse a socialisation perspective, urging the need for a sea change in the views of those on the nurture side. Thus, he criticises the vague talk about systems and emphasises the need for precision on the hypotheses to be tested. Unquestionably, he has a valid point. Stephen Ceci and his colleagues argue forcefully for the necessity of focusing on the "proximal processes" involved in the actualisation of genetic potential. This point, too, is valid and important (although they argue it more persuasively elsewhere) but Hunt is right that it is no good stopping there. How can proximal processes be operationalised and studied? Waldman takes the matter forward to some degree by his discussion of the role of gene-environment correlations and interactions but, again, there is much more that could be said.

It is not just the nurture protagonists who need to show a sea change in their views, however. Despite some lip service to other approaches, some of the geneticists in this volume adopt a one-sided deterministic model. Scarr argues that behaviour genetics is a "theory". It is nothing of the kind. Theories require explanations of processes and that is just what is missing. The findings of quantitative genetics have been tremendously useful in many ways but, in essence, it is a methodology for partitioning population variance. What is singularly lacking (but which will come from molecular genetics) is an understanding of how genes work, of the pathways by which genes lead (usually indirectly) to the trait (such as intelligence) being considered, and of the mechanisms through which nature-nurture interplay occurs.

Neither the chapters by geneticists nor those by psychosocial researchers consider the implications of what is known about such interplay. There is no discussion of the fact that gene-environment correlations and interactions are subsumed under genetic influences despite the fact that environmental mediation may be involved. Equally, no account is taken of the fact that genetic influences on the origins of a risk factor do not mean that the risks are genetically mediated (consider the example of smoking). The chapters by geneticists assume the need for genetic designs but provide no account of how they can be used to test environmental risk mediation hypotheses. Equally, they fail to consider what can be achieved by other designs for testing hypotheses on causal mechanisms. Worse still, they assume that the only causal question concerns individual differences. No attention is paid to effects on level (as the commentary chapters point out), despite their obvious importance. There are numerous dramatic examples of major changes in level over time (the increase in height and in IQ, but also the rise in crime and suicide in young people). These require causal explanations just as much as individual differences.

Another type of proximal process that is neglected is the translating of intellectual ability into academic or occupational achievement. There is well-justified interest in the hypothesised construct of general intelligence but the interest derives from its robust (but only moderate) power as a predictor of real life performance. The book has little to say about the influence of schooling, of motivation, of educational opportunities, or of self-esteem/self-efficacy. Equally, very little is said about the variety of cognitive skills or of the relationships among them. Bouchard, in the only mention of The Bell Curve, notes that virtually all valued behaviours correlate with IQ but fails to consider why or how. As he comments, all the correlations are weak, so the effects are unlikely to be direct. Which proximal processes are involved? The answers are crucial if any use is to be made of the fact that the correlations exist.

A further issue not discussed in the book concerns the connections between normality and disorder. To what extent are the influences on mild mental retardation the same as those on variations in IQ within the normal range? Both Robert Plomin and David Fulker have made important contributions to this topic but they have only a walk-on role here.

It is disappointing, too, that the book has no reasoned discussion of why the genetics of intelligence is a topic that continues to arouse controversy. Much of the furore is based on a misunderstanding of the science but it is necessary to be explicit on why that is so. This means a frank acknowledgement of the dangers of misuse of genetic concepts and an appreciation of the ways in which science is influenced by value judgements. The chapter by Grigorenko and Kornilova addresses this in an interesting manner by taking the case example of Russian psychology (with its political certainty that cognitive development is socially determined). It seems unfortunate that, apart from Plomin's clear statement that there are important ethical issues to be considered, the only genetic parallel discussed is provided by Bouchard's claim that the case against Burt has weakened considerably in recent years. In some respects, perhaps, but anyone who thinks Cyril Burt can be trusted should read Neil Mackintosh's very balanced detailed appraisal of the evidence. Equally, however, anyone who thinks that the elimination of Burt's data destroys the evidence on genetic influences should read the findings that deal only with evidence outside the (rightly) disputed Burt data.

Sternberg and Grigorenko have provided a clearly written, easily understood book that gives a reasonable account of the state of the art in 1993. It remains of interest just because of its balance and clarity but it is difficult to avoid coming away with a sense of disappointment that there is not a better vision of the way ahead. The potential of genetics is very great indeed and the advances in molecular biology have opened up new horizons. The potential, however, will only be actualised if the warring factions come together to meet the serious research challenge of really trying to understand the interplay between nature and nurture. Defending old positions will not get us very far, but taking up new challenges will. The book's success in the former, regrettably, is not quite enough to make up for its failures in the latter.

Sir Michael Rutter is director, Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, London.