GENES & BEHAVIOUR > INFORMATION > SHEET 4

Behavioural Genetics: Prospects and Challenges

Thomas Baldwin

1. In 1999 the Nuffield Council on Bioethics established a Working Party to report on ‘Genetics on Human Behaviour: the ethical context’. The composition of the Working Party was seven natural scientists, two social scientists, three lawyers and one philosopher (who is the author of this article). After several fact-finding meetings and a public consultation the report was published in October 2002 . In this article I present some of the main findings and conclusions of the report, but, necessarily, only in a summary and partial form (the report is two hundred A4 pages long, not including some substantial appendices).

The terms of reference for the Working Party began with the following aim:

To define and consider ethical, social and legal issues arising from the study of the genetics of variation within the normal range of behavioural characteristics
The phrase ‘normal range’ which occurs here was intended to be understood in a statistical sense (as the ‘bell’ of the familiar bell-shaped curve). Thus we were supposed not to attend much to the ‘tails’, which, at least at the left end, would often involve a behavioural disorder, since the Nuffield Council had already produced a report on Mental Disorders and Genetics: the ethical context . As will become clear below it was sometimes difficult to stick to this restriction. It is also important to bear in mind that, because the conception of what was normal was to be statistical, ‘abnormal’ behaviour was also to be thought of as only statistically abnormal behaviour, and not as undesirable behaviour.

The report comprises discussions of:

- the historical context for behavioural genetics
- the scientific background for current work
- some current work in the field
- its general implications for our understanding of human life
- potential applications of behavioural genetics
In what follows I shall run through some of these discussions

2. The Historical context

The idea of carrying across to humans some of the methods practised in the context of animal breeding is very old. Plato, writing in Athens in the fifth century BC, discusses it in his dialogue The Republic. As with animal breeding, the motivation for such a practice is typically one of ‘improving’ the people to be created on the assumption that one’s heredity has an important influence on the kind of person one is, in very much the ways now studied within behavioural genetics. In the 19th century this idea was especially promoted by the English scientist Francis Galton who coined the term ‘eugenics’ to characterise it. The ‘eugenic’ programme was taken seriously in the late 19th and early 20th century, especially by those early socialists (e.g. Sidney and Beatrice Webb) who were attracted by the idea of deliberately constructing a better society.

Within the ideas promoted by the eugenic movement it is important to distinguish ‘positive’ eugenics, practices designed to encourage the birth of better people, from ‘negative’ eugenics, practices intended to discourage the birth of ‘inferior’ human beings. Notoriously in the 1930s in Nazi Germany ‘negative eugenics’ led to the compulsory sterilisation of many thousands of those deemed unfit to reproduce (current estimates put the figure at about four hundred thousand). There was never any such practice of compulsory sterilisation in the UK, though the compulsory detention of the ‘feeble minded’ was practised, partly in order to control their capacity for reproduction. But on a smaller scale compulsory sterilisation was practised in North America and some European countries, in some cases up until the 1960s.

This background has irreparably damaged the concept of ‘eugenics’ and, by association, it has also made life difficult for behavioural genetics. The anxiety that behavioural genetics will furnish new reasons for eugenic practices has to be addressed.

3. The Scientific background

I begin with some elementary reminders.

(i) Genetics is a branch of biochemistry, and genes code for proteins, not abilities. Talk of ‘a gene for intelligence’ is a category mistake.

(ii) Insofar as interesting behavioural abilities and capacities depend on genes, they depend on multiple genes and on the environment.

(iii) Genes and environment are connected in many ways; in particular their effects often interact. Thus phenylketonuria (PKU) is a condition caused by a recessive allele of the PAH gene on chromosome 12 and when combined with a normal diet it leads to serious mental retardation. But once the diet of a child with this condition is changed to exclude the amino acid phenylalanine the child develops normally.

One of the central tasks of behavioural genetics is to separate the different contributions of genes and environment to behaviour. Within the study of quantitative genetics the genetic contribution is expressed as the degree of ‘heritability’ of a behavioural trait. This is a measure of the proportion of the variation in the trait within a given population that is attributable to genetic causes (which are usually taken to be additive, i.e. factors which do not interact causally). But a preliminary word of caution needs to be inserted here. Heritability does not measure the chance that an individual within the population has the trait in question on account of his or her genes. For one and the same individual falls within many different populations and the heritability of a phenotypical trait often varies with the population under study. A good example of this is height. Within present European populations the heritability of height is about 0.9 (i.e. about 90% of the variation within such a population is attributable to genetic causes); but we also know that in Europe average height increased by about 1 cm. per decade from 1910 to 1970. Since there is no reason to suppose that there was a widespread genetic change in Europe during this period, it is informally certain that this increase in height is not attributable to genetic factors but to environmental ones, especially diet and childcare, which have changed a good deal. So if one took a diachronic population, such as the children who attended a single school throughout the 20th century, and produced a heritability measure for height within that population it would be much less than 0.9. (This important point applies also to intelligence; I.Q. scores steadily increased in the latter half of the 20th century in a way which is not captured by measures of the heritability of intelligence within a population at any one time).

At the extreme, therefore, if one can fix it that one is dealing with a population who have shared very much the same environment, measures of heritability are likely to be very high. Similarly, if one could fix it that one is dealing with a population who have much the same genes, perhaps because they are clones, measures of heritability of any trait for that population would be very low since heritability measures the proportion of the variation in a trait that is attributable to genetic variation, and in this case there would no significant genetic variation. Thus measures of heritability are always relative to the population studied. In general, however, the larger and more varied the population studied, the more useful the measure is likely to be.

How are heritability measures reached?

There are two main types of study. One type involves studies which compare children within the same family, who are normally genetically related, with other children. By itself, however, such a study will not discriminate between the effects of the shared genes of the siblings and those of their shared family environment. So such studies are sometimes enriched by comparing the similarities between genetic siblings with those between genetic and non-genetic siblings (adopted and step children) in the same family, in the belief that differences here will show the differential effects of different genes in a shared environment. But considerable caution needs to be exercised in assuming that the environment is shared: within a single family different children will have different relationships with their siblings and parents, and these environmental differences will often be exacerbated where adopted and step children are involved. So there is plenty of scope for gene-environment interactions which confuse heritability measures.

The other, and potentially more revealing type of study, involves comparing the similarities between monozygotic (MZ) twins and those between other less closely related pairs (dizygotic (DZ) twins, siblings, and unrelated people). Since MZ twins have the same genes, it follows that where they are much more similar to each other than other pairs who have also shared similar environments, there is reason to attribute these similarities to genetic factors. Typically such twin studies involve twins who have been brought up together in a shared environment; but the aim is to factor out the role of this shared environment by comparing the similarities between MZ twins with those between DZ twins who have also shared the same family environment. Hence insofar as such MZ twins are significantly more similar (e.g. with respect to IQ scores) than DZ twins and other genetically related individuals there is reason to attribute this difference between MZ and DZ twins to genetic factors. A different type of study focuses on genetically related individuals, and especially MZ twins, who have been adopted apart in different families, on the hypothesis that in such cases the MZ twins no longer share the same environment. Such studies are necessarily limited, but where such twins and other genetic relatives show phenotypical similarities of a type not found between comparable unrelated people, there is reason to attribute the similarities to genetic rather than environmental factors.

Quantitative genetic studies of these types were carried out throughout the 20th century; and they typically do show significant measures of heritability within large populations for abilities such as IQ and dispositions to antisocial behaviour. As stressed earlier, however, historic improvements in IQ scores indicate the need for caution here. But from the 1980s onwards behavioural genetics has followed medical genetics in looking for particular genes (or, rather, alleles) which have significant effects on behavioural dispositions. The aim here is to get beyond the large scale comparisons used to generate measures of heritability and to look at individual cases in order to find detailed connections between alleles of particular genes and behavioural dispositions. Again there are broadly two types of method. One type, linkage studies, follows a strategy comparable to that often employed in medical genetics by concentrating on families in which some members display a distinctive behavioural trait. In such cases the research, often starting from some hypotheses about the genes likely to be relevant, will aim to find one or more alleles which are both characteristic of individuals with the distinctive trait and potentially relevant to the causation of the behavioural trait. The other method, association studies, starts from a population in which there is considerable variation in some behavioural phenotype (e.g. IQ scores) and then aims to connect this phenotypical variation with differences in the genotypes of the individuals in the population in order to identity relevant QTLs (Quantitative Trait Loci). As before, those who use this method often start with presumptions as to the genes that are relevant to the phenotype in question; but with modern computational techniques it can also be employed in a relatively presumption-free way by simply making whole-genome comparisons, though there is always a danger here of throwing up irrelevant associations.

4. Some current evidence

Before briefly discussing some recent work in behavioural genetics, I need to return to a point made earlier, that genes code for proteins, and not for behaviours. For this seems to imply that the whole project of ‘behavioural genetics’, particularly in the context of molecular genetics, is a misbegotten enterprise. The way to respond to this is to start from a comparison. Physics does not deal explicitly with tables and chairs; for the design of chairs and tables is a matter of human needs and cultural conventions, not physical science. But it does not follow that physics has no contribution to offer to the understanding and design of chairs and tables (i.e. to ergonomics); for, of course, physics includes information about the materials and structures involved. Similarly, then, behavioural genetics incorporates the hypothesis that the science of genetics will help to understand the ‘materials’ and ‘structures’ of human life, the ‘intermediate phenotypes’ which culturally defined abilities and dispositions make use of. Admittedly in some work the distinction between culturally defined dispositions and intermediate phenotypes is not drawn clearly; but in other work it is, and there seems no great difficulty in principle in making such a distinction.

Intelligence

One area in which this distinction is in fact drawn is the study of intelligence. For although there is a considerable variety of IQ tests which test for different abilities and, arguably, embody different cultural presumptions, there is a good deal of evidence to support Spearman’s hypothesis that there is an underlying general factor, general cognitive ability (‘g’), which systematically affects an individual’s performance in these tests. Quantitative genetics gives a figure of about 0.5 for the heritability of g in contemporary Western societies, although, as mentioned above, the systematic increase in IQ scores since World War II indicates that there are important environmental contributions too.

Can one be more specific about a genetic basis for g? In considering this it is important to bear in mind that the emphasis here is supposed to be on ‘normal’ variation in behavioural dispositions. Thus the well-known connections between some genetic conditions, such as Fragile X syndrome, and severe mental retardation are not to the point. The most important work in this area has been done by Robert Plomin, in which he has conducted a large-scale association study, comparing the genotypes of a group of children of average intelligence with those of another group with higher intelligence. His results so far point to a possible association with the presence of alleles of an insulin growth factor gene (IGF2R) on chromosome 6; but the evidence is mixed, and can only be part of a complicated story involving many genes some of which may well interact. There is no reason to suppose that further work will not clarify the situation further, especially as the structure of the human genome is better understood. But there is every reason to think that the eventual story will be a complicated one.

Antisocial behaviour

Not surprisingly there has been a good deal of work on the issue of the heritability of antisocial behaviour. But it remains a difficult topic to deal with. The point discussed earlier about the inappropriateness of looking for a genetic basis for cultural phenomena applies especially strongly here: criminal behaviour is clearly defined culturally and is therefore not a behavioural type for which one might seek a genetic basis. But even when one abstracts from such obviously cultural phenomena to underlying dispositions such as hostility and aggression, much work tends to focus, not on the genetic basis for variation within the normal range, but on the basis for abnormal behaviour which manifests extreme aggression etc. In fact, it seems clear that one does better psychology when one gets away from simple black-and-white classifications to the study of dispositions which are variably manifested, since it enables one to avoid arbitrary cut-off points and to consider a much wider range of cases; and by putting together a range of measures one can arrive at a complex measure of an individual’s disposition to antisocial conduct.

Research in this area suggests that, for contemporary western societies, the heritability of antisocial dispositions lies in the range of 0.4 - 0.5, which certainly suggests that there should be a genetic basis for them. But as indicated just now, most work then goes into looking for a genetic basis for a disposition for abnormally antisocial behaviour. One obvious candidate for such a basis is just the presence of a Y chromosome; can one do better than this? The most interesting work concerns an allele of a gene on the X chromosome for monoamine oxidase A (MAOA) which regulates serotonin. Study of a Dutch family with a long history of extreme violence by some males has shown that, in this family, the disposition to violent conduct correlates precisely with the presence of this allele in males . The presence of the same allele in members of other families, however, is not correlated with the same disposition; but a recent study has suggested that where there has been a history of child abuse, then this MAOA allele correlates with a disposition to antisocial behaviour (if confirmed, this would be an interesting case of gene-environment interaction).

Sexual orientation

Here too the issue of ‘variation within the normal range’ arises, but in a different way from that considered above. For although one can use the Kinsey scale to try to spread out sexual orientations in such a way that one arrives at a normal distribution, in fact individuals tend to classify themselves either at one end (strongly heterosexual) or the other (strongly homosexual). So in this case the evidence suggests that this is a behavioural disposition which does not display a normal variation.

The similarities between the sexual orientation of MZ and DZ twins have been compared, with a view to estimating the heritability of sexual orientation. These studies show that there is considerably greater similarity between MZ twins than between DZ twins, showing that there is some genetic influence in this area, but the evidence does not enable any precise conclusions to be drawn. In 1993, however, Dean Hamer suggested that homosexuality is associated with the presence of an usual pattern of genetic markers in the Xq28 region of the X chromosome. His evidence was that although one would expect brothers to have only a 50% chance of sharing an X chromosome with an unusual pattern, in fact 82% of homosexual brothers shared the unusual pattern. Subsequent research has weakened the association, but without completely undermining the hypothesis.

5. What does all this tell us about ourselves?

Behavioural genetics tells us that there is a genetic, and thus a physical, basis for many of our basic abilities and dispositions. Does this conclusion undermine established ways of thinking about ourselves?

(i) This conclusion should not be taken to be a challenge to religious faith. Most great religions insist on the physical nature of human personality (thus for orthodox Judaism, Christianity and Islam immortality goes along with the resurrection of the body). So behavioural genetics can be regarded as a way of filling out some of the details here, and not as a challenge to religious conceptions of human life.

(ii) Genetics does not imply that our genes fix our future unalterably: genetics is not an inherently ‘fatalist’ doctrine. Even where genes are deterministic (i.e. do not just increase the chance of some condition but guarantee its presence), it does not follow that we are not able to intervene to prevent some unwanted effect of the condition. Thus, a change of diet removes the bad effects of PKU. Of course there are exceptions, such as Huntington’s disease, where at present nothing can be done to prevent the effect of the genes involved; and perhaps sexual orientation is a case of a comparable behavioural predisposition. But there is no reason to think that this situation applies to most of the the genetic factors which affect variations within the normal range of behavioural dispositions. On the contrary, there is every reason to suppose that in this area ‘environmental’ interventions, such as education, make a great deal of difference.

(iii) Does genetics threaten ‘folk psychology’ - our ordinary conception of ourselves as rational animals, capable of acting for reasons? There is no good reason to think so (indeed it is not easy to understand how behavioural genetics could give us a good reason for thinking that we never act from reasons). Since the capacity to appreciate reasons for action and then to act on their basis is central to ‘free will’ and responsibility, the fact that behavioural genetics does not undermine this capacity shows that the belief that genetics is a ‘deterministic’ science that is incompatible with free will is a mistake. Genetics is not a deterministic science, and, anyway, the kind of free will we want is not an abstract, unconditioned, possibility of detaching ourselves from all physical limitations, but the ability to shape our lives in the light of our own rational deliberations.

(iv) One way to think of one’s genetic endowment is to think of it as providing biochemical ‘start-up’ tools which, in normal circumstances, enable us to develop the set of basic cognitive abilities and affective dispositions which constitute our ‘intermediate phenotype’. This basic kit makes it possible for us to be a ‘thinking subject’ capable of relating to other people; and equally to develop, refine and alter these very abilities and dispositions. Our genes equip us with the means to become rational animals; but they do not dictate how we exercise this rationality.

6. Potential applications?

The applications of behavioural genetics appear to be far off at present; but we need to start thinking about them now in order to address the anxiety that they would permit new and objectionable eugenic practices, albeit without state compulsion. Barbara Katz Rothman has put this point well:

“The scientists quickly speak up: that isn’t possible, they reassure us, you don’t understand the genetics involved. Five years later, of course, that is possible, and then it is too late to decide whether or not to do it: we wake up to find it done.”

So: we should imagine that behavioural genetics has advanced to the stage of providing genetic tests which indicate, with a reasonable degree of reliability, the chance that the person tested has a higher (or lower) than average degree of some significant ability or disposition; and that these tests can, in some cases at least, be associated with interventions (genetic, medical, environmental) which are intended to alter the abilities and dispositions in question. What are the issues to think about?

Medicalisation and stigma

There is the danger of the ‘medicalisation’ of behavioural dispositions that are regarded, for good or bad reasons, as undesirable. This danger is already manifest in the promotion of drugs to help people deal with shyness and in the excessive use of the drug Ritalin which is intended to help treat the genuine condition ADHD (Attention Deficit Hyperactivity Disorder). The anxiety therefore is that, with genetic tests for behavioural predispositions and the possibility of subsequent interventions to alter them, conditions that are in fact part of normal life will be regarded as quasi-medical disorders requiring treatment.
Similarly there is a possibility that those who are found to have a genetic predisposition to a disposition that some people judge to be inferior will be stigmatised as a result. This possibility was manifest in the reaction of some people to Hamer’s work on sexual orientation, though others found that this work confirmed their view that sexual orientation is fundamentally a matter of one’s ‘nature’ and not a matter for praise or blame.

Inequality
The possibilities for the enhancement of significant abilities give rise to important egalitarian anxieties. Suppose it becomes possible to enhance one’s general cognitive ability (or memory or some other useful ability) but only after expensive tests and interventions; there is then a danger that this opportunity for enhancement will be socially divisive unless the state intervenes to ensure that these tests and interventions are available to anyone. It is difficult, in advance and in the abstract, to assess the danger here, as compared, say, to the divisiveness which arises from private education. But the issue needs to be monitored.

‘Designer babies’

The most commonly debated issue concerns the use of pre-implantation genetic diagnosis (PGD) in conjunction with (we are here supposing) reliable genetic tests for behavioural dispositions to select embryos after in-vitro fertilisation (IVF) for implantation. This is often described as the potential for creating ‘designer babies’, though the term is misleading. There is a widespread consensus that it would unacceptable to terminate a healthy pregnancy merely because such tests indicate that there is a significant chance that the resulting child would have behavioural dispositions that somehow disappoint the parents although they fall within the normal range. But since PGD does not involve the termination of an established pregnancy but only the selection of which embryo to implant, the issues are different.

Those in favour of permitting PGD in this situation maintain that prospective parents have a ‘right to procreative autonomy’, such that it is not the state’s proper business to intervene to obstruct parental choice in this area. It is argued that this is as much a sphere of private life as is the matter of sexual orientation. But those opposed to PGD argue that the consequences of the exercise of the alleged right to procreative autonomy are not confined to the private sphere and threaten serious harm to others. It would reinforce prejudices against those with lower than average abilities and contribute to social divisions and inequalities. More fundamentally, it is argued, to use PGD in this way threatens to alter radically, and for the worse, the relationship between parents and their children. The proper attitude of parents to their children includes a loving acceptance of them as they are, as ‘a gift of God’, rather than a determination to use whatever means are available to ensure that they fulfil parental expectations. Hence the sphere of procreation is one in which we should curb our ‘will to power’ and learn to practice instead the virtue of humility.

The Nuffield Working Party was persuaded by these latter arguments to oppose PGD where it concerns behavioural traits that lie within the normal range. But just as the arguments against PGD are unpersuasive when one considers its use to prevent the birth of children with serious medical disorders, they are also less persuasive when one considers its use to prevent the birth of children with harmful behavioural traits that fall outside the normal range, for example in the case of the Dutch family afflicted by the MAOA allele.

Legal applications

As discussed earlier, the existence of a genetic basis for some behavioural dispositions does not of itself undermine one’s responsibility for one’s actions; and even where there is evidence of a strong behavioural predisposition to antisocial behaviour which falls outside the normal range, the law tends to start from a presumption of responsibility when assessing criminal liability. But when it comes to sentencing those who have been found to be guilty, a wider range of non-excusing, but mitigating, environmental considerations are already taken into account: hence reliable genetic evidence could easily be included here as a further mitigating factor. A further question concerns the use of such evidence to support preventive detention, in advance of the commission of any crime. The Working Party took the view that this would be an unacceptable infringement of individual liberty. It might be possible to use such data as a basis for interventions that are intended to be helpful. But the consent of those affected would still be required and the serious danger of stigmatisation attendant upon such intervention would need to be considered.