While there are strong biological reasons for avoiding sexual relations with close kin, social conditioning has always been a key reason, too.

At the back of the Anglican Book of Common Prayer is a table of Kindred and Affinity "wherein whosoever are related are forbidden by the Church of England to marry together." The table stipulates that a man may not marry his mother, his sister, his daughter, or numerous other relatives, and neither may a woman marry her uncle, nephew, grandparent or grandchild.

These elaborate rules are typical of incest taboos found in most cultures, although taboos usually relate to forbidden sexual relations rather than marriages. At first sight the rules look as though they make good biological sense, because they seem to minimise the genetic risks of inbreeding.

But the list in the Book of Common Prayer continues with some surprising prohibitions. For example, a man may not marry his wife's father's mother or his daughter's son's wife. The mind boggles at the thought of the need for such rules. What was life like in the 16th century if the church felt it necessary to specify such prohibitions? And how many women had the opportunity even to contemplate marrying their deceased granddaughter's husband?

At least six of the 25 relationships that are expressly prohibited from developing into marriage involve no genetic link at all. And yet the Church of England did not prohibit marriages between first cousins. Some other cultures do, but here again the inconsistencies are striking. In some cultures marriages between "parallel" first cousins are forbidden, whereas marriages between "cross" cousins are allowed and even actively encouraged. (A parallel cousin is the child of the father's brother or the mother's sister, whereas a cross cousin is the child of the father's sister or the mother's brother.)

To biologists with an over-simplified view of the evolutionary origins and biological functions of incest taboos, this distinction is puzzling, because in biological terms parallel and cross cousins do not differ. The genetic relatedness is the same in both cases.

An incest taboo is a culturally transmitted prohibition. This is not the same as a natural sexual preference for non-relatives. "Incest", like "rape" and "marriage", is a term that is sometimes mistakenly applied to animals in an attempt, perhaps, to lighten normally dull scientific discourse.

But the taboos are embedded in human institutions with all their associated restrictions, rights and responsibilities. There is no reason to suppose that anything comparable can be found in other animals. However, humans are like many other species in that we are normally disinclined to have sexual relationships with individuals with whom we have grown up from an early age. Such preference for novelty occurs even when children are encouraged by their family to marry a person who is familiar from early life.

Arthur Wolf, an anthropologist at Stanford University, has spent a lifetime studying the arranged marriage systems that used to be found in Taiwan. In some of the marriages, the wife-to-be was adopted into the family of the husband-to-be when she was a small child. These marriages were conspicuously unsuccessful when they were eventually formalised and were much more likely to end in divorce – which itself was deeply frowned upon. The partners simply did not find each other sexually interesting and broke up in the face of strong parental disapproval.

Interestingly, however, the marriages were only unsuccessful if the girl had been adopted into the family of her future husband when she was aged five or less. If she was older, they did not differ from marriages in which the partners met when they were mature.

The propensity to avoid mating with very familiar members of the opposite sex may be seen as a product of biological evolution. In the past, those individuals who learned about the people with whom they were associated in early life and chose mates who were a bit different would be more likely to have healthy children than those who didn't. This would have happened because they would have been less likely to inbreed with close kin.

Superficially, the cultural prohibitions that have been laid down in the incest taboos of churches and states look as though they have been designed to prevent inbreeding even though their evolution and their function are quite different from the unconscious sexual preferences arising from early experience. In our book Design For A Life, we describe one possible solution to the surface similarity between inbreeding avoidance and incest taboos.

Humans typically disapprove of other people behaving in ways in which they would not themselves behave. Left-handers were once forced to adopt the habits of right-handers because the majority right-handers found the minority left-handers' behaviour disturbing. In the same way, individuals who had sex with close kin – something few people would wish to do – were discriminated against.

Most people who had grown up with close kin were not sexually attracted to them, and disapproved when they encountered others who were attracted to theirs. This disapproval had nothing to do with society seeking to avoid being saddled with the half-witted products of inbreeding, because few societies had any idea that inbreeding was the cause. In any event, even if they did know about the ill-effects of inbreeding, having healthy offspring oneself is not the same as encouraging others to have healthy offspring.

We believe that the disapproval was about suppressing abnormal behaviour that could disrupt small, closely-knit societies. Such conformity seems harsh from a current perspective. But in the hunter-gatherer environment in which humans evolved, unity would have been crucial for survival. Wayward behaviour could have destructive consequences for everybody. It is easy to see why social conformity became a powerful trait.

Once in place, the desire for conformity on the one hand, and the unconscious inhibitions against inbreeding on the other, would have combined to generate social disapproval of inbreeding. Much later in human history, this disapproval was codified in elaborate social rules and taboos.

According to this theory, the particular incest taboos of a society should depend on which sorts of people were likely in practice to grow up together.

This explains at least some of the apparently strange proscriptions in the Church of England's marriage rules. Early familiarity also explains why some societies prohibit parallel cousins from marrying but favour marriages between cross cousins. In these cultures, parallel cousins tend to grow up together because in practice brothers tend to live near brothers and sisters near sisters. In contrast, cross cousins tend to be less familiar with each other because brothers and sisters usually live in different places after marriage.

If these ideas are correct, then human incest taboos did not arise historically from a conscious intention to steer others away from the biological risks of inbreeding. Rather, two quite separate mechanisms were at work during human evolution. One was a developmental process designed to strike an optimum balance between inbreeding and outbreeding when choosing an ideal mate. The other was social conformity.

When these two propensities were put together during the evolutionary history of our species, the product was social disapproval of individuals who chose mates from within their close family. Later in human history, more formal rules appeared which could be transmitted from generation to generation – initially by word of mouth and subsequently in written form. And hence a man may not marry his wife's father's mother or his daughter's son's wife.

The general message from this story is that when we seek to understand ourselves, biologists are not going to provide all the answers any more than the social scientists will. The biological processes that underlie each individual's development often have a regularity, but they can generate lives that are quite unique.

Consider the rules that are required for a game of chess. They are clear-cut and nobody is allowed to change them. And then think how complex are the games that arise from such simplicity. The strategies involved provide endless interest and nobody would suppose that you would be able to grind them out from a simple knowledge of the rules.

Incest taboos provide another example from human social behaviour where, to change the metaphor, if you want to admire a cathedral don't look at the shape of the bricks.

Professor Patrick Bateson is Provost of King's College. He and Dr Paul Martin are authors of 'Design For A Life'. The paperback is published by Vintage (£7.99) Up

© 2000 Independent Digital (UK) Ltd.