Ethnic minorities find the UK film and TV industry shut to them, says Sara Wajid, but Geoff Watts reports on an experiment proving that racism is not hard wired.

Stripped of its layers of ham, schmaltz and self-aggrandisement, the remaining few per cent of Halle Berry's recent Oscar award speech lifted the lid on race and the Hollywood establishment. Matters may be improving, she implied, but they have been and still are pretty bad.

Seen from this side of the Pond, Berry's choking observations about the hurdles facing "every nameless, faceless woman of colour" ring familiar bells. From the policing of London and the guarding of Buckingham Palace to the management of television and the governing of the country, racial inequality is a fact of everyday life. Change is not easy; but that does not mean it is necessarily as difficult as we tend to assume.

The ubiquity of the inequalities, and their apparent depth, has led many to suspect that discrimination on grounds of race is so deep seated as to be virtually ineradicable. Robert Kurzban, John Tooby and Leda Cosmides - researchers at the University of California with backgrounds in psychology and anthropology - dispute this view. Their case rests not only on theory but, unusually, on experimental evidence.

ÖйúAƬ

Conflict between groups - "the categorisation of the social world into us versus them", as the three authors put it - is found in almost all cultures. It can be reproduced under laboratory conditions, even when the experimenters randomly allocate their subjects to groups defined by trivial criteria of no social significance. In the real world, anyone who doubts the human taste for group loyalties that defy reason should think of football. Clubs in distant and sometimes never-visited cities, staffed by foreign players and run solely for the benefit of commercial enterprises indifferent to the interests of fans can still command a sometimes violent allegiance. Group loyalty comes naturally.

Race is self-evidently one of the characteristics used to define such loyalty. Indeed, along with sex and age, it is hard to avoid the suspicion that an instant tendency to identify and note someone's race must be programmed into the brain. Kurzban and colleagues point out that even if you recall nothing else about a new neighbour, you will probably know that she is young, black and female rather than, say, old, white and male. The evidence for this comes not only from everyday experience but from empirical studies. And as categorising people into groups, irrespective of criteria, almost invariably leads to discrimination, the suggestion that race might be a core recognition feature is depressing indeed.

ÖйúAƬ

This suggestion is precisely what the Californian researchers dispute. As evolutionary psychologists, they believe that any account of human thought and behaviour must reflect the long-gone hunter-gatherer circumstances under which our brains evolved and to which they are most likely adapted. Survival in those days would have depended on detecting and responding to a perpetually changing set of coalitions between individuals and bands. In managing these relationships, inferences drawn from observing the sex and age of others would have been valuable. Would noting the race of other individuals have been similarly helpful?

Hunter-gatherer bands moving on foot were hardly in a position to travel the world. Most would seldom, if ever, have encountered people from other racially distinct populations. Consequently, there would have been no selective pressure to develop brain mechanisms specifically attuned to spotting and noting racial differences. Although these differences are used nowadays to make distinctions that then become the basis of group loyalties, evolution could hardly have endowed the brain with a specific predisposition to look for them.

A far more likely outcome of evolution would have been a general mechanism for detecting coalitions and alliances. But patterns of coordinated action, cooperation and competition between individuals are often short-lived and leave no outward and visible trace. So another requirement would be an aptitude for identifying visible markers by which to spot membership of such coalitions. These might include dress, dialect, manner and other aspects of appearance and behaviour of the kind recorded by generations of anthropologists. Skin colour, of course, is one such marker. But that is all it is - a marker, not the marker, and of neither greater nor lesser significance than many others.

To demonstrate the point, Kurzban and colleagues used a "memory confusion protocol": an established laboratory method of relying on errors of recall to find out if subjects are subconsciously categorising individuals into groups and, if so, by what criteria. Their experiments involved showing subjects a succession of 24 photographs of eight members of two rival basketball teams, both made up of equal numbers of black and white male players.

In some tests, all the players were dressed identically; in others they wore shirts in team colours. In another set, the black men were replaced by white women. Each picture was accompanied by a short sentence. Taken together, these sentences added up to a group conversation telling the story of a dispute that had taken place between the two teams.

ÖйúAƬ

After a brief pause, subjects had to start matching the quotes to the photographs: a demanding task in which most people make many errors. It is these errors that allow the experimenters to draw their inferences. The basis of the tests is that subjects are more likely to confuse individuals whom they have mentally encoded as members of the same category than those they have pigeonholed as members of different categories. The object is to find out to what extent race acts as a core pigeonholing system.

The conclusions were clear: although sex was, as expected, a core feature for categorising team members, race was not. "We were surprised at how fast we got pretty drastic reductions in race encoding," Tooby says. Far from being hard wired into the brain, racial awareness is more like "a contagion that's passed back and forth".

When a more relevant method of categorising people becomes available - in the experiments this was a knowledge of team membership as indicated by the shirts - categorisation by race withers rapidly. "If you can get rid of the short-term feedback cycle," he says, "racism is not something that has to be there for ever."

ÖйúAƬ

Kurzban agrees that moving from laboratory experiments to a general conclusion about racial awareness is a big step. But he is confident that the findings are sufficiently robust to stand replication. Ideally he would like to see this done in one of America's southern states, where racial discord is more deeply entrenched than in California.

Kurzban and Tooby agree that if the findings stand up to scrutiny, they will have implications for future social policy. They are, however, understandably reluctant to speculate on what these might be. It does seem clear that purposely mixing the racial make-up of groups working together with a common purpose should minimise mutual suspicion. As a strategy, this is hardly new.

What the Californian findings offer is encouragement. Policies that foster racial mixing may confront not a locked door but one that will open with a surprisingly gentle shove. Even if some sins are original, racism is not among them. Like the thespian histrionics of an Oscar recipient, it has to be learnt.

Evolutionary psychology is often accused of providing a biological justification for all sorts of undesirable behaviours from aggression to adultery. Maybe so. But in this case at least, it seems to be on the side of the angels.

ÖйúAƬ