Facing protests about his straight-armed salutes last December, the Italian footballer Paolo di Canio declared that he was a fascist but not a racist – a case of political correctness gone mad if ever there was one.

What is behind this label that even a fascist nowadays feels the need to disavow? According to the Oxford English Dictionary, racism is ‘the theory that distinctive human characteristics and abilities are determined by race’.

For this to be a scientific theory, race in humans would have to be accepted as a scientifically meaningful concept; but for decades scientists dismantled it and denied they had any use for it. They passed it over to the administrators and the social scientists, in whose hands it thrived. Now the repressed is returning: race is back on the scientific agenda.

The transfer of responsibility for race was part of the post-war settlement. Many scientists already doubted that racial schemes were much use to them in their work. Now, in the light of how the Nazis had applied racial science, the idea looked far worse than useless. Over the years, arguments from genetics prised the idea of race ever further away from its fastenings in science. In 1972, the geneticist Richard Lewontin published the argument, frequently repeated since, that because 85 per cent of genetic variation can be found within a single population, the use of racial classification in biology could not be justified.
It readily found new justification, though, in identity politics and social administration. Concerns over the equitable treatment of ethnic groups gave race – socially defined – a presence in medicine. A number of drugs have been claimed, controversially, to have different effects on different groups. Some see patterns in the torrents of data surging from modern genetic projects that might be called racial; others seek ways of describing human biological diversity without restoring the framework of race. Marking a change in the wind, in 2003 the Cambridge statistician Anthony Edwards declared Lewontin’s argument to be ‘Lewontin’s fallacy’, because it overlooked patterns of genetic variation.
Suggestions that race might be biologically real after all are music to the ears of those who believe that mental characteristics can be quantified, that the genetic influences upon them are substantial, and that this explains differences not just between individuals but between groups. The core topic in this school of thought, initiated in 1969 by the psychologist Arthur Jensen, is the issue of differences in IQ scores between black and white Americans. Although its theory provokes outrage, its principles and methods are rooted in conventional psychology.

Despite its ally Charles Murray’s complaint of “Orwellian disinformation about innate group differences”, racial claims can now get a sympathetic hearing if suitably presented. Last year, a paper arguing that Ashkenazi Jews underwent selection for intelligence through their historical circumstances in Europe was reported respectfully and without a furore. Apparently it is acceptable to claim that one group is more intelligent than others, but unacceptable to claim that one group is less intelligent than others - even though in the two cases the idea, of racial superiority, is the same.