BELIEF in witchcraft collapsed among educated people in the early 18th century, but among the uneducated was still going strong at the end of the 19th century. Thomas Hardy's "conjurors" are one sign of this another is the favourite story by C.G. Jung about a Swiss magician weaving spells in a country meadow a few yards from where the Geneva Zurich express thundered by.

The great historical irony is that, just as even the unlettered abandoned hope in miracles from their "white magicians", a new literary coven of so called experts in the occult - principally Dennis Wheatley and Aleister Crowley - was instilling the now prevalent idea that witchcraft always involved diabolism and the worship of Satan.

James Sharpe is a serious, no nonsense historian and he has a brisk way with Wheatley's The Devil Rides Out, the monks of Cefalu and all the rest of the 20th century nonsense. Nor is he much more sympathetic to interpretations of witchcraft via Freud and the theory of hysteria.

He only allows three semi serious candidates on to the shortlist of plausible theory: Margaret Murray's idea that witches were adherents of a pre Christian fertility cult; the feminist notion that "witches" were really proto liberated women who fell foul of patriarchy and misogyny; and the argument that the victims of witchcraft trials were the losers in a game of economic and demographic pressures.

The first two interpretations Sharpe dismisses on grounds of scholarship in the case of the feminist theory, woefully inadequate scholarship, be it said. Only the last idea is taken seriously. By sifting through court records, county archives and trial proceedings, Sharpe shows that in the two centuries when witchcraft was a live historical issue the witch hunt was a form of social control operating to curb deviancy at village and hamlet level.

Women who did not behave in socially approved ways ran the risk of being tarred with the witch's brush. If they acted in any way "contrary to nature" - by being mannish, shrewish, disputatious, or lacking in deference for older women in the extended family - other actions that were "contrary to nature were laid at their door.

If a cow ceased to give milk, and there was a "difficult" woman in the village, the two phenomena were causally related, and the woman was accused of causing maleficium or harm, through spells, potions, the evil eye etc.

Meticulous and fair minded throughout, Sharpe admits that the witch trials in Norfolk in 1645-47, presided over by the notorious "witchfinder general", Matthew Hopkins, work against his thesis and in favour of the Wheatley/Crowley idea, for in these trials women admitted to sexual intercourse with the Devil.

Perhaps this was Hopkins importing alien ideas about witches from the Continent? Not so, says Sharpe. Although he is at pains to point out that he is writing about English witchcraft, not Scots or Irish - because of the sources he is using - he concedes that there was no significant difference between the varieties in these two centuries.

No, the true reason is likely to be the general feeling that the English Civil War had ushered in an age of disorder. Just as theologians who take the Devil seriously think it much more likely that he manifested himself at Auschwitz rather than in some Exorcist style single possession, so the Devil was thought to have been conjured up by a Europe convulsed by warfare in the 1640s rather than the individual action of a Norfolk witch.

This sort of insight is typical of a book rich in associations - erudite, wise, probing yet deeply human. It takes scholarship about witchcraft on to a new level.