The star system has changed since the heyday of Hollywood, but we still have stars, who carry the shades of their past films with them whenever they appear. Sex is clearly a major component of their prestige: they serve their audience, at least to some degree, as models of sexual behaviour or objects of desire. And their actual fame and sexual appeal make them more convincing in powerful roles. It is partly because Sean Connery used to be James Bond that we can believe in him as King Arthur.
On the other hand, a star's accumulated prestige can easily come into conflict with roles where he, or she, is not required to function as a sex object, and in particular with historical ones which reflect attitudes to love and sex different from our own. The makers of First Knight have opted for a medieval rather than a Dark Age setting - the architecture is Disneyland Gothic - but their text, the story of Lancelot and Guinevere is one of the defining narratives of courtly love. Someone might have reminded them of this before they decided to cast Richard Gere as the epitome of knightly virtue, and make him into a lone soldier of fortune who rescues Guinevere (Julia Ormond) from an ambush, engages in some urgent mouth-to-mouth resuscitation behind a tree and says: "I know you want me, I can see it in your eyes."
Failing that, they could have watched Robert Bresson's extraordinary Lancelot du Lac (1974, available on the Artificial Eye video label), which uses unknown actors and a 20th-century Catholic's sense of sin to create an austere and compelling vision of past times. The only ideas motivating First Knight are a nostalgic notion of "Camelot" which seems to have something to do with the musical, a concept of political justice which would have been equally meaningless to people in the Dark Ages and in medieval Christian Europe, and some cod psychology (we are asked to believe that the child Lancelot was deeply scarred by the violent death of his parents - at a time when, on the evidence of the film itself, pillage and murder were commonplace). Agreed, it is pedantic to fuss about such things when the sole aim is to entertain; but why bother with an historical setting, even a highly fictionalised one, only to treat the past as if it were the present using different clothes and technologies?
No one would accuse the team of Ismail Merchant, James Ivory and Ruth Prawer Jhabvala of lacking in reverence towards the past, but their latest effort has been coolly received. It is based on the experiences of Thomas Jefferson as ambassador of the United States to France on the eve of the Revolution. Against this background of conflicting political ideologies, it explores his relationship with two women: the married artist, Maria Cosway and the slave girl, Sally Hemings. If nothing else, it does acknowledge that the sexual manners of the 18th century were not those of our own time.
The affair with Sally Hemings is strongly denied by some historians, which explains part of the controversy: admirers of Jefferson cannot be sure that the audience will adopt a critical approach and distinguish between screen speculation and established historical fact. But there are other reasons why critics love to disparage Merchant-Ivory productions. Like the "well-made play" or the cinema de qualite in the 1950s, their work is open to the charge of bloodlessness - sacrificing energy to pretty costumes and glossy sets, to make a theme park of the past. In the case of Jefferson in Paris, though there are some glaring false notes, I think these criticisms are misplaced.
The film's greatest asset is Nick Nolte's Jefferson, the result of a palpable effort to project the personality of a man, a model of Puritan values, for whom the institution of slavery raises genuine dilemmas. This is not an attempt to justify an attitude which nowadays everybody finds unacceptable, but to show how a good man could adopt it. The result is not an audience-pleasing performance, but a central contribution to the film's attempt to penetrate Jefferson's world and to see his sexual relationships as determined by that, rather than by desire - and a promising pair of eyes.
The task of D'Artagnan's Daughter is easier, because Bertrand Tavernier's film is self-consciously a fiction derived from a fiction. It has little ambition except to be an enjoyable romp, giving its stars, Philippe Noiret and Sophie Marceau (as, respectively, Dumas' aging hero and his lively daughter), the opportunity to play themselves, in costume, and brandish a sword. But it is addressed to a far more knowledgeable and sophisticated audience than either First Knight or Jefferson (which has Dr Guillotin naively demonstrating a machine he has just invented for cutting off heads): expecting to raise a chuckle with a joke about Mazarin remembering, too late, that he has not advised Louis XIV to revoke the Edict of Nantes, argues a certain faith in the educational system that neither of the English-language films could assume.
This is one reason why, on the whole, the French enjoy historical films more than we do, and make them better.
A study guide to Jefferson in Paris is available from Film Education, 41-42 Berners St, London W1P 3AA.