IT TAKES 30 years for an idea to cross the Atlantic. Roland Barthes (1915-1980), French semiologist, structuralist, and epicurean, published his first book, Writing Degree Zero, in 1953. By the late '60s Nathaniel Tarn, the poet-anthropologist-semiologist, at that time editor of the elegant little paperback series of monographs called Cape Editions, had Writing Degree Zero translated (by Annette Lavers and Colin Smith) for the series (1967). Rodney Needham, professor of anthropology at Oxford, colleague and translator of Claude L,evi-Strauss (Barthes' master in the science of signs and symbolic structures), sent me a copy, characteristically without comment. I remember reading the thin little book while giving an exam at semester's end. I find that I've underlined passages on every page, for Barthes' prose is terse, epigrammatic, and seductive.

"Whatever its sophistication, style has something crude about it: it is a form with no clear destination, the product of a thrust, not an intention, and, as it were, a vertical and lonely dimension of thought." I underlined this, and can still feel the wonder of saying under my breath, "Thrust? Intention? Vertical? Lonely?" The first response a reader of Barthes is likely to have is that here is a writer who sees the world in a new way, and who writes with precision about things one did not think would be, or could be, talked about at all. Such as the manners of a racing cyclist, the amount of sugar in American food, fruit compotes in the utopia of Charles Fourier, the euphoria of the preterite tense.

And now, 29 years (things are speeding up) after Barthes began to publish, we have A Barthes Reader, compiled and introduced by Susan Sontag, the first attempt to integrate his diverse talents and concerns, to demonstrate the unity of his thought, and to place him in his tradition in French writing.

Sontag's introductory essay (and it is a full-scale, rigorous essay, with a title, "Writing Itself: On Roland Barthes," with no trace of the perfunctory gestures of a run-of-the-mill introduction) is charmingly Barthesque. She extricates Barthes from what, on this side of the Atlantic, is apt to seem a confusion of French intellectuals subtly different from each other, and all quarreling with each other the whole length of the Boulevard St.-Germain. Sontag constructs a lucid picture of Barthes. She shows us the components of his complex sensibility, the cleverness of his imagination, his evasions and intellectual passions. (Of the man himself, tubercular, transcendentally neurotic, opium addict, pederast, shy, reclusive, she says nothing at all, except that he was a brilliant teacher.)

Structuralism is a discipline almost as old as the century. It began as a method for analyzing unwritten languages (such as the Siberian ones), and developed in the hands of certain Russian critics and French sociologists as a way of finding patterns and coherences in such ungraspable complexities as the idea of kinship, systems of symbols (how we know to name farm animals, wild animals, and pets, for instance: no one ever taught us, and yet we know), why the lower classes prefer sweet chocolate and strong perfume and the upper classes bitter chocolate and subtle perfume.

The progenitor of structural anthropology, L,evi- Strauss (Barthes' colleague at the CollMege de France) has been contemptuous of non-anthropological structuralists ever since the science became a vogue among literary critics and pundits of every kidney. It is a joke on American campuses that the professors who chatter structuralism can never be pinned down as to what in the name of God they're talking about. The Times Literary Supplement, some 10 years ago, invented an imaginary structuralist, a mad Balkan whose idiotic works they regularly review. The parallel with the fate of Freudian analysis is obvious.

It is Susan Sontag's observation that Barthes' standing as a structuralist or semiologist is beside the point. He was a writer who contributed to these disciplines (I'd say he's the man who taught us vernacular structuralism, legitimately derived from the mandarin dialect of L,evi-Strauss), but most of all he was a writer. He belongs to the tradition of Montaigne and Voltaire, who wrote about any and everything with wit and intelligence. The fun of reading Barthes is precisely his wickedly imaginative insight, always a surprise, often infuriating, sometimes sterile (who cares that the Eiffel Tower "returns Paris to nature"?), but a gratifying number of times wonderfully fertile. Barthes is at his best talking about language and prose, cooking and games, photographs and spectacles. He can be maddeningly tedious analyzing every phrase in a story by Balzac (S/Z), and surprisingly tedious on the subject of his own self.

This admirable anthology (with which my only quibble is that it doesn't have all of Writing Degree Zero or the fine essay on bicycle racing) contains Barthes' masterpiece, the study of Fourier (from his book about de Sade, Fourier, and Loyola).

The art of explaining a large subject is a skill all to itself, and among the masters of it nowadays, there's Hugh Kenner without peer. Next, Northrop Frye. Next, Barthes. Kenner on Beckett, Frye on the nature of comedy in The Anatomy of Criticism, and Barthes on Fourier are the century's masterpieces of critical exposition. The Fourier essay is flanked by its negative image, so to speak: an analysis of the revolting Marquis de Sade. Barthes needs him and his monstrosities to show the beautiful paradox of de Sade's erotic hoe-down turning out to be an insane display of French logic. Fourier, on the other hand -- the Henri Rousseau of sociology -- planned a logical solution to all problems that turns out to be a delightful example of French charm.

The Fourier that Barthes writes about so beautifully and with such smiling pleasure is a recent discovery. A French publisher brought out all of Fourier in a new edition in eight volumes (1966-67), to include (volume 6) the unpublished "Nouveaux monde amoureux" that had lain lost in the BibliotMeque Nationale for a century and a half. This event played into Barthes' capable hands as significantly as L,evi-Strauss' writing had before. Barthes' genius is to see what others don't, to make the astoundingly clever observation that startles us with its obliqueness.

In L,evi-Strauss he found a master for the kind of intelligence he was cultivating, and in Fourier he found a companion. What Fourier added to Barthes' sensiblities was courage and charm. Structuralism itself has the charm of mathematics, and the humor in L,evi-Strauss (his description of India in Tristes Tropiques, for example) is unintended. Barthes has, in a sense, humanized structuralism, as he has, in another sense, enriched humanism (so bleak and full of duties and commitments in Sartre's hands) with an attractive renaissance of both Epicurean delight and an Erasmian critique of modern follies. Susan Sontag is right to place him among the French essayists for whom observation was a keen sport, introspection a spiritual exercise, and writing itself one of the greatest of civilized pleasures. There are several Roland Barthes -- the semiologist being thoroughly technical about signs and signifiers, the anatomist of popular culture, the dilettante, the critic, the philosopher. This welcome anthology gives us the opportunity to see them all, and to understand how they cooperate.