AN ARTIFICIAL WILDERNESS Essays on 20th-Century Literature By Sven Birkerts Morrow. 430 pp. $20.95

THERE ARE few books on world literature. This collection of bright and diverse essays by Sven Birkerts is unusual even among them. Its main purpose is surely to treat of unfamiliar writers. Most are German, Russian or French, with the odd Swede (Lars Gustafsson) or Italian (Eugenio Montale) thrown in. There are six essays of a more general nature.

There is, too, an overall purpose. However, since the essays are collected from various magazines, it is hard to say how much this is a convenient afterthought, how much a serious article of faith. No excuse was in fact needed to bring them together. But insofar as Birkerts is serious about his overall thesis, it is a weakness as it stands.

Let us first look at the essays themselves. Many deal with writers who should be better known, but are not because of some freak of chance. For example, the Swiss Robert Walser would probably be more read today had he not been dismissed as a disciple of Kafka. He is, as Birkerts shows, different: more "buoyant." In a few brief words Birkerts distills his chosen subject, and thus leads us towards him. Of course, the method has disadvantages. He has no room to discuss Walser's six novels, and especially Jacob von Gunten, which amounts to much more than "sublime quirks and crotchets." Still, we cannot have everything, and we are led in a similar manner into the work of such as the German Peter Schneider, the Israeli Yaakov Shabtai, and others.

There is much, as is right, to disagree with. For many tastes, Birkerts misses the essential pretentiousness of the French novelist, Michel Tournier. He exposes the lack of substantial answers to the grandiose questions asked in the portentous work of George Steiner, but spoils the exposure by his unironic treatment of this critic as a "superior person," which is perhaps the main impression Steiner seeks to make. Birkerts is frequently more starry-eyed than his alert intelligence will allow. . . . He is on the whole better on unfamiliar names, such as Shabtai, than on familiar ones, such as Marguerite Yourcenar, on whom he is merely conventional. He is better on prose than poetry, about which he is less confident. But for all that he is superb on Montale, from whose works he quotes revealingly. There is much of this sort to be grateful for.

Now to Birkerts' thesis. His quest was triggered by what he (rightly) takes to be the barrenness of most contemporary American fiction, inspired by the "postmodernist" assumption that because "all artistic modes and genres have been exhausted" so innovation is no longer possible, and "art is to be viewed as an arena for ingenuity rather than as an expressive necessity." He thinks his chosen authors "give the lie" to what he calls "the argument from exhaustion." For the first time "ever," the jacket rashly says, "we can begin to speak of a world literature." Clearly this "world literature" is to consist of translations. But there is little or nothing said about the notorious difficulties involved in translation: no discussion even of the view that essence is what eludes translation. One might thus imagine that the verdict of the excellent Clarence Brown, on Mandelstam, is the last word; or that there is no controversy in the matter of English versions of Borges, or that the versions from Montale of G. Singh have not been criticized! Birkerts is continually speaking of style in his authors, as though this offered no problem.

It is not as simple. An essay honestly confronting these difficulties is a glaring lack. In the case of a Mexican novelist not mentioned here, Carlos Fuentes, there has been an open controversy about the matter of translation. There is the notorious case, by no means isolated, of the first version of Pere'z Galdo's' Fortunata and Jacinta. There is as much to be said about versions of works by writers dealt with here. Not some but most poets have had their works so badly damaged that the English versions bear no resemblance to the originals. Literature is not a game. Style is at the very heart of the problem. Of course language must not be a barrier any more than it has to be. But the aim must be to get better and more responsible translations. This means that somebody with a well developed critical sense has, also, to have some working knowledge of the originals: that matter cannot just be skated over.

BIRKERTS IS hostile to "postmodernism." He is right to draw attention to attenuated and "affectless" work. But, make no mistake about it, modernism is exhausted. What I should have liked to hear is more about how the truly new, at a time of crisis, must develop from a reconsideration of the very old. A writer Birkerts deals with well, Borges, has an apprehension of this; but Borges was passionless and, indeed, "affectless." It is significant that Birkerts prefers to deal with Salman Rushdie, but not with Gunther Grass, whom he mentions only twice in passing. Rushdie has been seen in Great Britain not as a modernist (can there be modernists now?) but as a skillful cloner of Latin-American models long out of date. Grass, on the other hand, with roots in medieval literature and in myth, may have gone some way towards a truly postmodernist fiction. Birkerts does not mention the fact that Latin-American fiction has now become to a large extent a parody of itself, and that this applies even to the most recent novels of Garci'a Marque'z. The best Latin-American novelists are no longer the fashionable ones; one needs Spanish to get at them.

Still, that said, this knowledgeable collection leads us to demand a sustained and considered book from its author, perhaps arguing his case for the old modernism reborn. It would be formidable.

Martin Seymour-Smith is the author, most recently, of "The New Guide to Modern World Literature."