BOOK REVIEW / Putting earrings on crocodiles: Barbara Everett on Roberto Calasso's vertiginous retelling of the tales of the gods and heroes: The Marriage of Cadmus and Harmony - Roberto Calasso Tr. Tim Parks: Jonathan Cape, pounds 19.99

THE literate eight-year-old used to be handed at infant school a book called Tales of the Gods and Heroes. Roberto Calasso has rewritten it for grown-ups. Already a double prizewinner, widely translated (here with accomplishment) and given sumptuous treatment with a pale terracotta cover stamped with gold, The Marriage of Cadmus and Harmony is considerably more entrancing than the sort of thing we read at eight years old. But it is more difficult too. This perfectly written book will often be admired but may often be left unfinished.

Many familiar stories are told: Zeus and Europa, Theseus and Ariadne, Oedipus and Orestes, sometimes briefly and sometimes in detail and with great bravura. Evidently a formidable scholar, a cultural historian of myth with interests in both philosophy and anthropology (while professionally a publisher), Calasso masters a large field in more ways than one: his book is at home in Greek prehistory, at ease with Homer and the tragedians and the historians, takes in Ovid, and makes its farthest mark 'that great writer' Nonnus, who in the 5th century AD composed a Paraphrase of St John and a Dionysaca, the story of the travels in India of Dionysius.

The Marriage of Cadmus and Harmony includes a final list of references for its quotations; but no index. This is oddly immaterial. Having read through Calasso's book twice, partly from faint bafflement, though with appreciation, I can't imagine ever reading through it steadily again - or consulting it - with purpose: it is a book to pick up and open anywhere, and find unputdownable. It shares this characteristic with those highly individual Renaissance writers such as Montaigne and Burton and Browne, encyclopaedists of a humanist culture they saw as decaying, and tried to garner.

Calasso's book is, in short, not only no compendium of Tales for infants, but no scholarly thesis either. Nor is it a novel, nor any very recognisable kind of romance. Aesthetic it certainly is. One of The Marriage's innumerable godravished women, Leda, makes her way to Delos to give birth, clutching a palm tree: 'Apollo emerged, and everything turned to gold, from top to bottom'. Calasso's book is full of that light which Nabokov called 'aesthetic bliss'.

Cadmus invented writing for the Greeks. The Marriage may be seen as a myth about how works of art are made, and why civilisations rise and fall. Its glittering stories are inset into a commentary which slowly comes to predominate, and which takes on the lineaments of argument: 'We do not have to believe in gods . . . But we do have to believe in the reality of the human conditions and aspirations that are stored in myth'. As the book's more or less chronological account of Greek culture shows it disintegrating, so the narrative tone becomes embittered: 'We have arrived at an age that is neither pagan nor Christian, but that unknowingly continues to practise the same twin gesture of detachment and flight while sinking its claws into both earth and lunar dust.'

The Marriage of Cadmus and Harmony is highly individual and both sustains and justifies its own vision. In Greek myth Calasso sees a poetry that harmonises fragmenting elements of culture ('Only when we become aware of a sudden consistency between incompatibles can we we have crossed the threshold of myth'). The writer's beloved Nonnus is as much a poet of decadence as is Eliot in The Waste Land: both use myth to illuminate ruin. But the myth on which such imagination depends is, Calasso argues at several points, 'vertigo' to handle, always escaping, bifurcating, multiplying and metamorphosing: 'For every myth told there is another, unnameable, that is not told.' In such dissolving territory, the reader can rather easily get lost.

I'm not certain, though, that a reader finds this confusion in the Metamorphoses of Ovid, used by the book as a source. The difficulty lies, rather, in Calasso's own sophisticated understanding of myth, which he interprets in a fashion both his own and highly modern, even modernist.

One of Calasso's characteristically magical sentences runs: 'Achilles is time in its purest state, drumming hooves galloping away.' The book's gods and heroes are white light, timeless gallop and unbloodstained sovereignty. Persephone plucks not just a flower but a narcissus, that Valerian emblem of consciousness: 'She recognised, in the eye observing itself, the eye of an invisible other.' Never without authority, but often preferring variants on accepted interpretation, Calasso makes Helen double, phantom and character, not a woman; his Alcestis is a death goddess, who dies not from love of her human husband but from her need to die.

The second half of the book charts a kind of Fall in Greek culture, a descent into history. Calasso is never less than a fascinating guide, and there are riches here: in particular a wonderful excursus on Sparta, as archetype of vicious utilitarianism in politics; and an account of Egyptian superstition has a wild comic poetry ('They put earrings on crocodiles . . . When there was a fire, the only thing they worried about was saving the cats. And the cats, in turn, threw themselves into the fires'). But it is easy to lose direction in this part of the book, and lost direction is lost interest. The reason is a bewilderingly double system of time - the book's epigraph from Sallust, 'These things never happened, but are always', is not altogether on either the reader's or the writer's side; and an insistence on the definition of what is called 'the fatal nature of reality, its irreversibility'.

A helpful gloss on irreversibility in cultures is provided by Jonathan Bate's Shakespeare and Ovid (Clarendon Press, pounds 35): a welcome short study of a great mythmaker in relation to one of his favourite sources. It reminds us of an important general fact about cultural translation: 'Many important Elizabethan writers combine Ovidian allusions and analogies with a marked sense, usually determined by the Christian tradition, of difference and distance from their classical forebear.' Shakespeare, who had little Greek (and in any case, there is little direct Greek influence on English writing before the 19th century) got much of his knowledge of Greek myths from Ovid, and certainly used translations alongside the Latin, and made such alterations as he pleased in both Latin and translations. It's hard to see how such creative alteration differs from Calasso's vision of the modern search for 'lunar dust'.