How Fiction Works, By James Wood

An Olympian critic points out where major-league talents are getting it wrong

Whatever one may think about James Wood's constant ejaculations, his ceremonious name-dropping ("W G Sebald once said to me...") and his lecture-hall mannerisms – more of these in a moment – he really is an A-grade exponent of what university syllabi used to call "practical criticism". Some of the best bits of this brief but luminous primer – and they are very good indeed – come when Wood strips the engine of some fabled fictional juggernaut down to its component parts with the aim of establishing just how a piece of prose works to bring off its effects, the way in which, as he puts it, a novel "teaches us how to read its narrator".

The mark of Wood's dexterity, as chapters on Detail, Character, Language and Dialogue rapidly succeed one another, all broken up into a series of daintily numbered paragraphs, is his ability to extract useful lessons from the sight of a major-league talent getting it wrong. A closely inspected passage from Henry James's What Maisie Knew, in which he demonstrates how the authorial gaze oscillates between Maisie, the wider community in which Maisie operates and the writer himself, is made easier by James's mastery of the point of view. Even better, consequently, is Wood's dissection of an extract from John Updike's Terrorist, in which he convicts Updike of stuffing in so much character-forming detail that he ends up playing his creation false.

Supposedly intended for "writers, readers and anyone interested in what happens on the page", How Fiction Works belongs to the upper shelf of what is turning into a very considerable critical sub-genre: the literary user's manual. Professor John Sutherland got in first with his The Novel: A user's guide (2006), closely followed by Professor John Mullan's How Novels Work. Verse, meanwhile, was covered by Ruth Padel's Fifty-Two Ways to Read a Poem.

Naturally, the levels vary. Sutherland was chatty, informal and commercially attuned; Mullan more interested in the nuts and bolts side of literary composition. Wood, on the other hand, is positively Olympian, taking Flaubert, Nabokov, Joyce and other heroes of the modernist canon out into the book-lined study of his mind and hob-nobbing with them in the most companionable manner.

Style-wise, all this realises a quaint academicism not seen in English letters for the best part of a century. No one minds enthusiasm, of course, but when Wood admires a particular sentence he crows over it: "What a piece of writing this is!" (What Maisie Knew); "What an amazing opening!" (a Chekhov story); "How fine this is!" (Marilynne Robinson).

There are the intent prolepses ("I return to the question of artifice and lifelikeness in 'Truth, Convention, Realism', paragraphs 111-122"), the habit of pronouncing taxonomic judgment on absolutely everyone mentioned (poor old A C Benson, dragged into a foot-note, can't just be "the writer A C Benson", he has to be "the minor English writer A C Benson"), above all that de haut en bas Cambridge combination room use of the adjective "little". Wood apparently got his inspiration from Ruskin's "little book" The Elements of Drawing. His hero Henry Green once gave a "little talk" at the BBC. Why, Wood – modesty breaking out all over the place – admits to writing a "little book" himself.

The foregoing is merely a bad case of literary manners, of course, which one can take or leave while continuing to relish the verve of Wood's insights. Where Wood sometimes falters is in his progress from practical criticism to the mighty conclusion. Rather like Adam Thirlwell in Miss Herbert, he has a trick of coining grand aphoristic obiter dicta in a way that suggests no one had ever thought of the idea before. "Metaphor is analogous to fiction because it floats a rival reality." "The novel is the great virtuoso of exceptionalism: it always wriggles out of the rules thrown around it."

It's a pity, too, that the rigour brought to Updike can't be applied to some of the other titans of contemporary OK-ness. When Saul Bellow, for example, describes a river as "crimped, green, blackish, glassy", he is not so much describing a river as drawing attention to how good he is, contributing, in fact, to one of the great deceptions of modern literature, which is to offer to the reader not a plausible narrative but the less allowable spectacle of the writer writing.

The result is a mixed bag: full of top-notch observations from the coal-face, but weighed down by its self-consciousness. The publishers are marketing this book as a 21st-century update of Forster's Aspects of the Novel (1927). No harm in that, but the fustiness of some of the critical language ("marvellous alchemical translation" etc) occasionally makes it seem closer in spirit to some of the productions of late-Victorian belles lettres – Henley's Views and Reviews, say.