BOOK REVIEW / Loose thoughts in the flirtation chamber: 'On Flirtation' - Adam Phillips: Faber, 14.99 pounds

IT SAYS it is on flirtation, and its subject matter is psychoanalysis, with some excursions into literary criticism at the end. You anticipate the denouement - that psychoanalysis itself is flirtation; a means of endless play and deferment. It doesn't happen. Adam Phillips writes as he says he does, as a flirt, and so the lack of conclusion is a formal requirement. But his identity as a writer always defers to that of analyst: respect, rather than flirtatiousness, seems to be the reason he won't go that far. 'Flirtation, of course, stops when you take it seriously,' he remarks in the title essay.

Up to that point it is, as he also points out, a form of scepticism. So for the reader on the outside of the intricate networks that make up psychoanalytic thought, this collection of essays can serve to explore the question of how far it is possible, and worthwhile, to flirt with psychoanalysis. If flirtation can be 'sado-masochism with a light touch', does that make it bogus? Do we have to have heavy-duty relationships, or none at all?

Underlying these questions is one which wonders how open psychoanalysis is as a system of knowledge. Phillips's intriguing discussion of contingency and parapraxes - Freudian slips - illustrates Freud's urge to erase chance. Mistakes, deviations from orderly behaviour, don't just happen; they signify something. The quantity of random occurrence in mental life must be minimised; the explanatory power of the analytic method must be extended to all corners of the psychic terrain. There is a similarity here to the hard adaptationist school in Darwinian theory, which recoils from the idea that sometimes things just happen, insisting that there has to be a purpose to all evolutionary change. Another theoretical defence, Phillips quotes Bernard Williams as observing, is morality, which we hope will make us 'immune from luck'.

In outlining the possibility of what he calls a 'contingent self', one that acknowledges accidents, Phillips helps open up psychoanalysis and render it credible to outsiders. His dismissive comments on various standard criticisms of the discipline are a reminder of how stale the interaction has become. Psychoanalysts know all the arguments, about Freud's authoritarianism and so on, and are blase. Meanwhile, psychoanalysis gradually loses its influence in the therapeutic and popular domains, ceding ground to apple-pie theories of the self.

The justification is that psychoanalysis may be magnificent, but it isn't therapy. On Flirtation does not address the issue of whether psychoanalysis is any practical use to the psychiatrist faced with a patient in need of treatment. It does, however, illustrate a more general point. We need Freudian texts like this one because, as Maureen Freely said of its predecessor, On Kissing, Tickling and Being Bored, it 'does all the things that fiction is meant to do and usually doesn't'. The Freudian text rubs the reader's nose in the intractability of real life. And it doesn't allow the reader to collude in the fiction of the single, simple self. In analysis itself, 'the patient has to refuse himself the conventional satisfactions of narrative. Abrogating his need for beginnings, middles and ends, he often has to become a very bad story-teller and make a nonsense of his life.'

This makes it an impossible consciousness to live by, except for those who make a profession out of it. As a text, it is difficult: Phillips's style is weightless by the standards of the trade, but recondite by anybody else's. A flirtation with it thus departs with the aphorisms and bon mots that glisten on the tightly knit sinews of argument; mementos, not engagement rings.