An aimless wander on fertile ground fails to sort Mills and Boon romancer from philosopher

Geography, as every schoolboy knows, is about maps, but biography is about chaps. All the same, a certain amount of mapping is desirable in a biography, to help us find our way about the chap in question. Just as a road map abstracts from the limitless detail of the landscape those features that enable travellers to navigate, so a good biography makes the pullulating raw data of its subject's life graspable by means of the structure within which it presents them.

By this standard, Peter Conradi's biography of Iris Murdoch, for all its industrious merits, is not altogether successful. His oddly segmented, discontinuous account often reads more like a hastily tidied-up archive than a finished interpretation. The archival work is humblingly impressive: Conradi must have laboured long and hard over the years to amass this material. But he might have allowed himself more time to digest his findings, to abbreviate them savagely, and to turn them into a properly geographical "life".

The heart of Conradi's failure is that we leave his book feeling not properly acquain-ted with his subject. He does not really answer the main questions that we hope to find confronted. What was the core of Murdoch's personality? What was she like to be with? Why does he believe she is an important novelist rather than, as some prefer, the thinking person's Mills and Boon romancer? Was she really a considerable philosopher, and if so why?

Conradi provides a plethora of detailed new information, much of it relevant to these issues, some of it wonderfully diverting anecdote. This will delight and inform many readers, especially those with a previous interest in Murdoch and her worlds; but, again, there is not enough of an explanatory framework to fix the salient themes in our mind - lots of gaudy beads, but no string strong enough to thread them on. Perhaps Conradi knew her too well, and too uncritically?

Some of the excellent vignettes of the supporting cast, especially the men in her life - Frank Thompson, Franz Steiner, Elias Canetti, Donald Mackinnon, her husband John Bayley - leave us feeling better informed about these characters than about Murdoch, who remains opaque, inaccessible, out of focus. To the extent that we do form a picture of her, she seems, despite what we are told was a formidable intelligence, to be by turns muddled, mildly portentous, obscure, naive, stylistically unreliable (like Conradi), particularly in her later work, without edge as a philosopher, detached from everyday reality and inhabiting an over-intellectualised imaginative world of neurotic (and erotic) religiosity. Maybe the task of revealing her essence is hopeless. "There simply isn't very much to be said about your ideas," said the logician Georg Kreisel. "And your point is?", one asks, of Murdoch and Conradi, again and again. Making a point isn't always the point: but sustained pointlessness is wearisome.

One issue not frontally attacked by Conr-adi is the conflict between Murdoch's much-vaunted search for, and alleged achievement of, goodness, "which shines through so much of what Iris did and was", and her often (it seems) cruel, self-centred, inconsiderate behaviour to others, especially in the sexual arena. Not until page 538 is her promiscuity so referred to, and nymphomania is nowhere mentioned. (Nor, incidentally, are contraceptive arrangements - a regular but strange omission in accounts of colourful sex lives that do not issue in pregnancy. Was she fertile?) Murdoch herself appears to have seen the problem better than her biographer: "that business of falling in love with A, then with B, then with C (all madly) seems a bit sickening"; "Who am I? ... a male homosexual sado-masochist"; "Does this endless capacity for new loves shew ... that I am very shallow, unstable?"; and, in a poem to her husband, "So many others in my life have place / How can I dare to look you in the face?"

And what of the novels? Conradi has written another volume about these, and this is after all a biography: so the books properly get only walk-on roles. But I have always been puzzled by the standing of her work, and not enough is done to assuage the feeling that her reputation as a novelist may be overblown. Again, Murdoch is her own sternest critic: in The Black Prince she gives the fictional novelist Arnold Baffin a plot with "ludicrous scenes of an abbot felled by an immense bronze crucifix and a Buddhist nun with a broken ankle rescued from an overflowing reservoir". This is self-parody, as many of Conradi's summaries of Murdoch plots remind us. In The Bell , "Catherine, who turns out to be schizophrenic and in love with Michael too, attempts to drown herself. Nick shoots his head off." At moments like this the sceptical critic is inclined to rest his case.

Conradi's shortcomings are abetted by the publisher's. The illustrations are splendid and generous in quantity, and the index is well made. But is the "scrupulous editing" Conradi speaks of a polite fiction? The structural oddities, repetitions, inconsistencies, dangling modifiers, mistakes and plain misspellings (eg "Acropolos", "diminutative", Jenifer Hart's given name, throughout) make one wonder. Most, if not all, of the tedious catalogue of Murdoch's Irish family origins should have been omitted or at most confined to an appendix. "Mrs Walton, Belfast-born foster-mother to Iris's cousin Eva Robinson, seven years Iris's senior, and closer to her than Rene's sister Gertie's four sons": hello?

Similar pruning is needed throughout: do we need to know that one Gladys Short, who may or may not have been Murdoch's kindergarten teacher, was "capable and imaginative", and "died of a bee sting 20 years later"? Attention should have been paid, too, to Conradi's Wittgensteinian tendency to write in pretentious staccato telegraphese: "Mystic: one who has died and therefore fallen in love with the whole world." Perhaps Conradi was as resistant to editorial help as Murdoch, or perhaps he was left to his own deficient devices.

Why, too, was the now almost universal reader-friendly practice of cueing endnotes in biographies by the opening words of the passage in question, so that the text is undefaced by superior numerals, not followed? Conradi's pages are bespattered with numbers. Worse, his notes are numbered by chapter, sometimes reaching three ugly figures, and the information sourced is often of breathtaking triviality, while interesting quotations, such as Stuart Hampshire's put-down of Eduard Fraenkel as "pompous, Germanic, show-off, dominant, boastful, embarrassing, oppressive, humiliating those who were bad at writing Greek prose", or Isaiah Berlin's description of Murdoch as "a lady not known for the clarity of her views", remain unannotated.

Such defects are not fatal, but they do weaken the book. This "life" of Iris Murdoch is full of matter, an indispensable database. But it should not be the last attempt at its task, and for my money its successors should be more avowedly cartographical.

Henry Hardy is a fellow of Wolfson College, Oxford, and editor of the works of Isaiah Berlin.