The ghost in the machine. What happens in the brain when we remember? A new study of trends in neuroscience attempts to unravel one of the mysteries of consciousness

Memory

Patricia Fara and Karalyn Patterson (editors) <em>Cambridge University Press, 207pp, £17.95

"In our part of the country," Montaigne said, "we call a man who has no memory stupid." "Memories," sang Barbra Streisand, "like the corners of my mind/Scattered pictures full of mem-or-ies/Of the way we were . . ." I wonder if Terrence Sejnowski, author of the paper entitled "Memory and Neural Networks" which backs up this symposium on Mnemosyne's works, would concur more with the former statement or the latter? And further, how does he feel about Kris Kristoffersen as a romantic lead?

Sejnowski, ostensibly the hardest of hard scientists, emerges with the writing credits from this bizarre, interdisciplinary Festschrift. There's a haunting suggestibility about the coinages of neural networking, the discipline which attempts to create computational homologues of brain processes. I can't imagine dear old Antonia Byatt coming up with a trope half as good as "graceful degradation" (the method whereby a neural network continues to function after being damaged), or one as queerly euphonious as "error back propagation" (unfortunately a statistical, rather than moral form of feedback). And show me the tyro of the science-fictive who has originated anything as good as the "perceptron", a learning algorithm, which, when structured in a multilayered network with feedback connections, becomes the sublime "Boltzmann machine".

I felt rather at home among the corners of these geometric minds; indeed, it may be that Memory is best read backwards, starting with the join-the-dots modelling of Sejnowski, then going on to Steven Rose and his plaintive search for keys to the connecting doors of the lumber rooms of the mind. Rose, in his contribution "How Brains Make Memories", conjures up a most peculiar (yet paradigmatic) vision of the examined life: "I am convinced that if I had a personal cerebroscope, wired up through life to a portable MEG machine or whatever, I could make this connection [between subjective memory and verifiable data]. I am also pretty sure that the content of my memory will always be coded in the language of my mind."

I like that - it has a certain ring, don't you think, like a dimly remembered 1960s pop song, windmills, languages, minds, round and round. It's surprising to come upon it at the end of what is otherwise a competent and accessible attempt to marry the tough technicalities of the neuroscientists (themselves a diverse bunch of neurobiologists, artificial intelligence modellers, molecular biologists and quantum physicists), to the woollier abstractions of psychologists, philosophers and lay persons. Rose states that as we surge towards the millennium, "the sense of confidence, even arrogance, among the industrial world's neuroscientific community is tangible". He demonstrates with deft economy why that arrogance is justified. His prose isn't too bad, either.

Rose has a marvellous phrase with which to link the "language of brain" (in this case proliferating synaptic connections in the brains of day-old chicks) with that of mind. He speaks of such connections being "pruned away, leaving what we may presume to be only the minimum necessary for recall or recognition of the bitter bead". It's fantastically evocative, that "bitter bead", so much better than Proust's poxy old Madeleine - even if you don't know it refers to a literal bead, coated with some bitter substance, which is fed to the chick in order to promote the "cascade of biochemical reactions" we term memory.

Rose actually has hard suggestions for the treatment of patients with global amnesias, such as those associated with Alzheimer's: steroids rather than drugs which effect the neurotransmitters directly. Rose also provides us with a useful memory test as we work our way backwards. He employs an illustration of an Elizabethan "Memory Theatre" in his discussion of the implications of mnemonics for human memory; and, lo and behold, the very same illustration crops up some 70 pages earlier in A S Byatt's discussion of the same dynamic.

In another essay, Barbara Wilson doesn't want to give amnesiacs steroids; she wants to give them "neuropagers". These handy devices are "a simple and portable system usually worn on the belt. Reminders are programmed into a computer by a therapist. The computer is linked by telephone to a paging company, and once messages are entered into the computer, no further human interface is necessary." It sounds very like how the government's whip is organised, but is actually the downbeat conclusion to "When Memory Fails", Wilson's survey of the amnesiac's shadow land.

Wilson's case histories have none of the pizzazz of Oliver Sacks' expositions of twisted savants and maladaptive prodigies; but at least one of them does present us with curious questions about the author's own memory. Wilson describes the case of one "CW", who, she informs us, was, until he developed herpes simplex encephalitis, "chorus master of a leading contemporary music group, conductor of a professional Renaissance group, and one of the world's leading authorities on Orlando Lassus". Now, is this a load of cobblers? Or has Wilson rather forgotten not to reveal the identity of her patients? I actually went down and checked my Hyperion recording of Lassus' Requiem Mass for Six Voices to see if I could identify the impresario with the venereally diseased cognitive capability.

Wilson's piece is notable for some more great illustrations. There's an encephalitic's Cocteauesque cartoon which recalls the drawings done by A R Luria's patients. Luria, the founding father of neurology, receives only a few mentions in this volume, although his ground-breaking studies The Mind of Mnemonist and The Man with the Shattered World are masterpieces of forensic science and forensic prose.

Not so Juliet Mitchell's "Memory and Psychoanalysis", which manifests all the worst characteristics of second-order Freudian discourse - a tendency to recount childishly opaque case histories, and to make self-aggrandising propositions - "Psychoanalysis has progressed from understanding the amnesia of the hysterical girl as patient to the capacity to hold in mind an image of the psychoanalyst as mother" - and reach staggeringly arrogant conclusions, which are by no means worth quoting in full, but amount to the proposition that psychoanalysis is a form of transcendental idealism, and the analyst a magus who conjures up the very antimonies of mind.

Forget it. Better to move back to Jack Goody's down-home analysis of the Bagre myth of the LoDagaa of northern Ghana. This everyday tale of field-working folk tells us what, I suspect, we always should have known; that oral cultures' methods of information dissemination do not fulfill the same criteria of factuality as those of literate cultures. In other words, there is no rote learning without writing. Interestingly, it was the arrival of another analog device, the tape recorder, which made it possible for anthropologists to analyse these formalised recitations properly.

In his characterisation of oral memory as "simply experience reworked", Goody comes close to suggesting that there are profound correspondences between collective and individual methods of remembering; just as Steven Rose, with his description of the mind's polity as "an efficient anarchist commune, rather than a Stalinist command economy", pushes the analogy in the opposite direction.

For Catherine Hall collective memory needs to be subjected to a re-examination, so that "heritage experiences", such as Cadbury World, at Bournville outside Birmingham, are properly deconstructed. Hall sees our failure as a nation to move beyond our imperial past as crucially enmeshed with the sanitised versions of our history which we retell. I agree with her; we should implant our history in visitors to these isles like 16 inches of cold British steel; or alternatively they can go to Billing Aquadrome instead.

Hall and Richard Sennett (who begins - and thus concludes - this collection) are both good liberals; for them it is through memory that "we seek to possess ourselves". They are equally uneasy with the longue duree of personal private experience. In Sennett's "Disturbing Memories" he recounts the cost to redundant, middle-aged IBM computer programmers of becoming too "centred" in their own memories.

There's no danger of A S Byatt suffering from such a destructive centring through memory, because for her - as for Proust - certain childhood experiences have inculcated "a sense of duty towards them". It would be pretty easy to be snotty about Byatt's essay; after all, there are few more egregious spectacles in the known world than a writer quoting extensively from her own work - let alone placing those quotations alongside others from The Prelude, The Aeneid, Paradise Lost, The Mill on the Floss and - bien sur - A la Recherche du Temps Perdu. But this being noted, hers is an agreeable enough piece of whimsical donnishness, and I can agree well enough with her view of writing as "not unlike remembering" in its reticulation of experience. However, I can't pass all the way down her intertextual garden path, and unlike Byatt I don't view art as a glorified A-level subject.

Patricia Fara and Karalyn Patterson are to be congratulated for Memory, which is far more than the sum of its parts. Really, with the demise of philosophy's pretensions to true esemplasticity, it has to be to this kind of collection that we must turn.

Will Self is a novelist and newspaper columnist