I think I remember

I must declare a personal interest in this book: it exposes a possible false non-memory (otherwise known as amnesia) for me, or a false positive memory for the editor. The sole acknowledgement states that the original idea for the present volume was mine and offers gracious thanks. This is not how I remember its origins, although I was admittedly an adviser for the OUP's Debate series. The credit must belong to Martin Conway. But in a different role - as a member of the independent Scientific Advisory Committee for the British False Memory Society - I awaited the outcome with interest; our committee has long advocated rational discussion between therapists and experimentalists (and damaged clients and parents) of this heated and well-publicised issue that crosses so many tragic boundaries and touches so many lives.

How does the book turn out? Its high points are very good. There are only a few low points, among them those which could have been predicted to be adversarial. For one who wants a review of much of the relevant evidence on the vagaries and distortions of human memory, placed in the context of memories of putative or real childhood sexual abuse, this will be a good and authoritative source. For a review of clinicians' attitudes and relevant case material, this will also be valuable, especially that of the final chapter. Most of the authors are already well known, but it is useful to see their efforts when the predicated focus is supposedly a "debate". There are, however, even among such an array, some obvious omissions. Elizabeth Loftus, Ulrich Neisser, Stephen Ceci, Richard Ofse, Marcia Johnson, among others, are very well known for their pertinent studies. Lindsay, Poole, and colleagues, whose important surveys and analyses are frequently referred to, would have been appropriate. As much of the evidence and argument in the book turns on the issue of "encoding specificity" and the importance of specific cues for retrieval, Endel Tulving would have been an appropriate and distinguished contributor, as he was central to the conceptual and experimental development of the concept. Perhaps these persons were invited but declined.

The introduction by the editor, Conway, provides a cognitive psychological background of autobiographical memory and a helpful summary of points in individual chapters, even though as the convener of the debate he is a touch judgmental, nor can he resist assuming a didactic role, demonstrated in his own chapter later in the book. In it he argues that his "constructivist" cognitive theory of autobiographical memory is compatible with Freud's own views. It is a provocative and broad-ranging theoretical analysis, but not easy reading.

The other authors fall roughly into two categories, experimental cognitive or developmental psychologists of memory, and those whose main focus is either in the clinic or as commentators on that scene. As a group they grapple with the problem of relating what is known from formal memory research to reported memories of putative emotional clinical events. The result is not really a knock-down confrontational "debate" but, with a few exceptions, a dispassionate review of evidence and case material, and is very welcome. What would have been interesting, and would have counteracted the reflexive culling by readers with predetermined views, would have been to see published rebuttals and exchanges between the contributors, as in a genuine debate.

Some chapters are of special merit. The review by Daniel Schacter et al is, as one might expect, scholarly, erudite and replete. It deals especially with the evidence regarding memories about traumatic events, leading to "psychogenic" amnesia, and also with possible neurological changes associated with them in relevant structures of the brain. It also develops the concept of "source memory" as an identifier or misidentifier of recalled events.

The contribution by Henry Roediger et al is a review, tightly analysed, of their experimental work demonstrating the ready generation of false memories even with short delays in normal subject, and the effects of repeated practice and guessing, and on "reminiscence". It does not deal with trauma or emotive material as such, but extrapolations. The chapter by Sven-ke Christiansen and Elisabeth Engelberg ranges through both the clinical and experimental domains, with particular emphasis on traumatic effects on attention.

The final chapter, by Schooler et al, has such remarkable balance and breadth of argument, with almost unique corroborative evidence, that it is bound to be a source of reference. These authors acknowledge the great dangers and demonstrations of false memories, but they analyse four cases of recovered memories of abuse for which there is some measure of corroborative evidence. They note just where there evidence might be questioned and where there is residual ambiguity, and they probe the issues with sophistication. They take "the middle line", not a Magimix blend but a mixed and varied menu.

The developmental chapter by Robyn Fivush et al reviews evidence about the children's memories (as distinct from the memories of childhood by adults - so-called "infantile amnesia"). They usefully review the good evidence that very young children can learn and remember, but with more rapid forgetting; an appeal is made to the role of language in changing the accessibility and organisation of memory. But why is it that such reviews ignore the animal evidence? Even rats have infantile amnesia, and for pretty nasty events, as shown by Byron Campbell and coworkers. The parsimonious biological view that maturation of critical brain structures is important in infantile amnesia is not considered.

Chris Brewin and Bernice Andrews contrast the clinicians' perspective with that of the experimentalists: the former "represents a human science that is concerned with intentions and goals, the experimental perspective represents a natural science that is concerned with non-intentional mechanisms and processes". But as is clear from neuropsychology, no such contrast is necessary; one can integrate both mechanisms and intentions in the same domain. It may be that conclusive evidence is more difficult to obtain clinically and that laboratory tests are restricted for ethical reasons. It may or may not be that traumatic events demand special principles of memory dynamics. But it is a slippery slope if it is assumed that rigour can be relaxed for the clinician. Thus, "given our knowledge of the integrity of autobiographic memory . . . we may question why recovered memories have to be assumed to be false and proven to be true. It would seem strange to assume that ordinary memories were false unless proven otherwise, and there is no statement of the grounds on which recovered memories are believed to be different from ordinary memories".

Given the clinical context in which false memories have occurred, with families torn apart, this is a remarkable statement. There are clinical (and legal) situations in which the dangers and effects of false negatives and false positives are not symmetrical.

Christine Courtois argues along similar lines, although with an informative review and with a broad and tolerant therapeutic stance, "both therapist and client must be able to tolerate ambiguity and uncertainty".

A recurring message throughout is that memories - "recovered" or not - can either be true or false, and that without corroborative evidence one must remain agnostic; Michael Yapko, John Kihlstrom, Schooler et al, and Schacter et al, especially, insist on this point, both for "explicit" as well as "implicit" memories. It is a bit rash, in the circumstances even risky, for some of the authors (Conway, Christianson and Engelberg) to try to suggest or indicate ways in which true vs. false memories differ. There is simply no sure way in which corroboration can be bypassed.

A quibble: why no author index? It is irritating not to be able to trawl for the contributors' discussions of others' views, especially those that are controversial.

Lawrence Weiskrantz is emeritus professor of psychology, University of Oxford.