Cognisant conversations

Cognitive science is often defined as the interdisciplinary study of the mind. In some ways it is a mongrel field, drawing on (at least) cognitive psychology, computer science, linguistics, neurophysiology, philosophy of mind, and artificial intelligence in order to understand the workings of the mind. This hybrid background has led to a field which is full of debate, and this volume, Speaking Minds: Interviews with 20 Eminent Cognitive Scientists, attempts to make this debate accessible.

There is a clear role within the infrastructure of cognitive science for a book of interviews, and I congratulate Peter Baumgartner and Sabine Payr on undertaking a project fraught with both academic and administrative difficulties. Indeed, my enthusiasm for the project is immeasurable, but despite this (or perhaps because of it), I found the result something of an anticlimax. Although there is plenty for the educated reader to argue with in Speaking Minds and the conversational interview style provides an informal and very human feel to the subject (which is well complemented by the inclusion of photographs of all interviewees), there are several flaws that detract from its undeniable potential. First, the interviewer (Baumgartner) is sometimes naive, and often overly accepting. Only rarely do the major debates within the field bubble to the surface, and the highly structured nature of the interviews coupled with Baumgartner's tendency to adhere to his own conception of critical questions tends to stifle the issues dear to the interviewees.

Baumgartner pursues three primary issues: the utility of the Turing test, the validity of Searle's Chinese Room argument, and the problem of background knowledge. In the Turing test (originally devised by Alan Turing as a test of human-like intelligence) one person puts questions to both a computer and a second person (located in another room). The computer (or its program) is said to pass the Turing test if the first person cannot tell, from the answers to his/her questions, which of the other two is which. Searle's Chinese Room argument is a parable intended to demonstrate that understanding of language involves more than the manipulation of symbols, and hence that computers cannot achieve true understanding. The problem of background knowledge concerns the vast amount of knowledge that we require in our simplest real-world interactions, and the resultant computational difficulties (in terms of storage and access) that this appears to demand. These questions have had a major impact on AI and cognitive science, but the extent to which they shape the progress of working cognitive scientists today is open to debate.

A second flaw is that the interviews are reproduced essentially verbatim. While major editing would not be in line with the spirit of the interview-based approach, this does mean that there is a lot of repetition of basic common ground. Is it really necessary, for example, to include each interviewee's description of the Turing test? Surely a single quote from the original source (which is in any case included in a very useful glossary) would be sufficient.

More seriously, the picture presented by the interviews when taken together is biased. Obviously there will be omissions in any collection of interviews, but not only are all those interviewed American scientists (odd given that Turing was British), they are also predominantly from the AI side of cognitive science. Both cognitive psychology and linguistics fare very badly. None of those interviewed could list either of these disciplines as their primary background subject. Such a bias can only serve to encourage the confusion between AI and cognitive science.

Regardless of these criticisms, Speaking Minds should find a home on every cognitive scientist's bookshelf. The frank and friendly style of the interviews makes the book both an invaluable accompaniment to a standard text and an excellent educated layman's introduction to some of the more computational issues in the science of the mind.

Richard Cooper is a lecturer in psychology, Birkbeck College, London.