Articles for novices, an index for experts

Nature Publishing Group specialises in the production of large multi-volume encyclopedias. This four-volume set is NPG's attempt at providing the definitive reference work for cognitive science. At more than 4,000 pages and 700 articles, it is undeniably impressive, and NPG's experience translates into a well-produced collection with a comprehensive glossary and extensive index.

The project was overseen by Lynn Nadel, an active and respected researcher in the cognitive neurosciences at the University of Arizona.

Nadel also acted as section editor for the neuroscience articles. Other major disciplines are covered by four section editors -all specialist academics with solid reputations in their fields. These are David Chalmers (philosophy), Peter Culicover (linguistics), Robert French (computer science) and Robert Goldstone (psychology). Five ancillary section editors, again respected specialist academics, ensure that the more tangential areas were not neglected, and a further 15 academics acted on an advisory board.

Articles span the entire field, from the neuroscience of specific brain structures to the philosophy of artificial intelligence, and from linguistic theory to anthropology. Biographies of major figures are also included.

Nadel and Massimo Piatelli-Palmarini provide a lengthy introductory essay addressing the critical question "What is cognitive science?". They adopt a historical approach, describing how researchers from a variety of disciplines became involved in critical questions concerning the functioning of the mind. The history is important and a good way to answer the question, for the fundamental research questions and methods of cognitive science owe much to the discipline's origins in the decline of behaviourist psychology and the rise of information theory and artificial intelligence.

Cognitive science, Nadel tells us, addresses questions that have broad appeal, and the encyclopedia is deliberately aimed at a wide readership with introductory articles for intelligent novices, intermediate articles for those with more experience and advanced articles on specialist topics for the very dedicated. I do not fully understand the rationale for this.

While some topics are undoubtedly more technical, I imagine some users will be frustrated when they find their desired topic is not treated at their level. Having said this, the articles are well written and the introductory essays at least are generally very accessible.

Articles are ordered alphabetically by title but some titles seem a little idiosyncratic, and in any case most users will probably access the content through the subject index. This is itself a major reference work, consisting of more than 200 pages in small font. The problem is that the index is also idiosyncratic. Individual authors have clearly provided their own index terms. This is not unusual, but some have been more diligent than others, and different authors often index related topics under different terms. Furthermore, in attempting to be comprehensive, the index is nearly overwhelming. It is certainly of limited use for common terms or broad topics, which may be cross-referenced to many articles. Also, the index does not indicate the level of difficulty of the articles to which it refers, so users cannot focus their search.

Given the difficulties of providing the perfect index in print form, it is odd that there are no plans for an electronic version, which could offer greatly enhanced cross-referencing facilities. An online version, offering updates, would be even better.

Any new encyclopedia of cognitive science will inevitably be compared with the Massachusetts Institute of Technology Encylopedia of Cognitive Sciences , 1999 ( Mitecs ). Mitecs' articles are considerably shorter, and generally written for a more advanced audience. In contrast, the NPG articles have more space to introduce topics gradually, generally making fewer presuppositions. Mitecs is a serious reference work for academic and research-oriented cognitive scientists, while the NPG encyclopedia is better suited to student use. Beyond this, the quality of articles across the two is similar, and some contributors have even written for both encyclopedias.

The length and different levels of most of the NPG articles give it a textbook flavour. It is perhaps instructive, therefore, to compare it with existing cognitive science texts. Clearly, no textbook can compete in terms of coverage. What the encyclopedia lacks is coherence and something to draw the various disparate entries together. I suspect the alphabetical ordering of entries works against this, because hardcore neuroscience is interspersed with linguistic theory, and economics is interspersed with cognitive psychology and social cognition. Most cognitive science textbooks offer a more coherent format that draws the disciplines together and highlights common ground.

Turning to the content, articles are generally of a high standard.

Inevitably there are overlaps, some more stark than others. I was drawn to the article headed "Blindsight", for example, which is followed by one headed "Blindsight, neural basis of". More curiously, the former includes a section on anatomical bases of blindsight. What is the relation between the two articles? Perhaps surprisingly, the answer is very little. Both are informative in their own right. Arguably the justification for the two is that the former is pitched at an introductory level, the latter intermediate. But there is just one cursory cross-reference, and significant duplication between the two. It seems the first was commissioned by the philosophy editor, while the neuroscience editor commissioned the second. Certainly, no one person could be expected to read and cross-check every article, but oddities such as this do give the impression that there was not enough communication between editors.

Big things are often awe-inspiring and this encyclopedia is no exception.

It was clearly a huge part of many people's lives for the five years it was under preparation. But is there anything impressive beyond its sheer scale? I am not convinced that there is. It is hard to feel that it is anything more than a collection of a few hundred semi-related articles. Perhaps this is all an encyclopedia should be, but I cannot help feeling this is a missed opportunity.

One serious issue within cognitive science is that its component disciplines can appear to have little in common. This has serious sociological consequences for the discipline. It poses significant difficulties for teaching and acts against public understanding. This encyclopedia does little to challenge the apparently disparate nature of cognitive science. Indeed, glancing at the table of contents encourages it.

For example, an article on "asset market experiments" is followed by an article on "attention". The casual browser may be forgiven for thinking cognitive science is a hotchpotch of unrelated disciplines. At the other extreme, an article on "stress" is followed by one on "stress and cognitive function". While these might seem closely related, the former concerns stress within speech production while the latter concerns the psychophysiological phenomenon of stress. One wonders if these apparent discontinuities and false continuities could have been addressed by more interaction between editors. Such interaction might also have limited some of the overlap between articles and ensured greater uniformity of levels between the contributing disciplines.

I do not mean to end on such a negative note. Many institutions will rightly consider this a "must-have" publication. For many, the indisputable strength of the encyclopedia -the way it draws together so much information on so many issues concerning our current knowledge of the structure and function of the mind and brain -will far outweigh niggling concerns about the titles of some articles. Certainly, any institution with an undergraduate cognitive science programme will find the encyclopedia indispensable. I can also imagine it finding a good home in many public libraries.

Richard Cooper is reader in cognitive science, Birkbeck College, University of London.