Why two plus two can equal one

Debates about whether all humans "think alike", irrespective of their cultural background, are by no means new. The title of this book deliberately echoes that of a similar collection of papers published nearly a quarter of a century ago. (The two have just one contributor in common.) This already invites consideration of the extent to which thinking about thought has changed in the interim.

In the present volume there is a great deal of trumpet-blowing for "cognitive science(s)", but it remains unclear whether cognitive science itself is a distinct mode of thought. No one even seems very sure about whether "scientific thinking" in general is one either. In short, the question of what constitutes a mode of thought remains as vague as ever, although the introduction insists, with appropriate vagueness, that it is nevertheless an important "concept".

What is on offer here is certainly a good mix of approaches, perhaps a bewildering mix. Geoffrey Lloyd proposes a cautious comparison of astronomy, mathematics and medicine in Greek and Chinese antiquity. Stanley Tambiah examines "linguistic ethno-nationalism" and related attitudes. Brian Stock analyses "the self" in the autobiographical writings of Augustine and Petrarch. Ian Hacking discusses the 19th-century notion of "normal people". Yaron lzrahi tackles European thinking about authority in the modern state. Jerome Bruner detects four "universal" ways in which people make sense of themselves and the world. Carol Feldman and David Kalmar focus on western autobiography and fiction. Keith Oatley invites us to consider inference in narrative and in science. David Olson tries to characterise "literate mentalities". Cameron Shelley and Paul Thagard plunge into the murky depths of mythology. Susan Carey investigates the notion of "cognitive domains". Scott Atran asks how people from different cultures think about "living kinds". Deanna Kuhn queries the distinction between "scientific" and "ordinary" thinking. Myron Tuman looks into the idea of "collaborative thinking" in education.

It would doubtless be regarded as foolish and old-fashioned for a reviewer to ask what all these papers have in common. That would be dismissed as benighted essentialism. A more acceptable question might be to ask what "family resemblances" link them. But it is hard to get very far with that either, since the answer seems to boil down to bandying about words like "thought" and "cognition", often with little or no attempt at definition. The diversity of the contributions is in one way attractive, but the corresponding weakness of the volume is that it lacks focus.

The best of the papers are well written, lucid and sophisticated. At the other end of the spectrum, we have to cope with gobbledygook such as: "From a cognitive standpoint, each culture is constituted by a more or less loosely connected ecological network of cultural representations, which are themselves only more or less loosely connected ecological networks of their various public and private versions." Being circular as well as sesquipedalian, this is not a good advertisement for the "cognitive standpoint". Fortunately, such lapses are rare. But it is noticeable that most of the contributors are less critical of their own thinking than of the thinking by others which they analyse. Here one might expect to find a credible candidate for cognitive universality, but no one in this volume seems to have spotted it. (Kuhn perhaps comes closest.) One old chestnut that turns up predictably is the problem of the relation between thought and language. This is central to Atran's paper, is skirted rather than confronted in Tambiah's, and appears briefly in one or two other places. But some contributors are surprisingly reluctant to address it. Lloyd, for instance, writes about ancient science as if it was simply irrelevant that Greek scientists wrote in Greek and Chinese scientists in Chinese. Perhaps it is. But the point needs to be raised and the case argued. It cannot just be taken for granted, as when Lloyd gives the bland assurance that "mathematical truths do not vary across cultures". One of his examples is the "truth" that two plus two equals four. But if your native language lacks a numerical term for "four" or any higher integer, in what sense do you nevertheless grasp this particular truth as a truth?

The same example recurs in Carey's paper, where it is suggested that we learn that two plus two equals four "by observing two sets of two things combine into one set of four things". This seems highly unlikely: certainly no more likely than observing that some sets of two things can combine into a single object of the same kind leads us to learn that two plus two equals one. But Carey also talks airily about our "innately given system of knowledge of number". Since it has long been recognised that most of our serious thinking has to be articulated via language or some other form of semiological systematisation, it is surprising to find anyone now treating truth and semantics as so easily divorced. But perhaps readers are expected to accept that our innate mathematics is automatically made viable by our equally innate knowledge of language. Innateness is a notion often invoked in order to put an end to regresses; but it actually generates more regresses than it terminates.

It is also surprising that no contributor is keen to make much of the fact that the locus of human experience is the human body. The scope for cross-cultural variation on this front is very limited. Shakespeare realised this long before the first professor of cognitive science was ever appointed. If you prick us, do we not bleed? If you tickle us, do we not laugh?

Is it not everybody's experience that fire burns, that if you stand out in the rain you will get wet, and a myriad similar things? Here is an area where we have little - and usually no - room for manoeuvre about what to think. Nor is this surprising. For it would have been extraordinarily obtuse of God, nature, evolution to produce a Homo sapiens whose sapience failed to recognise, or perversely ignored, such lessons of experience.

Of course, the thinking manifested in giving explanations of experience can vary widely. But hardly our thinking about the experiences themselves. That is where all thinking begins, and whither it must return in terms of day-to-day accountability. A species that is going to survive does not have much option about this. But there is plenty of opportunity for divergence when it comes to thinking about how the world began or how to lead a good life.

Those who try to persuade us that a devout Christian and a devout Hindu think alike, in spite of holding quite different beliefs, merely demonstrate that there is something wrong with their own criteria for "thinking alike". If everyone thought alike in matters of religion the world would be a quite different - and a happier - place. All this may sound very much like plain common sense. But it needs to be said because common sense is sometimes conspicuous by its absence in Modes Of Thought.

One basic criticism of the book is that, in spite of the broad sweep promised by its title, it reflects a very blinkered view of the forms thinking can take. There is nothing in it about the sheer scope and creativity of human imagination. What the contributors, collectively, understand by "thinking" seems to be reasoning of some kind, based on itemisation and classification. Which is doubtless important. But if that were all the human mind were capable of, our mental lives would be infinitely poorer than they in fact are. In the end, Modes of Thought tells us less about modes of thought than about academic modes of writing about thought. One current vogue in psychology seems to be for couching whatever you want to say in terms drawn either from traditional rhetoric or from poetics, with a few references to postmodernism thrown in. The end product often reads like second-rate literary criticism dressed up as science.

Apart from being trendy, however, this reflects a more fundamental problem concerning thinking about thought. Only Olson comes anywhere near putting it in its proper perspective. All the contributors are dealing with controversial issues that arise - and can only arise - in a literate society. It is not simply that they happen never to arise in a preliterate society. Rather, literacy is what promotes the kind of thinking that engenders them, as well as the kind of thinking available for tackling them. Questions about universal cognition are eggs that stand no chance of hatching at all except inside the nonuniversal incubator that writing provides.

Roy Harris is editor, Language and Communication.