A review for the New Scientist of:
Grooming, Gossip and the Evolution of Language
by Robin Dunbar,
Faber and Faber, £7.99, ISBN 0571173977
The study of the origins of language has generally been rejected as a topic of serious study - the evidence, it is said, is too sparse for anything more than speculation. The generally accepted view is that language evolved because of the complexity of orchestrating hunting expeditions or as a side-effect of more general cognitive evolution. Linguists have virtually excluded the issue from their field, with only a few exceptions. Most notable of these is perhaps Derek Bickerton, the originator of the bioprogram approach to language evolution. Nevertheless, this subject captures and excites the imagination, and we can be grateful to Robin Dunbar for producing an interesting and readable book about his research on primates and his conclusions about language evolution.
The main thrust of Dunbar's argument rests on the correspondence between primate group size and the volume of the neocortex. His research shows that individuals in larger primate groups have larger neocortexes than those in smaller groups. The neocortex is the area of the brain associated with the higher processes (like language), and Dunbar argues that group size and language are connected, using this data to support this theory. What, then, is the nature of this connection? At this point Dunar draws on the research that he and his colleagues carried out on the topics of casual conversation. The conclusion of this research is that two-thirds of the content of casual conversations is based on matters of social import. This leads to the claim that language is primarily a tool for initiating and cementing social bonds.
At this point, comparative primatology is used to extend this social explanation for language evolution. In social primates, individuals use grooming as a way of establishing and maintaining bonds. This, Dunbar points out, is a very important way of showing who you expect to support in conflict and vice versa. However, when early hominid groups grew larger, it would have become increasingly difficult to find time to groom and do all the tasks necessary for survival. Therefore, a more efficient form of social bonding would have become necessary. The solution to this problem was, of course, language. The statistics seem to give credence to Dundar's hypothesis. The maximum group size in non-human primates is 55, while that of humans is clearly much larger. It is when the author attempts to calculate exactly how many individuals form a human``group" that a certain amount of doubt emerges. Dunbar draws on a diverse and eclectic range of evidence to back up the number which will fit his theory - from hunter-gatherers to military units, he claims, 150 is the magic number. This fits in neatly with the fact that while grooming is a one-to-one affair, conversation typically includes three or four individuals. Thus we would expect human groups to be around three times larger than those of grooming primates. This is one of the weaker points in the book. While there are certainly groups of 150 for some social purposes, there are also smaller (families) and larger (tribes) groups. Dunbar goes to great lengths to convince the reader that people can generally recognise or name 150 individuals and that it is the number of people we can ask to do favours for us, for example.
Despite the doubts that may arise, however, the main argument is not so badly damaged that it necessarily fails. However, as is often the case with popular science texts, Dunbar's book could have been reduced by one third without losing any real substance. The last third of the book consists of speculatory probes into various hypothetical extensions and elaborations of the basic arguments. It is suggested, for example, that language may have evolved as a kind of social lek. This is plausible, but attempting to follow Dunbar's frequent and seemingly random tangents becomes tiresome towards the end of the book. Although on the whole an enjoyable and potentially informative experience, reading it does at times resemble listening to a learned but slightly absent minded lecture. The main criticism of the book, then, must be that it lacks a rigorous structure. Two hundred pages would be better reduced to one hundred and forty well-structured and unsuperfluous ones.
Despite these criticisms, the central hypothesis of the book is interesting and plausible, and perhaps more importantly it implicitly suggests that some of the most useful approaches to questions about language origins may lie in the area of comparative primatology. One of the unfortunate results of the lack of enthusiasm shown by linguists in this area is that there is little collaboration between them and behavioural biologists like Dunbar. This sometimes becomes apparent, for instance when the author mistakenly writes that humans may think in the subject-verb-object framework because that is how language is structured. English and many other languages are, but at least as many are not.
Although Grooming, Gossip and the Evolution of Language
does not really do justice to its central theme,
the strength of the main hypothesis is sufficient to make the book an
interesting and important one.
Richard Wright is a postgraduate student at the School of Cognitive and Computing Studies, Universtity of Sussex