Clues to the origin of the human mind from primate observational field data

Prof. Richard W. Byrne (University of St Andrews, Scotland, U.K.)

This lecture uses two topics in primate cognition to show that observational field data provide crucial information for reconstruction of cognitive evolution. Firstly, learning to deceive familiar conspecifics (tactical deception) is a good indicator of cognitive ability, for which a unique corpus of data exists spanning all major primate taxa. To learn deceptive tactics, individuals must distinguish their companions as individuals, keep track of who is currently in view, and learn from a single trial. The frequency of deception varies dramatically across species and depends on the size of the neocortex: both absolute volume and neocortex ratio are good predictors of the rate of use. The animals may not understand the deceptions they practice, however: tactics may be learnt without comprehension of their mechanism. Evidence that primates deceive in an intentional way is much more sparse and limited to the great apes, where there are cases in all species. This monkey/ape difference is consistent with other lines of evidence suggesting that great apes have greater cognitive understanding than monkeys in particular, understanding of the intentions of other individuals. Great apes do not live in systematically more complex groups than those of monkeys. However, all great apes feed in complex ways, apparently giving them access to foods that monkeys cannot exploit. In the programs of manual action apes use for these tasks, several different stages of processing are reliably sequenced; the hands are used in complementary roles, coordinated together to achieve a single goal; and the overall plan is hierarchical, with modules repeated or omitted according to task demands. The overall program seems to be learnt by imitation (but not by slavish copying of details), which suggests that great apes are able to interpret the skilled actions of others: discerning the results to which particular actions are normally aimed and the steps necessary to achieve them. The ability to use regularities in skilled behaviour to work out the underlying correlational structure of intentional action, behaviour parsing, is a crucial first step in understanding mental states. Plausibly, the ability to interpret intelligent action by behaviour parsing derived in evolution from the need for great apes to compete effectively with (smaller, more agile) monkeys for similar foods, allowing the later development of an ability to read intentions of others.