Abstract: It is widely accepted that many species of nonhuman animals appear to engage in transitive inference,
producing appropriate responses to novel pairings of non-adjacent members of an ordered series
without previous experience of these pairings. Some researchers have taken this capability as
providing direct evidence that these animals reason. Others resist such declarations, favouring instead
explanations in terms of associative conditioning. Associative accounts of transitive inference have
been refined in application to a simple 5-element learning task that is the main paradigm for
laboratory investigations of the phenomenon, but it remains unclear how well those accounts
generalise to more information-rich environments such as social hierarchies which may contain scores
of individuals, and where rapid learning is important. The case of transitive inference is an example of
a more general dispute between proponents of associative accounts and advocates of more cognitive
accounts of animal behaviour. Examination of the specific details of transitive inference suggests
some lessons for the wider debate.