Which came first: language or high-order metarepresentation?
On standard Gricean and neo-Gricean accounts of language development in ontogeny and phylogeny (defended by Tomasello, Sperber and Wilson, and Scott-Phillips among others), language acquisition becomes possible only when speakers are able to act with and understand communicative intentions. Since acting with and understanding communicative intentions is thought to require very sophisticated socio-cognitive abilities, including the ability to entertain fourth order meta-representations, the possession of fourth-order metarepresentational abilities is held to be a pre-requisite of language development. On the standard view, it was the emergence of uniquely human abilities for high-order metarepresentation in phylogeny that led to the evolution of language; and it is because young children but not apes possess these abilities that they alone can acquire language.
For all that the standard view has been adopted by developmental psychologists, it faces potentially insurmountable empirical obstacles. This is because while children start to use words not long after their first birthday, current evidence suggests that they master fourth-order metarepresentations only around the age of 11. Given this, I argue, standard interpretations of the cognitive pre-requisites of Gricean communication must be wrong.
Against the standard view, I defend a 'minimally Gricean' account of intentional communication, according to which the socio-cognitive abilities required for Gricean communication are shared by both human and chimpanzees. I offer a new explanation of why humans but not apes acquired language, and I defend the view that human abilities for high-order metarepresentation are likely language dependent. In the course of developing my position, I will also argue against the view (defended by Gergely and Csibra) that ostensive eye contact constitutes a human-specific adaptation for understanding communicative intentions.