Keith Stenning

A logical characterisation of a human language phenotype finds a central role throughout cognition

Half of the problem of understanding evolved cognitive phenomena is characterising the phenotype. In understanding the evolution of the human capacity for language, the phenotype is probably even more than half of the difficulty. This talk is about what happens when one takes just a subset of the phenomenon (human narrative discourse) and uses a nonmonotonic logic to characterise it as a species of planning (van Lambalgen & Hamm, 2004). This first of all changes ideas about the relevant phenotypes of the ancestors, and second, it redraws the boundaries of the human cognitive faculties that are involved.

(Stenning & van Lambalgen, 2008) proposed a nonmonotonic logic at the heart of human reasoning—specifically the component of reasoning to an interpretation. It argued for a rethink of the biology of cognitive evolution in the mould of EvoDevo (Evolutionary Developmental Biology). This latter argument appeared in chapters 6 and 9. It contrasted the Chomskyan (and more broadly linguistic) conceptualisation of language that focusses on the structure (phonological, syntactic, semantic) of infinite sets of sentences, with the level of discourse processes at which language is a medium of social action, the basic action being the cooperative communication of the preferred model of the narrative to its recipient. This is the process of generating a new interpretation for a fragment of a language in the light of general knowledge recruited from semantic memory, through principles of cooperation. The nonmonotonic logic of this process is Logic Programming (LP). LP is also known as ‘planning logic’. All that you need to know about EvoDevo for this talk is that it says that evolution takes place by tweaking complex systems within phenotypes through changes in control genes, thus transforming one complex phenotype into another complex phenotype. This is contrasted with the idea (common in psychology) that evolution works by adding new modules.

So planning, rather than communication, is our one-word characterisation of (this part of) the modern language phenotype, and this changes where we look for the origins of this part of language in our ancestors. With this radical change of conceptualisation, two broad questions open up: Which of our other cognitive capacities can be usefully construed as planning? And what were our ancestors’ planning capacities that transmogrified into modern language? This talk will focus on the first of these two questions where recent psychological advances have made a redrawing of the map an exciting possibility (Hassabis & Maguire, 2007).


Hassabis, D., & Maguire, E. A. (2007). Deconstructing episodic memory with construction. Trends in cognitive sciences, 11(7), 299–306.

Stenning, K., & van Lambalgen, M. (2008). Human reasoning and cognitive science. Cambridge, MA.: MIT Press.

van Lambalgen, M., & Hamm, F. (2004). The proper treatment of events. Oxford and Boston: Blackwell.