Simon Kirby

From Santa Fe Institute Events Wiki

Language as an Adaptive System: are models of cultural evolution really modelling change?

Abstract. Traditionally, researchers studying the origins of human language have seen this as a primarily biological enquiry. The motivating question has been the origins of a putative species-specific faculty for language, and the mechanism of explanation has been biological evolution by natural selection. This approach has in part been motivated by the idea that language shows evidence of apparent design - in other words it is adaptively structured. In recent years, however, there has been increasing realisation that the role for natural selection in explaining adaptive structure in language may have been overstated. A growing body of research highlights the importance of cultural evolution instead in understanding the origins of linguistic structure. In this view, biological evolution in humans has provided the substrate upon which cultural evolution has been able to operate, and it is this latter mechanism that gives rise to adaptive complexity in language. In essence, language is an adaptive system in its own right.

One intriguing result of this theoretical move in evolutionary linguistics is the blurring of the divide between language evolution, and language change. In this talk I will argue that the exact same processes proposed in work on the cultural evolution of language are the drivers of language change. Fundamentally, language is constantly under "transmission pressures" arising directly from the fact it persists over time only by being learned and used by generations of speakers. These pressures - to be learnable and expressive - existed at the point when language first emerged, and they exist still today. The challenge for researchers is to understand what structures emerge from the adaptive process given different starting points (e.g. a protolanguage or a fully modern human language). I will propose a way forward that relies on a combination of computational models and experiments using human participants.