Accueil > Archives > Archives 2011

Conf. S. Kirby (U. Edimbourg) "Why language has structure..."

par Thierry POIBEAU - publié le

Conférence organisée par le laboratoire LATTICE-CNRS (,
en collaboration avec les labex TransferS ( et EFL (

université d’Edimbourg

Titre : "Why Language Has Structure :
New evidence from studying cultural evolution in the lab, and what it means for biological evolution"

Jeudi 8 Décembre 2011
Ecole Normale Supérieure
45 rue d’Ulm 75005 Paris
Salle des Actes (escalier A, 1er étage)

Simon Kirby

Simon Kirby is Professor of Language Evolution at the University of Edinburgh and cofounder of the Language Evolution and Computation Research Unit. He has pioneered the application of computational, mathematical and experimental modelling techniques to traditional issues in language acquisition, change and evolution. In particular, he has developed an approach to cultural evolution called Iterated Learning which treats language as a complex adaptive system operating on multiple interacting time-scales. His view is that a complete understanding of human nature requires an account of the complex interactions between individual learning, cultural transmission and biological evolution in human populations. In addition, he pairs his scientific research with artistic output by collaborating with sculptors and musicians to create interactive installations exploring issues of communication and cultural evolution in a socially and informationally promiscuous world.

Abstract :

Where do the characteristic design features of human language come from ? In particular, how do we come to have a language that allows us to express novel utterances and have them reliably be understood ? One answer is that this highly adaptive trait is simply an innately encoded feature of our biological endowment, tuned by natural selection under pressure for successful communication (e.g. Pinker & Bloom, 1990). In recent years, however, an alternative view has been set out that suggests language adapts not through a process of gradual biological evolution, but rather as a result of cultural evolution as it is transmitted in a population through repeated learning and use (e.g., Kirby, Dowman, & Griffiths, 2007). This process, known as iterated learning, can be studied in the lab by creating artificial languages and observing how they evolve as they are acquired and transmitted by experimental participants (Kirby, Cornish, & Smith, 2008).

In this talk, I will show that productive linguistic structure only emerges in these artificially evolving languages in certain specific conditions. These shed light on where the familiar design features of human language come from, and explain some emerging data from real languages in unusual circumstances. I con- clude that language structure is an emergent compromise between being compressible andexpressive, driven by pressures from learning and communication respectively.

These results have important implications for our understanding of the basic design features of human language (Hockett, 1960). I will argue that many of these design features arise inevitably as a result of two fundamental ones : semanticity and traditional transmission. The cognitive support for these two features represent the crucial biological pre-adaptations for language. At the end of the talk I will give a tentative suggestion for how and why these evolved, setting the stage for the inevitable emergence of structured language.