Formal and computational models of language evolution

Gerhard Jäger

[revised and updated September 16, 2006 ]



The term "language evolution" is used in two different ways in the literature:
  • the biological evolution of the human language faculty
  • the cultural evolution of human languages
The course will be concerned with the second notion. Natural languages can be considered as evolving entities, similar to species in biology. Accordingly, language change can be conceptualized as an evolutionary process. Linguistic universals are features of languages that are invariant in evolutionarily stable states.

We will study formal models of language evolution in this sense. Evolutionary Game Theory (EGT) -- a mathematical framework that emerged from collaboration between biologists and economists -- has proven a useful metatheory for this kind of endeavor. The first half of the course will consist of a crash course of EGT. The second part of the course deals with various linguistic applications, ranging from phonology via morphology and syntax to semantics. The investigations make heavy use of computer simulations, and the course will also deal with practical implementation issues.


To take the course for credit, it is required to attend classes regularly and to write an essay of 6-8 pages. Possible topics are

  • Equilibrium notions: Explain the concepts "Nash equilibrium", "Strict Nash equilibrium", "Evolutionarily stable strategy" and "Stochastic stability". How are they justified, and under what conditions can they be applied? [Literature: lecture slides, Jäger 2004]
  • Vowel systems: What is a vowel (in terms of acoustic phonetics), what are universal tendencies in the typological distribution of vowel systems, how can these tendencies be explained in an evolutionary model? [Literature: lecture slides, Jäger 2006: Spiel mit Lauten]
  • Case marking systems: What are accusative systems and ergative systems, what are universal tendencies in the typology of case marking systems, how are they derived in Jäger (in press), what is the empirical basis of the utility function used in that paper.
  • Horn strategies: What is a Horn strategy; what are linguistic examples for Horn strategies; what are the Nash equilibria of the Horn game; which strategy combinations are evolutionarily stable; how can the dominance of the Horn strategy over the anti-Horn strategy be explained [Literature: lecture slides, Jäger 2004]
The essays should be turned in via email to Gerhard Jäger no later than October 31st.