Dr Anna Kinsella

Dr Anna Kinsella is a British Academy Postdoctoral Fellow based in the Language Evolution and Computation Research Unit in the School of Philosophy, Psychology and Language Sciences at the University of Edinburgh. She is working on why language is the way it is and why only humans have language. Her research is examining both the biological and cultural evolutionary processes impacting on these questions, taking as a starting point a picture of language as a system of form-meaning mappings ranging from completely general grammatical rules to completely idiosyncratic lexical items. She is investigating the biological evolutionary plausibility of such a view by determining how close to typically evolving natural systems it is, while its cultural evolutionary plausibility is investigated using iterated learning experiments which explore the social transmission of language.

Dr Kinsella presented a poster entitled Towards an Evolutionarily Plausible Theory of Language at the 2009 British Academy Postdoctoral Symposium

TOWARDS AN EVOLUTIONARILY PLAUSIBLE THEORY OF LANGUAGE

In linguistics, theory is underdetermined by data. Consequently, theories of language are constructed with some constraining factor in mind. Here, I examine the implications for linguistic theorising of taking an evolutionary perspective. What would our theory of language look like if the constraining factor we adopt is evolvability?

Constraints on language theorising

  • Learnability
  • Parsability
  • Neurological plausibility
  • Cross-linguistic variation
  • Evolvability

The constraint of perfection

The Minimalist Program (Chomsky 1995) employs the constraint of perfection, but this rules out evolvability, as evolution more often leads to imperfection:

In nature

  • Mammalian recurrent laryngeal nerve
  • Human spine
  • Human appendix

In cognition (Marcus 2008)

  • Distortion of memories
  • Warping of objective reasoning

In language

  • Garden-path sentences
  • Ambiguity
  • Synonymy

An alternative approach

The Simpler Syntax framework (Culicover & Jackendoff 2005) rejects the problematic constraint of perfection, but is fully compatible with the constraint of evolvability. The properties assumed in this approach to language are precisely those associated with evolvability:

  • Redundancy
    • in lexical storage
  • Robustness
    • in the network structure of constructions
  • Modularity
    • in the parallel architecture (Jackendoff 1998, 2002)
  • Competing motivations
    • communication, acquisition, brain storage…
  • Gradual adaptive evolution
    • Toolkit hypothesis

Conclusion

Simpler Syntax marks the beginning of an exciting direction for evolutionarily plausible linguistic theorising.