Proceedings of the Annual Meeting of the Cognitive Science Society

Despite wide variation among natural languages, there are lin-guistic properties universal to all (or nearly all) languages. Animportant challenge is to explain why these linguistic universalshold. One explanation employs a learnability argument: seman-tic universals hold because expressions that satisfy them areeasier to learn than those that do not. In an exploratory studywe investigate the relation between learnability and complexityand whether the presence of semantic universals for quantifierscan also be explained by differences in complexity. We developa novel application of (approximate) Kolmogorov complexityto measure fine-grained distinctions in complexity between dif-ferent quantifiers. Our results indicate that the monotonicityuniversal can be explained by complexity while the conserva-tivity universal cannot. For quantity we did not find a robustresult. We also found that learnability and complexity patterntogether in the monotonicity and conservativity cases that weconsider, while that pattern is less robust in the quantity cases.