Rethinking classifier languages

Classifier languages have long been of interest to formal semanticists. That such languages do not allow nouns to be modified directly by numerals have lead researchers to propose a significant degree of cross-linguistic variation between classifier and non-classifier languages in either noun meaning (e.g. Chierchia 1998) or numeral meaning (e.g. Krifka 1989, Bale & Coon 2014). Classifier languages, on this view, show fundamental semantic differences from non-classifier languages. In this talk, I challenge this notion by examining the behavior of nouns and numerals in Tiwa, a Tibeto-Burman language of India. I argue that despite having the properties of a typical classifier language, nouns and numerals in Tiwa can and should be analyzed as having the same sorts of meanings as nouns and numerals in non-classifier languages like English. I argue that the differences between languages like Tiwa and English are largely morphosyntactic, and that this morphosyntactic variation is expected given standard assumptions about the syntax and semantics of nouns, numerals, and number marking. Most centrally, I argue that (sortal) classifiers in Tiwa are overt realizations of a morpheme that creates a predicate out of a degree-denoting numeral; a function which is often assumed to be a type-shift in languages like English (Bylinina & Nouwen 2020). What seem like major differences between the two languages, then, does not reflect fundamental semantic variation in noun or numeral meaning, but is exactly the sort of morphosyntactic variation we expect to see. I conclude by looking at the typological picture more broadly, examining the ways languages are predicted to differ based on the dimensions of variation I identify.

Location: Royce Hall 362