Analysis and falsifiability in practice: A commentary on Haspelmath’s ‘General Linguistics must be based on Language Universals’
Adam Tallman
February 2021

This paper is a commentary on Haspelmath’s “General Linguistics must be based on Language Universals”. Some practitioners in the natural-kinds programme argue that theoretical models developed and fit to data of one language make testable predictions about the next. I argue that this is not obviously true in cases where categories and relations presupposed by the hypothesis do not map deterministically onto the next language. While such a theoretical model can be fit, hypotheses couched in the model cannot be meaningfully tested. While I agree with Haspelmath’s paper in general, I suggest that the distinction between the Boasian/Greenbergian and natural-kinds programme may describe a normative tendency rather than a necessary division in the field.
Format: [ pdf ]
Reference: lingbuzz/005720
(please use that when you cite this article)
Published in: Theoretical Linguistics
keywords: philosophy of science, theoretical linguistics, morphosyntax, morphology, syntax, phonology
previous versions: v1 [January 2021]
Downloaded:559 times


[ edit this article | back to article list ]