Morphology is sometimes painted as the ‘here be dragons’ of the linguistic map: a baffling domain of idiosyncrasies and irregularities, in which Heath Robinson contraptions abound and anything goes. In his new book, Universals of Comparative Morphology: Suppletion, Superlatives, and the Structure of Words (MIT Press, 2012), Jonathan Bobaljik reassesses the terrain, and argues that there are hard limits on the extent to which languages can vary in the morphological domain.

The book is a comparative study of comparatives and superlatives with a broad typological base. Bobaljik’s contention is that, at an abstract cognitive level, the representation of the comparative is contained within that of the superlative. From this hypothesis, couched within the theoretical framework of Distributed Morphology, a number of generalizations immediately follow: for instance, in a language which, like English, has forms of the type “good” and “better”, the superlative cannot be of the type “goodest”. As he shows, these generalizations are solid candidates for the status of exceptionless linguistic universals.

In this interview, Jonathan outlines the generalizations and their evidential basis, and we go on to discuss apparent counterexamples (including the mysterious Karelian quantifiers), why the comparative should be contained within the superlative, how the generalizations extend to change-of-state verbs, and how similar generalizations can be found in domains as diverse as verbal person marking and pronominal case.