Karlos Arregi (University of Chicago): "Basque plural clitics: A case study in Crossmodular Parallelism"

12/09/2011
Sala Caro Baroja (1D8/1D23)

According to the Crossmodular Parallelism thesis, principles and rules are shared across modules, and differences among the latter are mostly due to the fact that each is based on a different basic alphabet. This talk argues for this thesis in a particular area of grammar: operations on abstract morphological structures are the same as ones that operate on phonological representations. Under this view, phonology and (part of) morphology share the same abstract grammar, but operate on different alphabets: phonological features and segments vs. abstract morphological features and morphemes. I illustrate the need for this view of morphology (and phonology) with plural
inflection in Basque finite verbs.

In joint work with Andrew Nevins, we propose a modular analysis of a particular plural morpheme (-e/-te) with two basic components: (i) generation and feature makeup, and (ii) placement within the word. With respect to (i), this morpheme is the result of Fission: pronominal clitics in finite verbs undergo a splitting operation that separates certain person and number features. Fission has non-trivial similarities with certain diphthongization phenomena, which provides evidence for an implementation of Fission based on Crossmodular Parallelism. Adopting ideas about diphthongization from Calabrese (2005), I argue that Fission is a repair operation triggered by a constraint that prevents the expression of certain person and number features in the same morpheme.

With respect to (ii), I argue that placement of the plural clitic -e/-te must be dealt with in a separate Linearization component. Specifically, a study of variation in the Bizkaian Basque dialectal area reveals that a correct account of the placement of this morpheme must take into account abstract displacement and copying operations that have important parallels with phonological metathesis and reduplication, once again providing evidence for the Crossmodular Parallelism thesis. The analysis is based on an extension of the formalism proposed in Halle & Harris 2005 for this type of operation.