The internal structure of compounds: a phase account of aphasia

Abstract

This study uses aphasia to support a phase-based derivation of compounds. Our research is nestled within the overarching and truly foundational debate between holists (Butterworth 1983, Bybee 2001, Starosta to appear) and atomists (Taft and Forster 1975, Rastle et al. 2004, Fiorentino and Poeppel 2007). The former camp maintains that compounds are stored devoid of any internal morphological structure; while the latter insist that compounds are derived by concatenation of constituent parts. Morpho-phonological analysis of the contrasting behaviour of simplex and compound words in Dinka and English (based on Kaye 1995) bears a striking similarity to the derivation by phase (Chomsky 2001) (cf. Newell and Piggott 2006, Newell and Scheer 2008, Scheer 2008, forth.). To confirm this novel phase-based account, contra the holists’ null-hypothesis, we ran an experiment. We tested an aphasic patient (RC), who produced high error rates with trisyllabic simplex words and negligible error rates with disyllabic simplex words. The divisive question: What would trisyllabic compounds pattern with? The surface inclined holists predict they should pattern with the long simplex words; conversely, the atomist, for whom a trisyllabic compound will be processed either [[σ σ] [σ]] or [[σ] [σ σ]], predict they should pattern with the short simplex words. The latter turns out to be correct. Our experiment shows a compound is derived by independently sending its constituent parts to spell out, once there the constituent parts are no longer accessible to grammatical operations.