Processing Cantonese light verb constructions and compounds : a masked priming study

Abstract

An important question in psycholinguistics is whether the semantic transparency of morphemes plays any role in de-compositional processing of compound words. Studies from different languages produce mixed results suggesting that the surface features of a language can produce differential effects of semantic transparency on word processing. The goal of this study was to investigate the possible effects of semantic transparency on processing of Cantonese light verb constructions (LVCs) compared with verb noun compounds (VNCs) using the masked priming paradigm. We predicted greater processing demands in LVCs in comparison with compounds due to non-canonical argument sharing and semantic under-specification of the verb in LVCs and we assumed semantic transparency would have an effect in character recognition. Native Cantonese speakers aged between 18 and 30 years were recruited for a fully within participants experiment varying semantic transparency, word type (opaque LVCs, transparent LVCs, opaque VNCs and transparent VNCs) and target presentation with two durations: 33 msec and 100 msec. Linear Mixed Effects (LME) modelling found a significant masked priming effect in semantically opaque and semantically transparent compounds and LVCs. We contend that semantic transparency is not an influential factor in Cantonese word processing and LVCs are processed like other compounds.