The present research explores the degree of morphological structure of compound words in the native and nonnative lexicons, and provides additional data on the access to these representations. Native and nonnative speakers (L1 Spanish) of English were tested using a lexical decision task with masked priming of the compound's constituents in isolation, including two orthographic conditions to control for a potential orthographic locus of effects. Both groups displayed reliable priming effects, unmediated by semantics, for the morphological but not the orthographic conditions as compared to an unrelated baseline. Results contribute further evidence of morphological structure in the lexicon of native speakers, and suggest that lexical representation and access in a second language are qualitatively comparable at relatively advanced levels of proficiency.Keywords: lexical representation; non-native speakers; compounds; morphological processing; masked priming.
IntroductionAmong the considerable issues in visual word recognition research, the role of morphology is hotly debated. While many agree that, at least in some cases, native visual word processing involves morphological decomposition for complex words (e.g., Marslen-Wilson, 2007), there is no consensus as to exactly how, when, and under what conditions this happens. As with any other aspect of language processing, access is crucially interesting for psycholinguists in that it is inextricably bound to representation.In fact, the existence of-or even the need for-a morphologically structured lexicon has been the subject of much discussion (see, e.g., Feldman, O'Connor, & Moscoso del Prado Martín, 2009). Focused on the visual processing of English compound words in isolation, the data reported in this article provide supporting evidence for models assuming some degree of morphological structure in the lexicon of native speakers, and suggests that lexical representations in relatively proficient second language users are equally complex.The role constituent morphemes play in the access to, and representation of, complex words has been explored systematically for the last 40 years, from the seminal studies of Forster, 1975, 1976) to the proliferation of research contrasting behavioural, electrophysiological, neuroimaging and clinical data (e.g., Bozic, Marslen-Wilson, Stamatakis, Davis, & Tyler, 2007;Fiorentino, Naito-billen, Bost, & Fund-Reznicek, 2014;MacGregor & Shtyrov, 2013;Marelli, Zonca, Contardi, & Luzzatti, 2014; Marslen-Wilson & Tyler, 2007, among others). Current positions range from models advocating exclusively morpheme mediated access (Stockall & Marantz, 2006;Taft & Nillsen, 2013), to dual/multiple-route accounts highlighting the competition (and/or cooperation) between whole-form and morpheme-based access (Hyönä, 2012), to proposals questioning the role of morphology in favour of a distributed-connectionist implementation of unmediated access to constituent and whole-word meanings through orthographic and/or phonological cues (e.g., Devlin, Jami...