摘要：The debate as to whether complex words (e.g., dealer, basketball) are processed as undivided wholes or parsed into their constituents is still vivid in the neuropsychological literature (Marelli et al., 2012). In fact, this conundrum is shared with the literature focusing on unimpaired speakers/readers, where data from masked priming experiments indicate that complex words are always decomposed into morphemes during visual identification, even when they are only apparently complex (e.g., comer\ Longtin et aL, 2003). Recent evidence, however, has shown that morpho-orthographic effects may not occur in tasks other than lexical decision (Dunabeitia et al., 2011), thus undermining the obligatoriness of morpho-orthographic segmentation. Duftabeitia et al. (2011) used a “same-different” task, which only requires comparing two letter strings, with no need for lexical analysis. It is thus still possible that a pre-lexical, semantically-blind decomposition of complex words is obligatory for lexical access, which is what we have tested in the present work.
