Reading morphologically complex words: Some thoughts from masked priming

Much research suggests that words composed of more than one morpheme (e.g, departure) are represented in a "decomposed" manner in the visual word recognition system, with morphologically complex words sharing representations with their stems (e.g., Rastle, Davis, Marslen-Wilson, & Tyler, 2000). In this chapter, we consider the extent to which semantic relationships influence morphological decomposition, especially with respect to those representations contacted in early visual word recognition. In two studies of visual lexical decision, we found that the recognition of stem targets (e.g., depart) was facilitated significantly and equivalently by the prior presentation of semantically transparent (e.g., departure) and semantically opaque (e.g., department) masked primes (using a 52-ms SOA). We found further that the recognition of stem targets (e.g., broth) was faster numerically when these targets were preceded by a morphemically structured semantically opaque masked prime (e.g., brother) than by a nonmorphemically structured masked prime (e.g., brothel). We believe that these results implicate the operation of a purely structural morphological segmentation system in early visual word recognition, which may enable the developing reader to capitalize upon higher-level regularities that morphology provides to the mapping between orthography and meaning (e.g., Plaut & Gonnerman, 2000).