“…Rare words emerge in spontaneous speech because frequency has not been encoded into the lexicon in the normal way (for example, because the base-rate activation of lexical entries has not been set at a level that appropriately reflects their frequency of usage; see e.g., McClelland & Rumelhart, 1981); or because the lexical retrieval process has developed inadequately to the extent that it mistakenly retrieves not-quite-appropriate, lower-frequency words in a given semantic context. Proposals of this sort include the following claims: (1) that word retrieval is deviant in WS ; (2) that while word knowledge may be well organised, inhibitory activation dynamics which integrate current contextual information are abnormal in lexical access (Rossen et al, 1996); (3) that lexical access is fast but sloppy in WS, with inadequately specified semantic representations (Temple, Almazan, & Sherwood, 2002); and more precisely, (4) that children with WS cannot access fine-grained semantic features of lexical items (Clahsen, Ring, & Temple, 2003). Claims of domain-specific atypicalities in the WS lexicon are often associated with a broader perspective of WS as a developmental fractionation of the normal cognitive-and in this case-language system, whereby the WS system is viewed in terms of the architecture of the normal system but with selective components of the system under-or over-developed (Temple & Clahsen, 2002;see Karmiloff-Smith, 1998, andKarmiloffSmith, 2002, for discussion).…”