Orchestration of Multiple Codes in Developing Readers: An Alternative Model of Lexical Access

Abstract

An alternative model of lexical access, based on recent PET scan studies showing localized coding operations in multiple brain sites, is proposed. Lexical access is viewed as a process of orchestrating multiple connections between corresponding visible language and oral language coding operations instead of a place--the internal lexicon--to which stimulus words gain access directly or indirectly. In both Study 1 (end of 1st grade) and Study 2 (mid 2nd and mid 4th grade) speed of naming phonically regular words, in which every letter can be phonemically coded, and of naming phonically irregular words, in which every letter cannot be phonemically coded, was not significantly different, indicating that naming can be based on a phonetic code connection rather than phonemic code connections. In both Study 1 and Study 2 speed of correctly accepting phonically regular real words was significantly faster than speed of correctly rejecting nonsense words on lexical decision, showing that absence of a semantic code is not detected as quickly as presence of a semantic code is. In both Study 1 and Study 2 speed of naming a phonically regular real word was significantly faster than speed of naming a nonsense word, demonstrating that naming is facilitated when two connections--between a printed word and its phonetic code and between a printed word and its semantic code--are activated. These findings held whether a key press (Study 1) or voice key (Study 2) response was used. Taken together, the results suggest that the whole word-phonetic code connection may be completed before the whole word-semantic code connection, both of which may precede the letter-phoneme connection in time.