摘要：Lexical retrieval is a multi-level process, which consists of two main lexical levels: semantic -word meaning, and phonological - word form. One of the intriguing questions of lexical processingis how homophones are represented in the model. Homophone is a phonological word-form that isshared by more than one semantic meaning (e.g., palm: hand; tree). The different meanings haveseparate semantic representations, but the question is whether they have also separate phonologicalrepresentations (Caramazza , 1997) or a shared phonological representation (Levelt et al., 1999).Biedermann and colleagues (2002, 2008a,b) explored this question, in German and in English, byconducting a treatment study for 3 individuals with aphasia. In their studies they treated onemeaning of the homophonic word and examined the improvement for treated and untreated items.They found an improvement for treated items, with generalization to the untreated meaning of thehomophone, but with no generalization to phonological and semantic related control words. Thesefindings support a shared phonological representation. In the current study we examined the samequestion in Hebrew.
