Lexical retrieval after Arabic aphasia: Syntactic access and predictors of spoken naming

Research into anomia has been carried out in English and many Indo-European languages extensively, but not in Arabic. Previous studies have investigated predictors of successful lexical retrieval after anomia, and access to syntax during lexical retrieval. The aim of the current study is to examine impaired lexical retrieval in Arabic at two levels: predictors of lexical retrieval, and access to syntax during lexical retrieval, via checking whether syntactic cueing (using the definite article/əl-/'the' prior to nouns) facilitates noun retrieval in Arabic aphasia, with regard to naming speed and accuracy, and establishing the determinants of aphasic noun retrieval in Arabic. Three participants with anomia following CVA named 186 pictures from a published Arabic database in two conditions: bare noun condition, and determiner + noun condition. Participants' accuracy and reaction times were compared in both conditions. Furthermore, a multiple regression analysis was carried out to test the effect of psycholinguistic variables (visual complexity, name agreement, age of acquisition, imageability and other intrinsic variables) on successful lexical retrieval to determine predictors of Arabic noun retrieval after anomia. The production of the determiner + noun in picture naming facilitated spoken naming in all three participants. Nouns produced with the determiner were produced faster and more accurately than their counterparts produced without the determiner. The two participants with agrammatism produced morpho-syntactic errors in the bare noun condition, but not in the determiner + noun condition, suggesting that the determiner sets up a noun phrase frame with a slot for the noun to be filled, resulting in responses that are faster and more accurate. Age of acquisition and imageability were the only two variables that had influence across the participants. These results have theoretical and clinical implications for lexical retrieval models.