Morphological Structure in the Arabic Mental Lexicon: Productivity and Priming Effects in Nominal and Verbal Patterns.

Abstract Semitic languages, including Arabic, are characterized by two types of discontinuous morphemes: roots and word patterns. However, the role of these morphemes in lexical access and representation remains debated. Roots have been found to exhibit robust priming effects across both nominal and verbal patterns. In contrast, word patterns have shown mixed results, with verbal patterns consistently demonstrating stronger priming effects than nominal patterns. While previous research, such as Deutsch et al. (1998), suggested that productivity might explain these differences, no study has directly investigated the role of productivity in morphological priming by Arabic word patterns. This dissertation addresses this gap by examining whether documented differences between priming between nominal and verbal word patterns are due to differences in productivity and whether morphological priming in word patterns can be distinguished from the effects of overlap in semantics and phonology. A total of 583 native Arabic speakers participated in five experiments, including cross-modal priming and masked visual priming tasks. Stimulus onset asynchronies (SOAs) of 32 ms, 42 ms, 48 ms, and 65 ms were varied to investigate the time course of lexical access in word patterns. Stimuli included high- and low-productivity nouns and verbs, compared with phonological, semantic, and unrelated controls.  The results showed that productivity plays a critical role in explaining the differential priming effects of nominal and verbal patterns. High-productivity word patterns, both nominal and verbal, exhibited robust priming effects, while low-productivity patterns did not, offering new insights into the interaction between productivity and morphological decomposition in Arabic. These findings support the dual-route model of morphological processing, highlighting that high-productivity patterns are accessed via decomposition, while low-productivity patterns rely on whole-word storage. Additionally, the findings showed that word pattern priming effects were found only at longer SOAs (42 ms, 48 ms, and 65 ms) but not at 32 ms, emphasizing that word patterns activate morphemic representations after roots. Furthermore, high-productivity patterns were distinguished from form and meaning overlap, providing clear evidence for their independent role in morphological decomposition.