A hallmark of language fluency is the ability to decode words as they are encountered in speech and print. The available research shows that morphological awareness and knowledge is a key factor in the ability to access and process complex lexical items. Since Taft and Forster’s (1975) seminal article on the principle of Obligatory Word Decomposition (OWD) whereby all complex words are analyzed in terms of their constitutes (i.e. affixes and stems), an entire line of research has emerged to test the validity of such claim. Alternative studies, have claimed the opposite; i.e. that complex words are recognized wholly without the need to fully decompose them (Butterworth, 1983). The present study seeks to fill in a critical gap in our understanding of how complex verbs in Arabic are handled in reading by engaging this question from both a first and a second language perspective. In the present study, I report on a visual masked priming experiment from two main populations: native speakers of Arabic (n=12) and advanced learners of Arabic (n=12). The experiment consisted of a lexical decision task where participants judged the lexicality of 192 items. The reaction times and error rates were analyzed using a By-subject and By-item ANOVAs. In particular, this study investigates whether affixed words in Arabic are processed in the same way when the affix is real versus when it’s part of the stem by manipulating the type of the attached affix (i.e. real vs. pseudo) across three verbal patterns in Arabic (i.e. VII, VIII, X). By parametrically varying the morphological affiliation of consonants /N/ and /S/, such that in half of the items they belonged to the root and in the other half they belonged to the verbal templatic pattern, this study seeks to measure the relative contribution of morphology during lexical access as well as the time course of activation for root morpheme and word-pattern morphemes in question. Preliminary results suggest that only certain categories of words are necessarily decomposed as suggested by the differential reaction times across verbal categories, supporting a hybrid account of lexical access in which some words are listed wholly while others are decomposed prior to lexical access. Results from second language learners of Arabic show similar patterns and provide evidence against the claim that second language processing is characterized by shallow processing as hypothesized by Clahsen and Felser (2006).