Pseudoword and Root Variation Effects on Lexical Decisions: A Study of Visual Word Recognition Processes of L2 Learners and L1 Speakers of Arabic

Abstract

A major question addressed in visual word recognition research concerns the “decomposition” of morphologically complex words into smaller parts. For example, do English readers process a word (e.g., unsafe) as a whole word, or is the word decomposed with the stem (safe) used as the access for locating the derivations (e.g., unsafe, safely, safety) in the mental lexicon? The nonconcatenative root and pattern structure of Arabic words provides a unique perspective for word-recognition testing. Previous findings with Semitic languages provide strong evidence that the mental lexicon in Arabic is organized differently than English, with the triconsonantal Arabic root serving as the organizational unit. The current study expands on previous research by examining visual word-recognition performance of beginner Arabic second-language (L2) learners, advanced Arabic L2 learners, and Arabic first-language (L1) speakers. The study uses invented, derived pseudowords that vary in root familiarity to determine whether 1) pseudoword primes with real roots facilitate real-world target recognition (at an SOA of 70ms); 2) pseudoword targets with real roots interfere with response times and accuracy; and 3) phonological contrasts in Arabic that are not present in English affect response accuracy. The study findings show that beginner L2 Arabic learners demonstrate morphological decomposition and sensitivity to real and nonexistent roots in lexical decisions with pseudoword targets. Moreover, a comparison between beginners with little or no Arabic background and beginners who grew up speaking an Arabic dialect at home reveals interesting findings. With response times, the “true” beginners respond consistently faster than the “household speakers,” in addition to showing temporal sensitivity to the root variations. On the other hand, the “household speakers” show greater sensitivity to root variations with their decisions. Advanced L2 learners demonstrate word-priming effects with pseudoword primes sharing the same root as real-word targets. These results are consistent with L1 participants and imply that – at 70ms of exposure – word-level semantic meaning is not required for readers (advanced L2 or L1) to decompose root letters from the prime, access the root entry in their mental lexicon, and facilitate recognition of the target word. For pseudoword targets, advanced L2 learners display a significantly more profound interference effect than both the beginner L2 and L1 speaker groups. The accuracy difference is evidence that advanced L2 learners rely heavily on Arabic roots to recognize words. A further difference in performance comes from the phonological contrast condition. One letter from an otherwise real Arabic root is changed to reflect a phonological contrast that exists in Arabic but not English (e.g., pharyngealized vs. non-pharyngealized consonants). For beginner L2 learners, roots in the phonological conditions are treated as if the roots are the real alternatives. Advanced L2 learners are continuing to struggle with the phonological contrast in certain lexical contexts. However, L1 participants treat the phonological condition roots as if the roots are nonsense. Therefore, the visual word-recognition task reveals the challenges L2 learners face with acquiring Arabic phonological contrasts that are not present in English. Findings from the current study are discussed in relation to previous research and theories in the fields of psycholinguistics and second language acquisition. Several pedagogical implications for the study are presented, including attention paid to root awareness and morphological structures in first-year Arabic courses, use of root-based dictionaries at all learner levels, and the importance of comprehensible reading input and vocabulary testing.