The purpose of this study was to investigate how second-language (L2) learners lexically encode confusable phonemes. Given the inconsistency of previous studies on whether and if so how learners can establish separate lexical representations of confusable categories, we examined (1) how phonetic categorization and lexical encoding abilities were developed at the early stage of learning and (2) whether there are any differences in those abilities between the words with a sound pair from a corresponding native language (L1)-dominant category and those lacking such category. Native speakers of Korean learned Arabic words with these two types of sound pairs for four days and then their phonetic categorization and lexical processing abilities were evaluated in AXB discrimination and lexical decision tasks, respectively. The results showed that phonetic categorization of the words with a sound pair from an L1-dominant category developed very early. With success in their discrimination abilities, L2 learners began to overcome lexical competition from the words with such a sound pair. By contrast, learners showed poor sound discrimination and lexical encoding skills for words with a sound pair lacking an L1-dominant category. This suggests that (1) L2 learners' accurate phonetic categorization abilities are prerequisite to success in L2 lexical encoding and (2) lexical representations of the L2 words with confusable phonemes depend on the distinct types of sound category matchup between L1 and L2.
Keywords: AXB discrimination task; Korean learners of Modern Standard Arabic; Phonetic and lexical learning; lexical decision task.