Unpacking the comparison of l1 and l2 glosses in vocabulary learning from reading

The appropriate amount of first language (L1) and second language (L2) to use in L2 learning has been constantly debated (e.g., Cummins, 2007; Hall & Cook, 2012). This study situates the debate of L1 and L2 use in the context of vocabulary learning from reading. By examining the potential moderating factors on the comparison of L1 and L2 glosses (i.e., short word definitions provided during reading), the study aims to provide a nuanced picture of how L1 and L2 input affects vocabulary learning in various circumstances. Investigating L1 and L2 glosses in the context of vocabulary learning also allows the study to contribute to the theories of bilingual lexicon (e.g., Kroll & Stuart, 1994; Jiang, 2000), i.e., the development of the bilingual lexicon as a function of input language.One hundred and eighteen L2 learners of English completed the study. Participants first read part of a graded reader, where 24 target words were embedded. Glosses for the target words were inserted through hyperlinks: participants could click the target words to access their glosses, written either in the participants' L1 or L2. Participants' time spent on reading each gloss was tracked. After reading, participants went through unannounced vocabulary posttests that measured receptive and productive meaning knowledge, and lexical retrieval fluency of the target words. Participants also filled in an exit questionnaire that aimed to further probe their reading and gloss access behaviors. Participants' gloss reading time, their vocabulary size, and the target words' frequency of occurrence (FoO) were analyzed as moderating variables on the comparison of L1 and L2 glosses.Findings revealed that L1 and L2 glosses benefitted the learning of different aspects of vocabulary. The comparative effects of L1 and L2 glosses were primarily moderated by participants' gloss reading time and target word FoO, suggesting that the initial depth of processing and subsequent memory reactivation were the keys to vocabulary learning. Results have pedagogical implications for how to choose the language for glosses, theoretical implications on the bilingual lexicon development, and methodological implications of using hyperlinks to track behaviors during learning.