Lexical input and categorization:: a study of vocabulary depth in second language learning

Abstract

This study investigates the association of vocabulary development and reading comprehension, focusing principally on the contribution of context to vocabulary depth learning. Research into reading for vocabulary learning has typically emphasized the learning of discrete items as though they were unrelated. Adopting the viewpoint of cognitive linguistics, this study assumes a cognitively motivated relationship inherent in these to-be-learned items, and thus defines vocabulary depth in terms of the ability to recognize the inclusion and membership properties of semantic categories occurring in a text. In other words, what distinguishes this study from others is its emphasis on extent to which learners are able to recognize the hierarchical relations (inclusion) and internal category structure (membership) of the lexical items found in a text. The research was conducted in a real classroom setting using procedures that are pedagogically valid. This is believed to be the optimal context for revealing the genuine nature of second language learning. Specifically, a Freshman English course incorporating extensive reading as one of its requirements was selected for this investigation since successful reading at advanced level depends upon recovering the instantial categories (i.e. those that exist in some particular instance) established in the text. To compare the effects of the original and alternate encoding options, a selected text was modified structurally and paratactically so as to enable a comparison of the extent to which text structure is associated with comprehension, when comprehension is defined in terms of the ability to recognize inclusion and membership. In addition, the rhetorical properties present in the original text were also taken into account in order to investigate whether common rhetorical properties could result in more successful recognition of category membership and whether distinct rhetorical properties could lead to more successful differentiation of degrees of category membership.