Algorithmic and subjective measures of lexical diversity in bilingual written corpora: a discussion

Abstract

Lexical development plays an important role in L2 acquisition/learning and has therefore been widely investigated, especially with regard to the lexical diversity of texts produced by L2 learners; as a result, several indices have been created to measure this feature. Nevertheless, L2 learner production, especially when children are concerned, is frequently relatively limited in scope, an aspect that makes it difficult to measure their lexical diversity. The aim of the study presented in this article is to discuss the applicability of several measures of lexical diversity on small texts samples (two algorithmic measures [HD-D and MTLD] as well as subjective ratings by untrained raters). The corpus comprises written productions from 105 sixth-grade Portuguese immigrants in the French and German-speaking parts of Switzerland. The results enable a deeper understanding of the very notion of lexical diversity and ways of measuring it.

Keywords

• French
• lexical diversity
• German
• subjective ratings
• HD-D
• MTLD