lexical sophistication

The literature on second/foreign language (L2/FL) discourse is replete with corpus-based studies into the use of various features representing lexical proficiency. Nonetheless, the lexical construct of English for academic purposes (EAP) texts developed by postgraduates majoring in teaching English as a foreign language (TEFL) still sounds like a relatively unexplored domain that merits further multi-dimensional investigation. To narrow the gap, the authors in the current study set out to evaluate the lexical richness of a corpus containing doctoral dissertations written by Iranian TEFL students in terms of lexical density, diversity, and sophistication. Taking advantage of the computational tool Coh-Metrix to analyze the lexical features, the corpus was analyzed in comparison with a first language (L1) baseline containing doctoral dissertations written by English native speakers. The comparative analysis of the L1 and L2 corpora revealed that the texts written by Iranian TEFL learners were lexically less diverse but more sophisticated. Additionally, the lexical density of the L2 corpus exceeded that of the L1 one in terms of nouns and adjectives. Based on the results drawn from a discriminant function analysis (DFA), the features representing lexical sophistication and density were found to be the best predictors of lexical richness since they could significantly discriminate between the two sub-corpora. The findings may provide new insights into the ways of evaluating and enhancing the lexical richness of FL/L2 written discourse.

EFL learners’ ability to use advanced and varied vocabulary is a crucial issue in writing performance. This study aimed to investigate the effect of writing task type on lexical diversity and sophistication. The relationship between lexical sophistication and lexical diversity in the narrative, argumentative, and descriptive task types written by upper-intermediate EFL learners was also explored. To this end, 70 EFL upper -intermediate learners, enrolled in advanced writing tasks, were selected as the participants. They undertook narrative, descriptive, and argumentative writing tasks in a counterbalanced way. Then, Vocabulary Diversity (Voc-D) and Measure of Textual Lexical Diversity (MTLD) were used to measure lexical diversity. Lexical sophistication was measured using CELEX log frequency and Beyond 2000 Words (K3+). The writings were pasted into Coh-Metrix measure MTLD, Voc-D, and CELEX. VocabProfiler was also used to measure K3+ of the writings. In lieu of the purposes of the study, the research design involved repeated measures and correlation. Using Multivariate Analysis of Variance and Pearson’s Correlation Coefficient, the findings revealed that lexical diversity varied significantly across the three writing tasks. The highest scores of MTLD and Voc-D belonged to the narrative, argumentative, and descriptive writing tasks, respectively. Moreover, EFL learners used more advanced words in argumentative task type according to the measurement of CELEX and K3+ level words. However, learners' lexical sophistication in narrative and descriptive writings did not emerge as significantly different when measured by K3+. The significant positive relationship between lexical sophistication and lexical diversity was evident in all three task types.

This study compared the Discussion section of theses written by M.A. English L1 and L2 students regarding lexical sophistication. Certain linguistic features were applied to investigate whether texts written in English L2 had any similarities or differences with native speakers’ (NSs) texts. To achieve this, 20 English L2 theses authored by Iranian M.A. students of Islamic Azad University of Isfahan (Khorasgan), were sampled. As such, 20 English L1 theses written by M.A. English NSs of the same major in the US and UK were randomly downloaded as a comparison corpus of English L1. The corpora were later uploaded to Coh-Matrix that processes text and discourse at different levels of language. Two main statistical procedures for data analysis were the MANOVA (Multivariate Analysis of Variance) and Discriminant Function Analysis. According to Coh-Metrix analysis, the results revealed certain similarities and differences between the corpora. Specifically, more CELEX content words were used in the NSs theses. However, the differences between the two corpora did not reach a statistical significance in terms of other indices of lexical sophistication (i.e., polysemy, concreteness, hypernymy, age-of-acquisition scores, and lexical diversity). Concerning these similarities and differences, academic writing pedagogy and theses writing abilities of Iranian English L2 learners can be improved by applying both word processing tools like Coh-Metrix, appropriate writing strategies and techniques according to the given results of current contrastive corpus study.