"To Whom it May Concern": A Study on the Use of Lexical Bundles in Email Writing Tasks in an English Proficiency Test

Abstract

Lexical bundles are worthy of attention in both teaching and testing writing as they function as basic building blocks of discourse. This corpus-based study focuses on the rated writing responses to the email tasks in the Canadian English Language Proficiency Index Program® General test (CELPIP-General) and explores the extent to which lexical bundles could help characterize the written responses of the test-takers of different English proficiency levels. Three subcorpora of email writing responses were created based on test-takers’ proficiency levels. AntConc 3.4.4 was used to identify 2- to 6-word lexical bundles, which were then manually coded for their discourse functions. The results showed that test-takers of higher proficiency levels used more lexical bundles in terms of both bundle types and tokens compared with test-takers of lower proficiency level. The writing samples of different proficiency levels shared some lexical bundles, and overall they had similar proportional distributions of lexical bundle functions. Nevertheless, noticeable differences among the proficiency levels were observed in the proportional distributions of the subfunctions of stance bundles, discourse organizing bundles, and bundles of other functions. The identification of differential use of lexical bundles can contribute to a better understanding of English learners’ email writing performance. Les expressions figees meritent notre attention tant en enseignement qu’en evaluation de l’ecrit puisqu’elles jouent le role de composantes de base du discours. Cette etude basee sur des corpus porte sur des reponses ecrites par courriel a des tâches et evalue la mesure dans laquelle les expressions figees pourraient aider a caracteriser les reponses ecrites des participants de differentes competences en anglais. Trois sous corpus de reponses ecrites par courriel ont ete crees en fonction des niveaux de competence des participants. AntConc 3.4.4 a servi dans l’identification d’expressions figees composees de 2 a 6 mots qui ont ensuite ete codees manuellement selon leurs fonctions discursives. Les resultats indiquent que les participants de niveaux de competence plus eleves employaient plus d’expressions figees et de types plus varies comparativement aux participants de niveaux de competence plus bas. Les echantillons de textes de differents niveaux de competence partageaient quelques expressions figees et, globalement, avaient des distributions proportionnelles d’expressions figees similaires. Toutefois, des differences notables entre les niveaux de competence ont ete observees dans la distribution proportionnelle des sous fonctions discursives. L’identification des differences dans l’emploi des expressions figees peut contribuer a une meilleure comprehension du rendement des apprenants d’anglais quand ils ecrivent des courriels.