Compiling a Corpus of Taiwanese Students’ Spoken English

Compiling a Corpus of Taiwanese Students’ Spoken English Lan-fen Huang Language Centre, Shih Chien University 200 University Road Nei-men, Kaohsiung 845 Taiwan Lanfen.huang@gmail.com Abstract This paper reports the compilation of a corpus of Taiwanese students’ spoken English, which is one of the twenty sub- corpora of the Louvain International Database of Spoken English Interlanguage (LINDSEI) (Gilquin et al., 2010). LINDSEI is one of the largest corpora of learner speech. The compilation process follows the design criteria of LINDSEI so as to ensure comparability across sub-corpora. The participants, procedures for data collection and process of transcription are all recorded. Sixty third- or fourth-year English majors in Taiwan are interviewed and recorded in English. Each interview is accompanied by a profile which contains information about such learner variables as age, gender, mother tongue, country, English learning context, knowledge of other foreign languages, amount of time spent in English-speaking countries and such interviewer variables as gender, mother tongue, knowledge of foreign languages and degree of familiarity with the interviewees. Another variable, the learners’ English proficiency level based on the results of international standardised tests is collected; this is not available in other sub-corpora of LINDSEI. The participants’ proficiency is similarly distributed across B1 to C1 levels in the Common European Framework of Reference. This paper concludes with a discussion of the contributions and research potential of the corpus. 1 Introduction Corpus compilation, as it has developed, can be traced back to the 1960s (Sinclair, 1991). Research on corpora has mostly focused on written English and contributed a great deal of corpus-based grammatical description and explanation. In contrast, relatively few studies have emerged of corpora of spoken languages, which call for a time-consuming and laborious transcribing process. Yet it is widely acknowledged that this is an area which needs to be further explored (Carter and McCarthy, 1995). A similar trend is found in the field of learner corpora. Learner corpora have been used to study the written language of learners from different backgrounds, in terms of mother tongue. However, little research has been done on the spoken language produced by learners. One of the few major accomplishments in the corpus studies of learners’ spoken English is the compilation of the Louvain International Database of Spoken English Interlanguage (LINDSEI) version 1 (Gilquin et al., 2010), which includes spoken English produced by learners from eleven different first languages (L1s). The present paper first introduces LINDSEI and then reports the compilation process of the Taiwanese sub-corpus, before discussing its contributions and potential for future research. 2 Overview of LINDSEI The LINDSEI project began in 1995 and in 2010 published its first version, which includes sub- corpora of eleven L1s: Bulgarian, Chinese, Dutch, French, German, Greek, Italian, Japanese, Polish, Spanish and Swedish1. It involved 544 informal interviews and roughly one million tokens in total, with an average of 1,949 tokens in each one. About one third of the spoken data comes from the interviewers and two thirds from the learners. In order to have comparable data across sub- …

Content truncated to 500 words